Synthetic, mechanistic, and computational investigations of nitrile-alkyne cross-metathesis

Article abstract of DOI:10.1021/ja800020w

The terminal nitride complexes NW(OC(CF₃)₂Me) ₃(DME) (1-DME), [Li(DME)₂][NW(OC(CF₃) ₂Me)₄] (2), and [NW(OCMe₂CF₃) ₃]₃ (3) were prepared in good yield by salt elimination from [NWCl₃]₄. X-ray structures revealed that 1-DME and 2 are monomeric in the solid state. All three complexes catalyze the cross-metathesis of 3-hexyne with assorted nitriles to form propionitrile and the corresponding alkyne. Propylidyne and substituted benzylidyne complexes RCW(OC(CF₃)₂Me)₃ were isolated in good yield upon reaction of 1-DME with 3-hexyne or 1-aryl-1-butyne. The corresponding reactions failed for 3. Instead, EtCW(OC(CF₃)Me₂) ₃ (6) was prepared via the reaction of W₂(OC(CF ₃)Me₂)₆ with 3-hexyne at 95°C. Benzylidyne complexes of the form ArCW(OC(CF₃)Me₂)₃ (Ar = aryl) then were prepared by treatment of 6 with the appropriate symmetrical alkyne ArCCAr. Three coupled cycles for the interconversion of 1-DME with the corresponding propylidyne and benzylidyne complexes via [2 + 2] cycloaddition-cycloreversion were examined for reversibility. Stoichiometric reactions revealed that both nitrile-alkyne cross-metathesis (NACM) cycles as well as the alkyne cross-metathesis (ACM) cycle operated reversibly in this system. With catalyst 3, depending on the aryl group used, at least one step in one of the NACM cycles was irreversible. In general, catalyst 1-DME afforded more rapid reaction than did 3 under comparable conditions. However, 3 displayed a slightly improved tolerance of polar functional groups than did 1-DME. For both 1-DME and 3, ACM is more rapid than NACM under typical conditions. Alkyne polymerization (AP) is a competing reaction with both 1-DME and 3. It can be suppressed but not entirely eliminated via manipulation of the catalyst concentration. As AP selectively removes 3-hexyne from the system, tandem NACM-ACM-AP can be used to prepare symmetrically substituted alkynes with good selectivity, including an arylene-ethynylene macrocycle. Alternatively, unsymmetrical alkynes of the form EtCCR (R variable) can be prepared with good selectivity via the reaction of RCN with excess 3-hexyne under conditions that suppress AP. DFT calculations support a [2 + 2] cycloaddition-cycloreversion mechanism analogous to that of alkyne metathesis. The barrier to azametalacyclobutadiene ring formation/breakup is greater than that for the corresponding metalacyclobutadiene. Two distinct high-energy azametalacyclobutadiene intermediates were found. These adopted a distorted square pyramidal geometry with significant bond localization.

Full text of DOI:10.1021/ja800020w

Published on Web 06/20/2008
Synthetic, Mechanistic, and Computational Investigations of
Nitrile-Alkyne Cross-Metathesis
Andrea M. Geyer, Eric S. Wiedner, J. Brannon Gary, Robyn L. Gdula,
Nicola C. Kuhlmann, Marc J. A. Johnson,* Barry D. Dunietz, and Jeff W. Kampf
Department of Chemistry, UniVersity of Michigan, 930 North UniVersity AVenue,
Ann Arbor, Michigan 48109-1055
Received January 2, 2008; E-mail: mjaj@umich.edu
Abstract: The terminal nitride complexes NW(OC(CF3)2Me)3(DME) (1-DME), [Li(DME)2][NW(OC(CF3)2Me)4]
(2), and [NW(OCMe2CF3)3]3 (3) were prepared in good yield by salt elimination from [NWCl3]4. X-ray
structures revealed that 1-DME and 2 are monomeric in the solid state. All three complexes catalyze the
cross-metathesis of 3-hexyne with assorted nitriles to form propionitrile and the corresponding alkyne.
Propylidyne and substituted benzylidyne complexes RCW(OC(CF3)2Me)3 were isolated in good yield upon
reaction of 1-DME with 3-hexyne or 1-aryl-1-butyne. The corresponding reactions failed for 3. Instead,
EtCW(OC(CF3)Me2)3 (6) was prepared via the reaction of W2(OC(CF3)Me2)6 with 3-hexyne at 95 °C.
Benzylidyne complexes of the form ArCW(OC(CF3)Me2)3 (Ar ) aryl) then were prepared by treatment of
6 with the appropriate symmetrical alkyne ArCCAr. Three coupled cycles for the interconversion of 1-DME
with the corresponding propylidyne and benzylidyne complexes via [2 + 2] cycloaddition-cycloreversion
were examined for reversibility. Stoichiometric reactions revealed that both nitrile-alkyne cross-metathesis
(NACM) cycles as well as the alkyne cross-metathesis (ACM) cycle operated reversibly in this system.
With catalyst 3, depending on the aryl group used, at least one step in one of the NACM cycles was
irreversible. In general, catalyst 1-DME afforded more rapid reaction than did 3 under comparable conditions.
However, 3 displayed a slightly improved tolerance of polar functional groups than did 1-DME. For both
1-DME and 3, ACM is more rapid than NACM under typical conditions. Alkyne polymerization (AP) is a
competing reaction with both 1-DME and 3. It can be suppressed but not entirely eliminated via manipulation
of the catalyst concentration. As AP selectively removes 3-hexyne from the system, tandem NACM-ACM-
AP can be used to prepare symmetrically substituted alkynes with good selectivity, including an arylene-
ethynylene macrocycle. Alternatively, unsymmetrical alkynes of the form EtCCR (R variable) can be prepared
with good selectivity via the reaction of RCN with excess 3-hexyne under conditions that suppress AP.
DFT calculations support a [2 + 2] cycloaddition-cycloreversion mechanism analogous to that of alkyne
metathesis. The barrier to azametalacyclobutadiene ring formation/breakup is greater than that for the
corresponding metalacyclobutadiene. Two distinct high-energy azametalacyclobutadiene intermediates were
found. These adopted a distorted square pyramidal geometry with significant bond localization.
Introduction
tions of the system to the formation of symmetrical alkynes,
unsymmetrical alkynes, and an arylene-ethynylene macrocycle
Recently, we communicated a new approach for synthesizing
symmetrical alkynes via the cross-metathesis of a sacrificial
alkyne (R′CCR′) with a nitrile (RCN) catalyzed by the metal-
nitride complexes NtW(OC(CF3)2Me)3(DME) (1-DME), [Li-
(DME)2][NtW(OC(CF3)2Me)4] (2), and [NtW(OCMe2CF3)3]3
(3).1 This effected coupling of the R′C and RC units of the
nitrile and alkyne to produce RCCR′, which underwent subse-
quent reactions to form RCCR. Herein, we describe mechanistic
variations in nitrile-alkyne cross-metathesis (NACM) as a
function of the catalyst identity and choice of nitrile as revealed
through synthetic and kinetic investigations. Computational
analyses of NACM using density functional theory (DFT)
indicate nearly isoenergetic tungsten-nitride and tungsten-
alkylidyne catalyst states along with a moderate barrier to
metalacycle formation and subsequent ring opening. Applica-
are also detailed.
Advances in the development of catalysts for alkyne metath-
esis (AM) have sparked an increased interest in the application
of AM to the synthesis of a wide variety of materials.2–5
Previously, application of alkyne metathesis was infrequent, in
part because of the difficulty of synthesizing the required
catalysts.6–9 However, several alkyne metathesis precatalysts can
(2) Schrock, R. R.; Czekelius, C. AdV. Synth. Catal. 2007, 349, 55.
(3) Zhang, W.; Moore, J. S. Angew. Chem., Int. Ed. 2006, 45, 4416.
(4) Furstner, A. Davies, P. W. Chem. Commun. 2005, 2307 and references
therein.
(5) Zhang, W.; Moore, J. S. AdV. Synth. Catal. 2007, 349, 93.
(6) McCullough, L. G.; Schrock, R. R.; Dewan, J. C.; Murdzek, J. C.
J. Am. Chem. Soc. 1985, 107, 5987.
(7) Freudenberger, J. H.; Schrock, R. R.; Churchill, M. R.; Rheingold,
A. L.; Ziller, J. W. Organometallics 1984, 3, 1563.
(8) Listemann, M. L.; Schrock, R. R. Organometallics 1985, 4, 74.
(9) Schrock, R. R.; Clark, D. N.; Sancho, J.; Wengrovius, J. H.; Rocklage,
S. M.; Pedersen, S. F. Organometallics 1982, 1, 1645.
(1) Geyer, A. M.; Gdula, R. L.; Wiedner, E. S.; Johnson, M. J. A. J. Am.
Chem. Soc. 2007, 129, 3800.
9
8984 J. AM. CHEM. SOC. 2008, 130, 8984–8999
10.1021/ja800020w CCC: $40.75
2008 American Chemical Society

Investigation of Nitrile-Alkyne Cross-Metathesis
A R T I C L E S
Scheme 1. NACM via Reversible Interconversion of Nitride and
Alkylidyne Ligands
Figure 1. Impact of alkoxide on relative energies of alkylidyne and nitride
complexes.
now be accessed readily through a variety of methods.2,10–18
These precatalysts range in identity, but in general, the proposed
active species is a metal-alkylidyne complex. As a result of the
increased accessibility of metal-alkylidyne complexes, large
macrocycles, polymers, and natural products were successfully
synthesized via alkyne metathesis.3–5,19 Unfortunately, one of
the drawbacks of alkyne metathesis is the requirement for pre-
existing carbon-carbon triple bonds, which can be difficult to
install. Therefore, an alternative method for the synthesis of
alkynes that utilizes more readily accessible precursors is
desired.
Scheme 2. Known Interconversions of Alkylidyne and Nitride
Complexes
In general, the nitrile functionality can be incorporated into
a molecule more readily than the alkyne moiety.20,21 This
spurred us to develop a method to convert a nitrile into an alkyne
cleanly and efficiently. This transformation necessarily requires
the ability to interconvert alkylidyne and nitride ligands.
Reversibility permits productive catalysis. The proposed cycle
for tungsten-catalyzed NACM is detailed in Scheme 1.
A precedent for the alkylidyne-to-nitride conversion is
provided by Schrock’s work with RCtW(O-t-Bu)3, in which
NtW(O-t-Bu)3 is formed upon treatment with PhCN.22 Ch-
isholm similarly demonstrated the stoichiometric conversion of
PhCtW(OC(CF3)Me2)3 into the trimeric nitrido complex
[NtW(OC(CF3)Me2)3]3.23 Recently, we provided the first
Seeking a catalyst that would permit the reversible formation
of a metal-alkylidyne complex from the corresponding nitride
species, we chose tungsten-alkoxides as a starting point. As a
less electronegative atom than molybdenum, tungsten should
favor ligation of nitride relative to alkylidyne moieties, given
the same set of ancillary ligands.24 More importantly, the greater
spatial extent of the d orbitals in W as compared to Mo should
lead to a diminished barrier to metalacycle formation for W
complexes.24 This effect is noted both experimentally25,26 and
computationally27 for AM. Further catalyst tuning was ap-
proached through variation of the ligand set. Employing weakly
donating ligands such as OC(CF3)2Me and OCMe2CF3 serves
two purposes. First, for a given metal, the use of weaker donor
ligands favors the formation of an alkylidyne relative to the
nitride. That is, for an identical set of ancillary ligands, there is
a greater positive charge at the metal in the nitride complex
than in the alkylidyne complex.24 This charge is better supported
when the ancillary ligands are strong donors. Thus, by stabilizing
the alkylidyne relative to the nitride, the reformation of the
alkylidyne from the nitride would be possible, as illustrated in
Figure 1. Second, Schrock’s investigations of alkyne metathesis
by W complexes revealed that [2 + 2] cycloaddition to a
terminal alkylidyne moiety becomes progressively more favor-
able as the degree of fluorination of the t-butoxide ligands
increases because of increasing Lewis acidity of the W
center.25,26
example of
a conversion in the opposite direction:
NtMo(OC(CF3)2Me)3 undergoes metathesis with 3-heyxne at
high temperatures to produce the corresponding propylidyne
complex EtCtMo(OC(CF3)2Me)3.10 A summary of known
interconversions of alkylidyne and nitride species is provided
in Scheme 2.
(10) Gdula, R. L.; Johnson, M. J. A. J. Am. Chem. Soc. 2006, 128, 9614.
(11) Zhang, W.; Kraft, S.; Moore, J. S. J. Am. Chem. Soc. 2004, 126, 329.
(12) Furstner, A.; Mathes, C.; Lehmann, C. W. J. Am. Chem. Soc. 1999,
121, 9453.
(13) Tsai, Y. C.; Diaconescu, P. L.; Cummins, C. C. Organometallics 2000,
19, 5260.
(14) Bunz, U. H. F. Science 2005, 308, 216.
(15) Fu¨rstner, A.; Mathes, C. Org. Lett. 2001, 3, 221.
(16) Pschirer, N. G.; Bunz, U. H. F. Tetrahedron Lett. 1999, 40, 2481.
(17) Kaneta, N.; Hikichi, K.; Asaka, S.; Uemura, M.; Mori, M. Chem. Lett.
1995, 1055.
(18) Mortreux, A.; Delgrange, J. C.; Blanchard, M.; Lubochinsky, B. J.
Mol. Catal. 1977, 2, 73.
(19) Brizius, G.; Pschirer, N. G.; Steffen, W.; Stitzer, K.; zur Loye, H. C.;
Bunz, U. H. F. J. Am. Chem. Soc. 2000, 122, 12435.
(20) North, M. Nitriles: General Methods and Aliphatic Nitriles. In
ComprehensiVe Organic Functional Group Transformations II, 1st
ed.; Katritzky, A. R., Taylor, R. J. K., Eds.; Elsevier: Amsterdam,
2005; Vol. 3, p 621.
This effect is also borne out by DFT calculations of model
systems, which indicate a stabilizing effect on the transition state
for metalacyclobutadiene formation as the ancillary ligands
(21) Tyrrell, E. Alkynes. In ComprehensiVe Organic Functinal Group
Transformations II, 1st ed.; Katritzky, A. R., Taylor, R. J. K., Eds.;
Elsevier: Amsterdam, 2005; Vol. 1, p 1083.
(24) Nugent, W. A.; Mayer, J. M. Metal-Ligand Multiple Bonds; John
Wiley and Sons: New York, 1988; p 26.
(22) Schrock, R. R.; Listemann, M. L.; Sturgeoff, L. G. J. Am. Chem. Soc.
1982, 104, 4291.
(25) Freudenberger, J. H.; Schrock, R. R. Organometallics 1986, 5, 398.
(23) Chisholm, M. H.; Folting, K.; Lynn, M. L.; Tiedtke, D. B.; Lemoigno,
F.; Eisenstein, O. Chem. Eur. J. 1999, 5, 2318.
(26) Schrock, R. R. Acc. Chem. Res. 1986, 19, 342.
(27) Zhu, J.; Jia, G.; Lin, Z. Organometallics 2006, 25, 1812.
9
J. AM. CHEM. SOC. VOL. 130, NO. 28, 2008 8985

A R T I C L E S
Geyer et al.
Scheme 3. Syntheses of Tungsten-Nitride Complexes
become poorer donors.27 Degenerate N-atom exchange between
nitriles, catalyzed by related complexes of Mo and W (Scheme
2), is thought to proceed via an analogous mechanism;10,28 here
too, the rate is enhanced by a more Lewis acidic metal center.
Accordingly, we anticipated enhanced rates of reaction with the
more highly fluorinated complex NtW(OC(CF3)2Me)3 (1) with
respect to [NtW(OCMe2CF)3]3 (3).
Figure 2. 50% thermal ellipsoid plot of catalyst precursor 1-DME.
Results and Discussion
Syntheses of Tungsten-Nitride Complexes. Our syntheses of
the tungsten-nitride catalysts are depicted in Scheme 3. Treat-
ment of [NtWCl3]4 · 1.1DCE29 (DCE ) 1,2-dichloroethane)
with 16 equiv of LiOC(CF3)2Me and 8 equiv of 1,2-dimethoxy-
ethane (DME) in toluene afforded [Li(DME)2]-
[NtW(OC(CF3)2Me)4] (2) in 63% yield. Subsequent treatment
of 2 at low temperatures with CF3SO3H (HOTf) in diethyl ether
afforded the DME adduct of NtW(OC(CF3)2Me)3, 1-DME, in
78% yield. Alternatively, 1-DME can be synthesized directly
via the salt metathesis of (NtWCl3)4 ·1.1DCE with 12 equiv
of LiOC(CF3)2Me in the presence of 4 equiv of THF at 65 °C
to form a noncrystalline THF adduct of the desired species.
Filtration of the reaction mixture followed by the addition of 4
equiv of DME to the filtrate then afforded crystalline 1-DME
in 60% yield from diethyl ether-pentane. This is the preferred
method for multigram synthesis of 1-DME, owing to the fact
that the reaction of 2 with HOTf under standard conditions
becomes less reliable when the quantity of 2 used exceeds 0.5 g.
Similarly, [NtW(OCMe2CF3)3]3 (3)23 can be accessed via salt
elimination with LiOCMe2CF3 in a 53% isolated yield. Ad-
ditionally, we found that metathesis of W2(OCMe2CF3)6 with
2 equiv of propionitrile at 95 °C (but not at room temperature)30
resulted in complete conversion to 3.
Figure 3. 50% thermal ellipsoid plot of catalyst precursor 2.
Table 1. Selected Bond Lengths and Angles for 1-DME and 2
1-DME
2
Bond distances (Å)
1.680(5) W-N
1.906(3) W-O(17)
1.954(3) W-O(18)
1.949(3) W-O(19)
2.192(4) W-O(20)
2.483(4) N-Li
W-N
1.672(1)
W-O(3)
W-O(5)
W-O(4)
W-O(1)
1.9703(10)
1.9569(10)
1.9665(10)
1.9640(10)
2.063(3)
W-O(2) trans to N
Bond angles (deg)
Structures of Tungsten-Nitride Complexes. Single-crystal
X-ray diffraction confirmed that 1-DME is monomeric, unlike
3, which exists as a trimer.23 A thermal ellipsoid plot (Figure
2) reveals approximately octahedral coordination about W, with
a characteristically short WtN triple bond (Table 1). The DME
ligand is bidentate, spanning sites trans to alkoxide and nitride
ligands. Unsurprisingly, the W-O bond trans to the nitride
ligand is substantially (0.305 Å) longer than that trans to
alkoxide (Table 1), owing to the large trans influence of the
triply bonded nitride ligand.31 Complex 2 was likewise found
to be monomeric, as determined by single-crystal X-ray dif-
fraction. In this case, there were four crystallographically
N-W-O(3)
105.20(14) N-W-O(17)
102.00(14) N-W-O(18)
100.70(14) N-W-O(19)
97.85(14) N-W-O(20)
93.21(13) O(18)-W-O(17)
94.92(12) O(18)-W-O(19)
81.37(12) O(20)-W-O(19)
81.39(12) O(20)-W-O(17)
W-N-Li
102.23(5)
N-W-O(5)
100.17(5)
102.18(5)
100.31(5)
87.81(4)
88.10(4)
87.96(4)
87.53(4)
173.60(10)
N-W-O(4)
N-W-O(1)
O(3)-W-O(5)
O(3)-W-O(4)
O(1)-W-O(4)
O(1)-W-O(5)
independent molecules in each asymmetric unit. The Li(DME)2
unit is directly coordinated to the nitride ligand as depicted in
the thermal ellipsoid plot (Figure 3). Both 1-DME and 2, which
are pale yellow in solution and colorless in the crystalline state,
serve as sources of 1 in reactions with alkynes.
Syntheses of Tungsten-Alkylidyne Complexes. Both 1-DME
and 3 were dissolved in toluene-d8 containing 3-hexyne to test
for formation of the desired propylidyne complexes at room
temperature. When 1-DME was treated with 2 equiv of
(28) Chisholm, M. H.; Delbridge, E. E.; Kidwell, A. R.; Quinlan, K. B.
Chem. Commun. 2003, 126.
(29) Close, M. R.; McCarley, R. E. Inorg. Chem. 1994, 33, 4198.
(30) Freudenberger, J. H.; Pedersen, S. F.; Schrock, R. R. Bull. Soc. Chim.
Fr. 1985, 349.
(31) Nugent, W. A. Metal-Ligand Multiple Bonds; John Wiley and Sons:
New York, 1988; p 156.
9
8986 J. AM. CHEM. SOC. VOL. 130, NO. 28, 2008

Investigation of Nitrile-Alkyne Cross-Metathesis
A R T I C L E S
Scheme 4. Formation of Benzylidyne and Alkylidyne Complexes
from 1-DME
Scheme 5. Formation of Benzylidyne and Alkylidyne Complexes
3-hexyne, an equilibrium mixture composed of 5 mol %
tungstenacyclobutadiene complex W(C3Et3)(OC(CF3)2Me)3 (4)7
and
95
mol
%
propylidyne
complex
EtCt
W(OC(CF3)2Me)3(DME) (5-DME) was obtained after 20 h.1
In the 1H NMR spectra, the organic byproduct of the formation
of 4 and 5-DME, propionitrile, also was observed. In contrast,
treatment of 3 with 10 equiv of 3-hexyne at room temperature
did not result in the formation of detectable quantities of
EtCtW(OCMe2CF3)3 (6). Furthermore, heating of the reaction
mixture to 95 °C for several days resulted in only 9% conversion
to 6.
Taking advantage of the facile metathesis of 1-DME with
alkynes, we prepared several new alkylidyne complexes for
further study. As depicted in Scheme 4, metathesis with
3-hexyne, 1-(4-methoxyphenyl)-1-butyne, and 1-(4-trifluoro-
methylphenyl)-1-butyne led to the successful isolation of 5-DME
(52%), 4-MeO(C6H4)CtW(OC(CF3)2Me)3(DME) (7-DME, 67%),
and 4-F3C(C6H4)CtW(OC(CF3)2Me)3(DME) (8-DME, 57%),
respectively. Unlike with 1-DME, the alkylidyne and benzyli-
dyne analogues of 3 could not be synthesized directly from 3,
as very little conversion to the desired complexes was achieved
even at 95 °C. Instead, EtCtW(OCMe2CF3)3 (6) was obtained
via reaction of 3-hexyne with the ditungsten complex,
W2(OCMe2CF3)6, as previously alluded to by Schrock.30 Next,
6 was converted into the desired benzylidyne complexes
ArCtW(OCMe2CF3)3 [Ar ) C6H4OMe (9) or C6H4CF3 (10)]
via metathesis with the corresponding symmetrical alkynes in
83% and 73% yield, respectively (Scheme 5).
24% yield occurred when 3 was heated at 95 °C in the presence
of 3-hexyne. No other tungsten-containing products were
identified.
Once reversible formation of 1-DME was established, we sought
to exploit this property for the exchange of R and R′ units in an
alkyne (RCtCR) and a nitrile (R′CtN). The nitride complexes
were chosen over the corresponding tungsten-propylidyne com-
plexes as the starting form of the catalyst. This was due to the
facility with which the metal nitrides could be synthesized in
comparison to their propylidyne counterparts. Initial test reactions
were conducted with p-methoxybenzonitrile and 0.5 equiv of
3-hexyne in the presence of 1-DME (5 mol% based on p-
methoxybenzonitrile) at 95 °C. After 21 h, 1H NMR spectroscopy
indicated conversion to bis(4-methoxyphenyl)acetylene (77%),
1-(4-methoxyphenyl)-1-butyne (15%), with some remaining p-
methoxybenzonitrile (8%) along with the N-containing byproduct,
propionitrile (eq 1).
General Considerations for NACM. Because reversible
formation of a nitride complex is required for the desired
catalytic cycle, metathesis of propionitrile with the propylidyne
species was analyzed. Metathesis of 5-DME with propionitrile
at 95 °C led to a 28% conversion to 1-DME in 3 h. Unlike
5-DME, complex 6 was completely converted to 3 upon addition
of propionitrile and heating at 95 °C, although some decomposi-
tion to OdW(OCMe2CF3)4 (11) was noted. This decomposition
likely occurred via C-O bond scission, as unidentified volatile
fluorine-containing materials were observed upon vacuum
transfer of the volatile materials from the reaction mixture. C-O
bond scission in alkoxide ligands is known to afford terminal
oxo complexes.32 Redistribution of the alkoxide ligands is
required for the formation of 11. Successful isolation of 11 in
Since 2 could potentially serve as a source of 1 in solution,
2 also was tested for NACM activity. Complex 2 was found to
be active, albeit with lower yields and longer reaction times
than 1-DME.1 Similarly, given that trace amounts of propylidyne
complex were observed upon prolonged heating of 3 in the
presence of 3-hexyne, 3 was tested for NACM activity.
Successful conversion of the aryl nitrile to aryl alkyne was
observed with 3, although overall yields were lower and the
metathesis reaction occurred at a slower rate than with 1-DME.1
The slower rate of metathesis is consistent with trends in AM
rates, where the rate of metathesis increases as the pKa of
conjugate acids of the alkoxide ligands decreases.7
Surprisingly, eq 1 shows that the symmetrical diaryl alkyne
is the major alkyne produced. This is not the primary product
(32) Mayer, J. M. Polyhedron 1995, 14, 3273.
9
J. AM. CHEM. SOC. VOL. 130, NO. 28, 2008 8987

A R T I C L E S
Geyer et al.
Scheme 6. Three Possible Cycles for Formation of Symmetrical Alkyne
Scheme 7. NACM Cycle I (Formation of ArCtCEt from ArCtN)
of NACM but must arise from a secondary transformation. In
principle, three different triple-bond metathesis cycles can
contribute to the formation of the symmetrical alkyne product.
These are illustrated in Scheme 6. For simplicity, the organic
reactants and products are denoted as a-f. Key metal complexes
are denoted as A-C. This focuses attention on the numerous
metal complexes involved in the three coupled cycles. At the
beginning of the catalytic cycle (cycle I), metathesis of nitrido
complex A with 3-hexyne (a) led to the formation of propyli-
dyne complex B with concomitant release of propionitrile (d).
The unsymmetrical alkyne (e) then was formed initially by
NACM via reaction between the propylidyne complex B and
aryl nitrile (b), with concomitant regeneration of the nitride
complex (Scheme 6, cycle I). The symmetrical alkyne (f) is a
secondary product that can be subsequently formed via either
NACM (cycle II) or ACM (cycle III) of the unsymmetrical
alkyne (e). Only NACM can account for the transformation of
the aryl nitrile (b) into an aryl alkyne (e); ACM can only
equilibrate the various alkynes after they are formed. Thus,
NACM is critical to the complete catalytic cycle.
At this point, it was unclear as to whether the symmetrical
alkyne (f) is formed predominantly through ACM or NACM
or a combination of the two processes. To better understand
the role of ACM and NACM in the generation of symmetrical
alkyne, the aforementioned alkylidyne and benzylidyne com-
plexes ligated with either OC(CF3)2Me (5-DME, 7-DME, and
8-DME) or OCMe2CF3 (6, 9, and 10) were used to investigate
the feasibility and reversibility of each proposed cycle.
Reversibility of NACM and ACM Steps. NACM Cycle I.
NACM Cycle I is depicted separately from the other cycles in
Scheme 7. Table 2a-d summarizes the results of stoichiometric
investigations of the proposed pathways for the formation of
unsymmetrical alkynes via NACM with both 1-DME and 3 at
95 °C (Scheme 7). As expected, each step contributing to the
formation of the unsymmetrical alkyne operates reversibly for
both catalysts. However, the catalyst resting state varies greatly
as a function of the ancillary ligands. Use of the less electron-
donating OC(CF3)2Me ligand leads to preferential formation of
alkylidyne or benzylidyne complexes. In contrast, the nitride
complex is favored when OC(CF3)Me2 ligands are employed.
Furthermore, formation of benzylidyne was favored over that
of propylidyne with both catalysts. Table 2c,d reveals no
significant electronic influence on the catalyst resting state as a
function of the para-substitution on the aryl group of the
substrates. Table 2a examines the formation of the tungsten-
propylidyne (5-DME and 6) complexes and propionitrile upon
treatment of the corresponding tungsten-nitride complexes with
3-hexyne. Table 2b examines the reverse reaction. The fact that
the final compositions differ for these forward and reverse
reactions is a consequence of the operation of two other
processes, in addition to the highlighted triple-bond metathesis.
First, alkyne polymerization (AP) preferentially consumes small
alkyl-substituted alkynes (i.e., 3-hexyne) from the mixture over
time, as noted previously.1 Second, close inspection of the
reaction mixtures reveals some decomposition of the tungsten
complexes, resulting in the formation of oxo complexes of the
form W(O)(OR)4 (11). As noted, C-O bond scission in the
alkoxide ligands appears to be the origin of the oxo ligands.
Ligand redistribution must account for the extra alkoxide ligands
in 11. Not all the tungsten can be accounted for. In comparing
the reversible formation of nitride from alkylidyne species for
the W(OC(CF3)2Me)3 fragment (Table 2a vs b), a large
9
8988 J. AM. CHEM. SOC. VOL. 130, NO. 28, 2008

Investigation of Nitrile-Alkyne Cross-Metathesis
A R T I C L E S
Table 2
Table 3
(a) Nt[W] + p-XC6H4CtCEt f p-XC6H4Ct[W] + EtCtN
(a) Nt[W] + EtCtCEt f EtCt[W] + EtCtN
entry
catalyst
WtN
WdO
WtCR, W(C3R3)
entry
catalyst
X
WtN
WdO
WtCAr
WtCR, W(C3R3)
1
2
3
63
0
26
0
9
100
1
2
3
4
3
OMe
OMe
CF3
75
6
42
13
14
0
33
0
9
80
25
87
0
14
0
1-DMEa
1-DME
3
1-DME
CF3
0
(b) Nt[W] + EtCtCEt r EtCt[W] + EtCtN
entry
catalyst
WtN
WdO
WtCR, W(C3R3)
(b) Nt[W] + p-XC6H4CtCEt f p-XC6H4Ct[W] + EtCtN
1
2
6
89
28
11
0
0
72
entry
catalyst
X
WtN
WdO
WtCAr
WtCR, W(C3R3)
5-DME
1
2
3
4
9
OMe
OMe
CF3
70
42
61
17
7
0
11
0
14
39
21
83
0
19
0
7-DME
10
8-DME
(c) Nt[W] + p-XC6H4CtCEt f EtCt[W] + p-XC6H4CtN
CF3
0
entry
catalyst
X
WtN
WdO
WtCR, W(C3R3)
WtCAr
1
2
3
4
3
OMe
OMe
CF3
75
6
42
13
14
0
33
0
0
14
0
9
80
25
87
1-DME
3
1-DME
(c) Nt[W] + p-XC6H4CtC-p-C6H4X f p-XC6H4Ct[W] + p-XC6H4CtN
entry
catalyst
X
WtN
WdO
WtCAr
CF3
0
1
2
3
4
3
OMe
OMe
CF3
76
46
100
67
17
0
0
7
54
0
1-DMEa
3
(d) Nt[W] + p-XC6H4CtCEt r EtCt[W] + p-XC6H4CtN
1-DME
CF3
0
33
entry
catalyst
X
WtN
WdO
WtCR, W(C3R3)
WtCAr
1
2
3
4
6
OMe
OMe
CF3
60
12
64
13
9
0
5
0
0
10
4
25
78
19
87
5-DME
6
5-DME
(d) Nt[W] + p-XC6H4CtC-p-C6H4X r p-XC6H4Ct[W] + p-XC6H4CtN
entry
catalyst
X
WtN
WdO
WtCAr
CF3
0
1
2
3
4
9
OMe
OMe
CF3
90
49
86
65
10
0
14
0
0
51
0
a Room temperature.
7-DME
10
8-DME
Scheme 8. NACM Cycle II (Secondary Formation of ArCtCAr)
CF3
35
a Room temperature.
Scheme 9. ACM (Cycle III, Formation/Consumption of ArCtCAr)
difference in catalyst resting state is observed. The conversion
of the nitride to the alkylidyne species readily occurs at room
temperature. In contrast, the conversion of the alkylidyne to
the nitride unit is slow and occurs only at elevated temperatures.
Trace amounts of nitride complex are probably formed un-
der these conditions, with 3-hexyne being removed from the
system via slow AP, which in turn shifts the solution composi-
tion. Thus, the same equilibrium is not established in these two
cases, owing to varying quantities of CEt units present in the
various soluble species. The impact of AP on the system is
discussed in greater detail later.
NACM Cycle II (Scheme 8). Scheme 8 focuses on NACM
cycle II. Similar catalyst resting state patterns were observed
when analyzing the steps of NACM to form the symmetrical
alkyne at 95 °C, as shown in Table 3a-d. With 3, larger
amounts of the terminal oxo complex 11 were generally
observed at extended reaction times under heating. Interestingly,
unlike NACM cycle I (to form the unsymmetrical alkyne), one
segment of cycle II appeared to be irreversible with catalyst 3.
As indicated in Table 3c (entry 3), treatment of 3 with bis(4-
trifluoromethyl)acetylene resulted in no conversion to 10. The
corresponding entry in Table 3d indicates that the reverse
reaction occurs essentially quantitatively except for some catalyst
decomposition. Only slight conversion is observed with the
p-methoxy derivative, but entry 1 in Table 3c,d clearly indicates
reversibility in this latter case. In contrast, the use of 1-DME
affords substantial quantities of both nitride and alkylidyne
complexes at equilibrium, without the formation of the terminal
oxo decomposition product. Once again, differential AP and
C-O bond scission prevents establishment of the same equi-
librium for forward and reverse reactions in the three cases
where both forward and reverse reactions operate.
ACM (Cycle III, Scheme 9). The ACM cycle (cycle III) is
highlighted in Scheme 9. Table 4a-d summarizes the results
obtained for cycle III, tungsten-catalyzed ACM. Tungsten-
catalyzed ACM is well-studied both experimentally and
computationally,5,7,27 so it was unsurprising to find that these
steps are reversible at room temperature. An analysis of the steps
of ACM in the production of the symmetrical alkyne revealed
that the two catalyst systems behaved similarly at room
temperature. In the case of the catalyst based on OC(CF3)Me2
ligands, the ratio of benzylidyne to propylidyne complexes was
often not determined because of the presence of multiple
1
overlapping resonances in the H NMR spectra. These arose
from other unidentified species (likely metalacycles) that
appeared upon cooling the samples. As a result, only the
presence of these species in solution is indicated in Table 4a-d.
9
J. AM. CHEM. SOC. VOL. 130, NO. 28, 2008 8989

A R T I C L E S
Geyer et al.
Table 4
(a) EtCt[W] + p-XC6H4CtCEt f p-XC6H4Ct[W] + EtCtCEt
Scheme 10. Formation of ACM versus NACM Products
entry
catalyst
X
WtCR, W(C3R3)
WtCAr
1
2
3
4
6
OMe
OMe
CF3
present
31
35
present
69
65
5-DME
6
5-DME
CF3
21
79
(b) EtCt[W] + p-XC6H4CtCEt r p-XC6H4Ct[W] + EtCtCEt
entry
catalyst
X
WtCR, W(C3R3)
WtCAr
1
2
3
4
9
OMe
OMe
CF3
present
52
33
present
48
67
7-DME
10
8-DME
CF3
49
51
(c) EtCt[W] + p-XC6H4CtC-p-C6H4X f p-XC6H4Ct[W] + p-XC6H4CtCEt
entry
catalyst
X
WtCR, W(C3R3)
WtCAr
1
2
3
4
6
OMe
OMe
CF3
trace
6
0
100
93
100
84
5-DME
6
5-DME
CF3
16
(d) EtCt[W] + p-XC6H4CtC-p-C6H4X r p-XC6H4Ct[W] + p-XC6H4CtCEt
entry
catalyst
X
WtCR, W(C3R3)
WtCAr
1
2
3
4
9
OMe
OMe
CF3
present
20
trace
0
present
80
100
7-DME
10
8-DME
CF3
100
ylbenzonitrile, bis(4-methyoxyphenyl)acetylene, and 3-hexyne
in the presence of 1-DME or 3 at 95 °C in toluene was
1
monitored via H NMR spectroscopy (Scheme 10). An equi-
Note that as usual, the difference in ratios in catalyst resting
states for the forward and reverse reactions is due to the lack
of well-defined equilibrium conditions as a result of the
competing irreversible AP. Overall, the favored resting state of
the catalysts was the benzylidyne complex. In contrast with
NACM (cycles I and II), all ACM reactions occurred rapidly
at room temperature. Furthermore, the identity of the para-
substituent on the aryl substrate did not appear to exert a
significant influence on the catalyst resting state with ACM.
Overall, these stoichiometric studies suggest three major
conclusions. First, the large difference in reaction rate between
NACM and ACM at room temperature implicates ACM as the
primary process by which the symmetrical alkyne is formed
over NACM (i.e., the symmetrical diaryl alkyne is formed
principally via cycle III and not cycle II, at least in the case of
ligation by OC(CF3)2Me). Second, given the rapidity of ACM
relative to NACM irrespective of the alkoxide ligand used, the
difference in the ratio of symmetrical to unsymmetrical alkynes
observed with the two catalysts (1-DME vs 3) under NACM
conditions is likely due to the difference in the resting states of
the catalysts. In particular, the use of OC(CF3)Me2 ligands
suppresses ACM by virtue of the fact that the NACM catalyst
resting state is the terminal nitride complex. This means that
the alkylidyne complexes required for ACM are present in only
very small quantities. When the alkoxide ligands used are
OC(CF3)2Me, a much larger fraction of the catalyst exists in
the alkylidyne form, thus favoring rapid ACM. Last, although
electronic influences of the aryl substrate do not appear to have
a strong impact on the catalyst resting state in ACM, more subtle
effects did render one step in NACM irreversible with catalyst
3, underscoring the complexity of the system.
librium mixture of ACM products was observed within 10 min
when catalyst 1-DME was used; there was no evidence of
NACM products at this time.1 In the case of catalyst 3, only
ACM products were noted after 10 min, but the reaction had
not yet reached equilibrium. After 3 h, NACM products began
to predominate in the reaction mixtures for both catalysts. This
revealed that ACM occurred at a much faster rate than NACM
for both catalysts under these conditions. Therefore, the overall
process for producing a symmetrical alkyne can be described
as NACM to produce an unsymmetrical alkyne followed by
rapid ACM to generate the symmetrical alkyne as depicted in
Scheme 11.
DFT Calculations of NACM Mechanism. Calculations on a
model system support a relatively low-energy [2 + 2]
cycloaddition-cycloreversion mechanism for NACM (Scheme
12). The calculations were carried out at the DFT level,
employing the commonly used B3LYP functional33–35 and the
LACVP* basis set.36 The Jaguar 6.5 package37 was used to
implement the calculations. Coordinates for all atoms in the
optimized structures and associated bond lengths and angles are
listed in the Supporting Information. The experimental
OC(CF3)xMe3-x (x ) 1, 2) ligands were modeled by OMe. No
constraints were employed in calculating the transition states;
these stationary points involve only a single imaginary fre-
quency. All transition states were verified to form the corre-
sponding reactant and product by geometry optimization along
(33) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
(34) Vosko, S. H.; Wilk, L.; Nusair, M. Can. J. Phys. 1980, 58, 1200.
(35) Lee, C. T.; Yang, W. T.; Parr, R. G. Phys. ReV. B: Condens. Matter
Mater. Phys. 1988, 37, 785.
To confirm the relative facility of NACM and ACM, another
substrate set was examined. Here, the reaction of 3,5-dimeth-
(36) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299.
(37) Schrodinger, L. L. C. Jaguar 6.5; New York, 2005.
9
8990 J. AM. CHEM. SOC. VOL. 130, NO. 28, 2008

Investigation of Nitrile-Alkyne Cross-Metathesis
A R T I C L E S
Scheme 11. Preferred Pathways for Formation of Symmetrical
Alkynes
Figure 4. Calculated vibrational transition state motions. Me groups of
OMe ligands were removed for clarity if associated atoms showed no
significant motion during the transition state.
account for significant molecular motion in this transition state.
This transition state is the maximum energy on the reaction
coordinate diagram and thus controls the rate of the NACM
reaction.
TS 2 is the transition state that connects the two distinct
azametalacyle intermediates Int 1 and Int 2. The geometry of
TS 2 is intermediate between the trigonal bipyramidal (tbp) and
square pyramidal (sp) extremes. In this case, use of τ is
somewhat deceptive, owing to the constraints of the azameta-
lacyclobutadiene ring. The ring enforces a small N-W-CR
angle (in TS 2, 79.9°), which necessarily leads to at least one
large (>140°) angle in the N-CR-Oe plane. Consequently, the
maximum value of τ possible in this case is 0.67 rather than 1.
Examination of Scheme 12 reveals that it is most instructive to
regard TS 2 as the tbp structure that connects the two sp
azametalacycles Int 1 and Int 2. In TS 2, the Oa-W-Oa angle
assumes its maximum value (165.6°), dramatically larger than
the corresponding angle in Int 1 (144.5°) or Int 2 (148.4°). In
TS 2, these O atoms define the axial sites in the distorted tbp
structure.
The atomic motions involved in the Int 1 f TS 2 f Int 2
conversion are best described as consisting of the methoxide
ligand OeMe moving from a basal site in sp Int 1, where it is
trans to CR, to a basal site trans to the N atom in sp Int 2 by
passing between the other two OMe ligandss a “swing-through”
mechanism. The bonds in the metalacycle change as shown in
Figure 4 such that the “apical” atom in the sp structures has
the greater multiple-bond character in its interaction with W.
The vibrational motions described by the single imaginary
frequency in TS 2 consist of an “umbrella inversion” of the
CtO3 unit about W as the Oa-W-Oa angle opens to permit
OeMe to pass through. Metrical parameters for TS 2 are more
similar to those of Int 1 than Int 2; TS 2 resembles Int 1 more
closely than Int 2. This also is reflected in the (admittedly
deceptive) τ parameters for these three structures.
TS 3 defines the [2 + 2] cycloreversion to yield the tungsten-
ethylidyne complex and acetonitrile. As seen for TS 1, the
nitrogen and two carbon atoms of azatungstenacyclobutadiene
show significant motion in the transition state vibration; some
rotation about the W-Oe bond also contributes to this transition
state. To the extent that W(tN)(OMe)3 is electronically similar
to W(tCMe)(OMe)3, the principle of microscopic reversibility
requires that the entering/departing organic substrates in TS 1
and TS 3 must approach/leave the WtE (E ) N, CMe) bond
in analogous orientations. This criterion is met by these two
transition states, wherein each E occupies the apical position
in an approximately square-pyramidal structure. This in turn
necessitates the swing-through interconversion of sp Int 1 with
sp Int 2 via tbp TS 2.
the associated reaction coordinate. Energy values are reported
as ∆G values that have been corrected for zero point energies.
Examination of Scheme 12, which corresponds to the tungsten
nitride f tungsten alkylidyne transformation in Scheme 7,
reveals several important points. Two distinct azatungstena-
cyclobutadiene intermediates (Int 1 and Int 2) are discernible.
These are relatively high-energy species, not more than 5.1 kcal
mol-1 more stable than the corresponding transition states for
the approach of MeCtCMe to W(tN)(OMe)3 (TS 1) or the
approach of MeCtN to W(tCMe)(OMe)3 (TS 3). Both Int 1
and Int 2 feature significant bond localization in the W-N-C-C
ring. The calculated metrical parameters correspond approxi-
mately to WdNsCdCsW bonding in Int 1 in contrast to the
WsNdCsCdW bond alternation in Int 2. These intermediates
are best described as having distorted square pyramidal struc-
tures (τ ) 0.34 and 0.14, respectively38) with the three
methoxide ligands occupying basal positions. In Int 1 and Int
2, the atom with the tungsten-element double bond occupies
the apical position. The trans influence is a likely explanation
for this preference.
The reactants NtW(OMe)3 + MeCtCMe are isoenergetic
with products MeCtW(OMe)3 + MeCtN in this model
system, which, together with reasonable activation barriers,
explains the reversibility of the NACM reaction. Note that in
the experimental systems, the weaker electron-donating ability
of the fluorinated alkoxide ligands as compared to OMe (the
model ligands) should decrease the barriers to metalacycle
formation and stabilize the alkylidyne complex relative to its
nitride analogue. Indeed, in the experimental systems, the use
OCMe2CF3 ligands leads to the nitride complex as the only
detectable W-containing species under catalytic conditions. In
contrast, alkylidyne complexes are generally favored when the
ancillary ligands are OC(CF3)2Me.
Analysis of the transition state motions allows for better
understanding of the mechanism (Figure 4). TS 1 defines the
formation of the azatungstenacyclobutadiene via [2 + 2]
cycloaddition of alkyne to the tungsten-nitride complex. Only
the nitrogen and two carbon atoms that form the azametalacycle
Directly comparable DFT calculations for (degenerate) alkyne
(38) Addison, A. W.; Rao, T. N.; Reedijk, J.; Vanrijn, J.; Verschoor, G. C.
J. Chem. Soc., Dalton Trans. 1984, 1349.
metathesis of the model complex W(tCMe)(OMe)3 with
9
J. AM. CHEM. SOC. VOL. 130, NO. 28, 2008 8991

A R T I C L E S
Geyer et al.
Scheme 12. Calculated Structures and Relative Gibbs Energies (kcal mol-1) for Nitride-Alkylidyne Interconversion in a Model Systema
a H atoms of OMe ligands have been removed for clarity.
MeCtCMe were reported recently.27 In that case, two distinct
metalacycles with bond alternation also were found,39 although
they are substantially more stable than the azametalacycles that
we calculate for NACM. Likewise, the energy of the highest-
lying transition state in that model ACM system is only 22.3
kcal mol-1 greater than the reactants,27 as compared to 29.6
kcal mol-1 in the case of NACM. This agrees qualitatively with
our experimental findings that ACM proceeds much more
rapidly than NACM with current catalysts.
Direct Monitoring of Catalytic Reaction Progress. The catalyst
resting state and reaction progress were monitored via NMR
spectroscopy to examine the practical consequences of the differ-
ence in relative stability of the metal-nitride and metal-alkylidyne
forms of catalysts 1-DME and 3. 1H NMR spectroscopic monitor-
ing of most NACM reactions proved unsuitable because of
difficulties in deconvoluting multiple overlapping resonances.
Therefore, these studies involved the metathesis of p-trifluorom-
ethylbenzonitrile and 3-hexyne at 95 °C and were monitored by
19F NMR spectroscopy. Spectra were collected every 5 min over
a 2 h reaction time for catalysts 1-DME and 3.
Owing to technical limitations as well as the complexity of the
coupled NACM-ACM-AP processes operating in these systems,
these initial studies have resisted attempts to extract rate constants.
Nevertheless, these 19F NMR studies revealed several important
qualitative differences in reactivity between the two catalysts,
1-DME and 3. The rate of formation of the symmetrical alkyne
from an unsymmetrical alkyne was found to be much more rapid
with 1-DME than with 3. The symmetrical and unsymmetrical
alkyne distribution was nearly that established by ACM alone
throughout the reaction catalyzed by 1-DME. In contrast, the
reaction with 3 resulted in a buildup of an unsymmetrical alkyne
prior to its conversion to a symmetrical alkyne; the alkynes were
not in the equilibrium ratio expected on the basis of rapid ACM.
Furthermore, the catalyst resting state of 1-DME under these
catalytic conditions consists of a mixture of 1-DME, 5-DME, and
8-DME, with gradual funneling toward the benzylidyne complex
8-DME as the reaction proceeds. Examination of the spectra
obtained for the reaction catalyzed by 3 does not permit unambigu-
ous assignment of the resting state in this system. However, it is
not 3, 6, or 10. Instead, the catalyst resting state under these
conditions appears to be an adduct of 3, as no aryl or alkyl peaks
corresponding to benzylidyne or alkylidyne complexes were
1
observed in the H NMR spectrum. Additionally, the aforemen-
tioned studies of NACM pathways further support a nitride complex
as the catalyst resting state. The presence of a nitride complex as
the sole catalyst resting explains the greater accumulation of the
unsymmetrical alkyne. This is because ACM requires a metal-
alkylidyne catalyst, which is present only in very low concentrations
when the bulk of the catalyst exists in the nitride form. Thus, the
use of 3 results in significant decoupling of the NACM and ACM
processes. This suppresses the rate of formation of the secondary
product, the diaryl alkyne.
Preferential Formation of Symmetrical Alkynes. From our
initial study of NACM (eq 1), it can be noted that the alkyne
product is largely composed of the symmetrical alkyne. While
ACM accounts for the formation of the symmetrical species, it
cannot by itself account for the selectivity of the system. There
are two reasonable explanations for the preferential formation
of the symmetrical alkyne. From a thermodynamic perspective,
the formation of an unsymmetrical alkyne from 3-hexyne and
benzonitrile is enthalpically favored in the gas phase: ∆H° )
-5.8 kcal mol-1 (eq 2).40 More importantly, the formation of
the symmetrical alkyne from the same precursors is more than
twice as favorable, with ∆H° ) -13.2 kcal mol-1 (eq 4)40
(39) Metalacycles are equivalent due to the degenerate nature of the ACM
reaction modeled.
9
8992 J. AM. CHEM. SOC. VOL. 130, NO. 28, 2008

Investigation of Nitrile-Alkyne Cross-Metathesis
A R T I C L E S
Scheme 13. Harnessing AP to Form Symmetrical Alkynes
In addition to the thermodynamic preference of the system,
a concurrent side reaction, AP, also favors the formation of the
symmetrical alkyne. AP shifts the equilibrium of the system
toward the formation of symmetrical alkynes by removing
3-hexyne from the reaction mixture as a polymer. Indirect
evidence of AP can be found under conditions where complete
consumption of 3-hexyne occurs with incomplete conversion
of the nitrile into alkyne, suggesting that some of the 3-hexyne
must have been removed from the system via an alternative
mechanism. The direct evidence, polymer, is observed as
copious insoluble material under the reaction conditions.
Effects of Catalyst Loading and Temperature. Although AP
drives the formation of symmetrical alkynes, it can also be
detrimental to the system in that it consumes 3-hexyne without
producing the desired alkyne product. Accordingly, the system
was optimized in toluene at 95 °C with 5 mol % catalyst
loadings and 1 equiv of 3-hexyne to determine the conditions
that would favor NACM over AP. Optimal NACM yields and
rates were obtained when catalyst concentrations of 6 and 35
mM 1-DME and 3, respectively, were employed.1 Additional
studies with the more reactive catalyst, 1-DME, were completed
in which the temperature and catalyst loading in toluene were
varied (eq 5, n ) 1).1 As the temperature was decreased,
reaction yields and rates decreased dramatically. Meanwhile,
increasing catalyst loadings led to increased rates of reaction,
as anticipated
be some form of the nitride complex. If so, the low concentration
of the metal-alkylidyne complex must preclude AP under the
tested conditions. It has been suggested that AP occurs via ring
expansion of metalacyclobutadiene complexes,7,11 although there
are other possibilities such as the formation of metal-alkylidene
complexes.41,42
Effect of Excess 3-Hexyne. The impact of the quantity of
3-hexyne on yields and rates of the reaction defined by eq 5 in
toluene also was investigated.1 Use of a moderate excess of
3-hexyne (2-3 equiv) with nitriles that otherwise reacted
sluggishly increased overall rates and yields of alkyne products.
Furthermore, reactions could be performed at lower temperatures
with excess 3-hexyne to afford increased yields of the desired
alkynes. Naturally, as the amount of 3-hexyne introduced into
the system was increased, the selectivity for formation of
symmetrical alkynes greatly decreased as a direct result of the
ACM equilibrium. When 1-DME was used as the catalyst, good
yields of the symmetrical alkyne could nevertheless be obtained
in these cases by taking advantage of AP as described next.
Consequences of AP. As expected, reaction mixtures contain-
ing even a slight excess of 3-hexyne often afforded increased
quantities of unsymmetrical alkynes relative to symmetrical
alkynes. However, simply removing the volatiles from the
system after complete conversion of the nitrile to a mixture of
alkynes, followed by resubjection of the resulting residue to
standard reaction conditions (toluene, 95 °C), resulted in nearly
complete conversion of the unsymmetrical alkyne to the
symmetrical alkyne via tandem ACM-AP with 1-DME (Scheme
13). Interestingly, with catalyst 3, no conversion of the unsym-
metrical alkyne to the symmetrical alkyne was observed under
these reaction conditions. The inability to shift the product ratio
with 3 is likely due to the catalyst resting state in the residue.
As noted previously, the resting state of catalyst 3 appears to
Comparison of Catalysts. The divergent behavior of the
reaction mixtures with respect to tandem ACM-AP following
apparent completion of NACM for the two cases of catalysis
by 1-DME and catalysis by 3 prompted us to examine this
behavior more closely. After the NACM reaction catalyzed by
1-DME had neared its end point as judged by 1H NMR
spectroscopy, to the resulting mixture of alkynes and propioni-
trile was added excess propionitrile. The NMR tube was then
resealed and heated to 95 °C. Reformation of 18% aryl nitrile
after 23 h clearly indicated the reversibility of NACM (Scheme
14) in this case. In contrast, when 3 was used as the catalyst,
extensive broadening of the NMR signals owing to polymer
formation necessitated filtration and removal of volatile materials
from the filtrate. Upon reconstitution in toluene and addition
of propionitrile to the resulting product mixture, the reformation
of aryl nitrile did not occur. Under these conditions, the “back
reaction” of the alkynes ArCtCEt and ArCtCAr with EtCN
did not occur; NACM was irreversible with this catalyst. This
result is consistent with our earlier observation that the resting
state of 3 appears to be a nitride complex, as indicated by the
buildup of unsymmetrical alkyne with 3 (diminished ACM) and
the much more sluggish NACM activity of catalyst 3 in
comparison to 1-DME. It also is consistent with failure to
observe any reaction of 3 with ArCtCAr under stoichiometric
conditions (Table 3c, entry 3).
(41) Mortreux, A.; Petit, F.; Petit, M.; Szymanskabuzar, T. J. Mol. Catal.
(40) NIST Chemistry WebBook-National Institute Standard Reference
Database Number 69; June 2005 release. http://webbook.nist.gov/
chemistry/
A: Chem. 1995, 96, 95.
(42) Masuda, T.; Sasaki, N.; Higashimura, T. Macromolecules 1975, 8,
717.
9
J. AM. CHEM. SOC. VOL. 130, NO. 28, 2008 8993

A R T I C L E S
Geyer et al.
Scheme 14. Reversibility Test
Ketones and aldehydes destroyed the catalysts (Table 6,
entries 1 and 2). However, testing of acetals and ketals revealed
an intriguing difference between 1-DME and 3. The more Lewis
acidic 1-DME was deactivated by these substrates, whereas 3
produced the desired materials (Table 5, entries 10 and 12). A
similar departure was noted in reactions of t-butyl ester with
1-DME and 3: these substrates were tolerated by 3 but not by
1-DME (Table 5, entry 11). This illustrates that catalyst
destruction can be overcome by tuning the Lewis acidity of the
tungsten center with the appropriate ancillary ligands. Interest-
ingly, thiophenes, which were previously incompatible with
tungsten-based alkyne metathesis catalysts,5,43 afforded products
in excellent yield. Overall, however, the functional group
tolerance of NACM catalysts 1-DME and 3 is more restricted
4
than that of the alkyne metathesis catalyst t-BuCtW(O-t-Bu)3
as a result of the increased Lewis acidity of the metal center
needed for NACM activity.
Solvent Studies. To determine the optimal conditions for
NACM, a variety of solvents was investigated at standard
conditions of 95 °C using 5 mol % of 6 mM 1-DME with 2
equiv of anisonitrile and 1 equiv of 3-hexyne (eq 5 and Table
6). Toluene is the optimal solvent for reactions, typically
exhibiting the greatest preference for formation of the sym-
metrical alkyne as well as the most rapid rate of reaction.
Reactions performed in benzene and bromobenzene exhibit a
decreased alkyne preference and an increase in 3-hexyne
polymerization, as indicated by the complete consumption of
3-hexyne with anisonitrile yet remaining. Unlike toluene and
bromobenzene, dichloromethane affords similar yields of NACM
products. However, no alkyne selectivity was observed. Fur-
thermore, the necessarily lower reaction temperature necessitates
a much longer reaction time. Reactions performed in THF
proceed extraordinarily slowly. This is most likely a result of
coordination of THF to the catalyst. Chloroform and 1,2-
dichloroethane afford only moderate yields of alkyne products
as a result of relatively rapid catalyst decomposition in these
solvents.
This suggests that the slowest step in NACM catalyzed by 3
is the formation of the propylidyne complex 6 upon reaction of
3 with 3-hexyne. Other alkynes such as ArCtCEt and ArCtCAr
are even less reactive toward 3, such that they fail to yield any
alkylidyne complex under the conditions employed. This is not
the case when 1-DME is used, however. Formation of the
propylidyne complex 5-DME upon reaction with 3-hexyne is
relatively facile, taking place at room temperature. Significantly
higher temperatures are needed to complete the NACM cycle,
thus indicating that a different step is slow.
Preparation of Alkynes by NACM-ACM or NACM-ACM-
AP. By appropriately harnessing the varied reactivity seen based
on the quantity of 3-hexyne employed as well as the ability to use
AP to drive to symmetrical products, NACM can be used for the
preparation of either unsymmetrical or symmetrical alkynes with
good selectivity (Scheme 15). For instance, standard NACM
conditions with 1-DME, 3,5-bis(trifluoromethyl)benzonitrile, and
1 equiv of 3-hexyne resulted in the initial formation of 1-(butyn-
1-yl)-3,5-bis(trifluoromethyl)benzene in 80% yield. Following
removal of volatiles and subjection to AP conditions, the
reaction mixture afforded bis(3,5-bis(trifluoromethyl)phenyl-
)ethyne in 95% conversion. Similar results were obtained in
the reactions of 3,5-dimethylbenzonitrile with 2 equiv of
3-hexyne and 4-cyanostyrene with 1.5 equiv of 3-hexyne
(Scheme 15).
Functional Group Tolerance. Tables 5 and 6 summarize an
assay of the functional group tolerance of the NACM catalysts
1-DME and 5.1 Several nitriles were tested for functional group
compatibility with the catalysts. Those substrates containing
alkyl, alkyl halide, vinyl, and many substituted aryl groups
resulted in good yields (Table 5). Unfortunately, many Lewis
basic substrates including pyridines, most amines, alcohols, and
nitroarenes were found to deactivate the catalysts, probably via
strong coordination to the W center (Table 6). Table 6 shows
that there are at least two different modes of catalyst deactiva-
tion. In many cases (Table 6: entries 2, 4, 5, 7, and 9–11),
conversion of 1-DME to 5-DME occurred stoichiometrically,
but 1-DME was not reformed; therefore, NACM did not occur.
Addition of anisonitrile to these reaction mixtures confirmed
that the catalyst had been deactivated. In other cases (Table 6:
entries 2, 4, 5, 7, and 9–11), 5-DME was never formed, but the
catalyst was likewise deactivated. Some bulky substrates (entries
13 and 14 in Table 6) failed to react but did not destroy the
catalyst.
Preparative Application of NACM. To demonstrate the
synthetic utility of NACM, we examined the syntheses of some
alkynes on a 0.1 mmol scale, as depicted in eqs 6 and 7.1 We
also began to investigate the application of NACM to the
synthesis of large molecules. To this end, we completed the
synthesis of the cyclic conjugated arylene ethynylene 12 as
depicted in eq 8. Compound 12 and other arylene ethynylene
macrocycles are of interest for applications in nanomaterials
and nanodevices.44 Use of NACM to prepare 12 streamlines
its synthesis by eliminating two steps (one of which requires
Pd) and affords the desired product in slightly improved overall
yield as compared to the best alternative route.45 1H NMR
monitoring of the course of the NACM reaction to prepare 12
provided evidence for initial oligomerization prior to macrocycle
formation. This situation is frequently observed in ring-closing
metathesis reactions.46,47
Conclusion
In summary, the tungsten-nitride complexes 1-DME, 2, and
3 and their alkylidyne analogues 5-DME and 6 catalyze the
(43) Furstner, A.; Seidel, G. J. Organomet. Chem. 2000, 606, 75.
9
8994 J. AM. CHEM. SOC. VOL. 130, NO. 28, 2008

Investigation of Nitrile-Alkyne Cross-Metathesis
A R T I C L E S
Scheme 15. Procedures for Selective Formation of Symmetrical versus Unsymmetrical Alkynes
Table 5. Alkyne Formation
95 °C, toluene. a catalyst ) 3.
Many substrates afforded good conversion to alkynes. However,
Lewis basic functionalities usually precluded reactivity. Interest-
ingly, 1-DME displays an unusual tolerance of thiophenes with
respect to previously known tungsten-based AM catalysts. We
are currently pursuing the development of catalysts that display
enhanced activity and functional group tolerance, including non-
W-based catalysts.
cross-metathesis of internal alkynes with nitriles. DFT calcula-
tions on a model system indicate that a [2 + 2] cycloaddition-
cycloreversion mechanism is feasible. Variation of ancillary
alkoxide ligand identity leads to changes in the NACM catalyst
resting state: use of OC(CF3)Me2 affords nitride 3, whereas use
of OC(CF3)2Me favors benzylidyne/alkylidyne species. All steps
in NACM are reversible for the reaction of 3-hexyne with
anisonitrile catalyzed by 1-DME, unlike with 3.
Owing to the rapid ACM equilibrium relative to NACM, use
of excess 3-hexyne in NACM reactions favors the formation
of unsymmetrical alkynes. Taking advantage of AP, tandem
NACM-AM-AP with 1-DME leads to the formation of the
symmetrical alkyne with increased selectivity. Substrate toler-
ance is catalyst-dependent, with 3 exhibiting the broadest
functional group compatibility, although 1-DME is more active.
Experimental Procedures
General Procedures. 1H NMR spectra were recorded at 499.909
MHz on a Varian Inova 500 spectrometer, 399.967 MHz on a
Varian Inova 400 spectrometer, or 300.075 MHz on a Varian Inova
300 spectrometer and referenced to the residual protons in C6D6
(7.15 ppm), toluene-d8 (2.09 ppm), CD2Cl2 (5.32 ppm), and CDCl3
(7.26 ppm). 13C NMR spectra were recorded at 100.587 MHz on
a Varian Inova 400 spectrometer or at 75.465 MHz on a Varian
Inova 300 spectrometer and were referenced to naturally abundant
13C nuclei in C6D6 (128.00 ppm), CDCl3 (77.16 ppm), or CD2Cl2
(54.00 ppm). 19F NMR spectra were recorded at 282.384 MHz on
a Varian Inova 300 spectrometer or 376.326 MHz on a Varian Inova
9
J. AM. CHEM. SOC. VOL. 130, NO. 28, 2008 8995

A R T I C L E S
Geyer et al.
Table 6. Incompatible Substratesa
bis(4-methoxyphenyl)acetylene1 were prepared according to litera-
ture procedures. LiOCMe2CF3 was prepared in a manner analogous
to that used for the preparation of LiOC(CF3)2Me. NMR solvents
were obtained from Cambridge Isotope Laboratories and were dried
over 4 Å molecular sieves for at least 24 h. DME was obtained
anhydrous from Aldrich and was further dried over 4 Å molecular
sieves for 48 h and run through a plug of alumina before use.
1-Bromotetradecane, sodium hydride (60% in mineral oil), bro-
mobenzene, copper(I) cyanide, and 3-hexyne were obtained from
Acros. Propionitrile and 3,6-dibromocarbazole were obtained from
Aldrich. Propionitrile and 3-hexyne were dried for 24 h using 4 Å
molecular sieves. DMF was dried for 1 month over 4 Å molecular
sieves. All other reagents were used as received.
Catalyst Syntheses. NW(OC(CF3)2Me)3(DME) (1-DME).
[NtWCl3]4 ·1.1DCE (4.00 g, 3.02 mmol) and LiOC(CF3)2Me (6.81
g, 36.2 mmol, 12.0 equiv) were slurried in toluene (40 mL) in a
bomb flask. THF (982 µL, 12.1 mmol, 4 equiv) was added via
syringe, and the bomb flask was sealed. The reaction mixture was
heated with stirring for 19.5 h at 65 °C. The reaction mixture was
heated to nearly boiling and filtered through Celite. The Celite was
washed with hot toluene (40 mL). The volatiles were removed in
vacuo from the filtrate. The resulting residue was taken up in Et2O
(12 mL) and DME (1.25 mL, 12.07 mmol, 4 equiv). Pentane (8
mL) was added, and the solution was cooled to -35 °C. Complex
1 was collected as yellow crystals via filtration (6.01 g, 7.23 mmol,
60% yield). Characterization data agreed with the literature.1
[NW(OCMe2CF3)3]3 (3). W2(OCMe2CF3)6 (10.0 mg, 0.00885
mmol) was dissolved in toluene-d8 (0.5 mL). Propionitrile (1.2 µL,
0.0173 mmol, 2.0 equiv) was introduced via syringe, and the
resulting reaction mixture was heated at 95 °C for 18.5 h. At this
1
point, H NMR and 19F NMR indicated complete conversion to
3.23
CH3CH2CtW(OC(CF3)2Me)3(DME) (5-DME). 1-DME (450
mg, 0.541 mmol) was dissolved in toluene (10 mL), and the solution
was transferred to a bomb flask. To this solution 3-hexyne (61.5
µL, 0.541 mmol, 1 equiv) was added via syringe. The bomb flask
1
was sealed and heated at 95 °C for 16 h. H NMR spectroscopy
indicated incomplete conversion to the alkylidyne. Additional
3-hexyne (20.0 µL, 0.176 mmol, 0.325 equiv) was syringed into
the reaction mixture. This was then heated at 95 °C for 2 h. The
resulting mixture was then filtered through Celite to remove poly-
3-hexyne. The volatiles were removed in vacuo from the filtrate.
The red residue was then taken up in 5 mL of pentane and cooled
to -35 °C. A red powder, 5, was collected via filtration (242.8
mg, 0.283 mmol, 52%). Characterization data agreed with the
literature.6
a Y: This catalyst form is deactivated.
400 spectrometer and were referenced to an external standard of
CFCl3 in CDCl3 (0.00 ppm). MALDI-TOF data were collected on
a Micromass TofSpec-2E matrix-assisted laser-desorption time-of-
flight mass 33spectrometer.
CH3CH2CtW(OCMe2CF3)3 (6). W2(OCMe2CF3)6 (570.4 mg,
0.505 mmol) was dissolved in toluene (10 mL). 3-Hexyne (68.8
µL, 0.605 mmol, 1.2 equiv) was added via syringe, and the reaction
mixture was stirred for 1 h. The volatiles were removed in vacuo,
and the resulting residue was taken up in 10 mL of pentane. The
pentane solution was filtered, and the filtrate was reduced in volume
to 2 mL. The resulting solution was cooled to -35 °C. A powder
of 6 was collected via filtration (528.2 mg, 0.871 mmol, 86%). 1H
NMR (300 MHz, tol-d8): δ 3.40 (q, 2H, tCCH2CH3, J ) 8.2 Hz),
1.31 (d, 2H, ArH, J ) 8.4 Hz), 1.65 (s, 18H, OC(CH3)2CF3). 19F
NMR (400 MHz, C6D6): -82.59(OC(CH3)2CF3). 13C{1H} NMR
(400 MHz, CD2Cl2): δ 274.18 (t, WtC, JW-C ) 148.1 Hz), 127.39
(q, OC(CH3)2CF3, JC-F ) 284.0 Hz), 82.29 (q, OCCF3(CH3)2, JC-F
) 28.8 Hz), 40.71 (s, CCH2CH3), 25.62 (s, OC(CH3)2CF3), 16.89
(s, CCH2CH3).30
4-MeO(C6H4)CtW(OC(CF3)2Me)3(DME) (7-DME). 1-DME
(200 mg, 0.241 mmol) and 1-(4-methoxyphenyl)-1-butyne (38.8
mg, 0.241 mmol, 1 equiv) were slurried in toluene-d8 (1 mL) in a
J. Young tube. This mixture was heated at 95 °C for 2.5 h. The
volatiles were removed in vacuo. The resulting residue was
reconstituted in toluene-d8 (1 mL) and heated at 95 °C for 2 h. The
volatiles were removed in vacuo. The remaining material was
dissolved in a total of 4 mL of 1:1 Et2O/pentane and cooled to
Materials and Methods. All reactions were performed in an
atmosphere of dinitrogen, either in a nitrogen-filled MBRAUN
Labmaster 130 glovebox or by using standard air-free techniques.48
All solvents used were dried and deoxygenated by the method of
Grubbs et al.49 [NtWCl3]4 ·1.1DCE,29 LiOC(CF3)2Me,50 1-(4-
methoxyphenyl)-1-butyne,11 W2(OCMe2CF3)6,30 [Li(dme)2][NW-
(OC(CF3)2Me)4] (2), [NW(OCMe2CF3)3]3 (3),1 1-(4-trifluorometh-
ylphenyl)-1-butyne,11 bis(4-trifluoromethylphenyl)acetylene,51 and
(44) Naddo, T.; Che, Y. K.; Zhang, W.; Balakrishnan, K.; Yang, X. M.;
Yen, M.; Zhao, J. C.; Moore, J. S.; Zang, L. J. Am. Chem. Soc. 2007,
129, 6978.
(45) Zhang, W.; Moore, J. S. J. Am. Chem. Soc. 2004, 126, 12796.
(46) Zhang, W.; Moore, J. S. J. Am. Chem. Soc. 2005, 127, 11863.
(47) Conrad, J. C.; Eelman, M. D.; Silva, J. A. D.; Monfette, S.; Parnas,
H. H.; Snelgrove, J. L.; Fogg, D. E. J. Am. Chem. Soc. 2007, 129,
1024.
(48) Shriver, D. G.; Sailor, M. J. The Manipulations of Air-SensitiVe
Compounds, 2nd ed.; Wiley-Interscience: New York, 1986.
(49) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.;
Timmers, F. J. Organometallics 1996, 15, 1518.
(50) delaMata, F. J.; Grubbs, R. H. Organometallics 1996, 15, 577.
(51) Shirakawa, E.; Kitabata, T.; Otsuka, H.; Tsuchimoto, T. Tetrahedron
2005, 61, 9878.
9
8996 J. AM. CHEM. SOC. VOL. 130, NO. 28, 2008

Investigation of Nitrile-Alkyne Cross-Metathesis
A R T I C L E S
-35 °C. The product (7), a deep red-orange powder, was collected
3H, OCH3), 1.64 (s, 18H, OC(CH3)2CF3). 19F NMR (400 MHz,
CD2Cl2): δ -83.17 (s, OC(CH3)2CF3). 13C{1H} NMR (400 MHz,
CD2Cl2): δ 264.99 (s, WtC), 158.34 (s, Ar), 138.94 (s, Ar), 132.83
(s, Ar), 125.94 (q, OC(CH3)2CF3, JC-F ) 284.4 Hz), 112.23 (s,
Ar), 81.83 (q, OCCF3(CH3)2, JC-F ) 28.8 Hz), 54.76 (s, OMe),
24.36 (s, OC(CH3)2CF3). Anal. calcd for WO4C20H25F9: C, 35.11;
H, 3.68. Found: C, 34.86; H, 3.43.
1
via filtration (124.2 mg, 0.133 mmol, 67%). H NMR (500 MHz,
C6D6): δ 6.81 (d, 2H, ArH, J ) 6.8 Hz), 6.71 (d, 2H, ArH, J ) 6.8
Hz), 3.65 (br s, 3H, DME), 3.25 (s, 3H, OCH3), 3.20 (br s, 4H,
DME), 2.92 (br s, 3H, DME), 1.88 (s, 9H, OC(CF3)2CH3). 1H NMR
(300 MHz, toluene-d8, -20 °C): δ 6.80 (d, 2H, ArH, J ) 9.0 Hz),
6.65 (d, 2H, ArH, J ) 9.0 Hz), 3.65 (s, 3H, DME), 3.21 (s, 3H,
OCH3), 3.10 (s, 3H, DME), 3.07 (t, 2H, DME, J ) 4.3 Hz), 2.86
(t, 2H, DME, J ) 4.3 Hz), 1.89 (s, 9H, OC(CF3)2CH3). 19F NMR
(400 MHz, CD2Cl2): δ -77.05 (s). 13C{1H} NMR (400 MHz,
CD2Cl2): δ 278.26 (s) (WtC), 160.05 (s) (Ar), 137.84 (s) (Ar),
135.47 (s) (Ar), 124.7 (q) (OC(CF3)2CH3, JC-F ) 289.00 Hz),
112.76 (s) (Ar), 83.11 (m) (OC(CF3)2CH3), 75.34 (s) (DME), 73.00
(s) (DME), 70.60 (s) (DME), 59.50 (s) (DME), 55.64 (s) (OCH3),
18.97 (s) (OC(CF3)2CH3). Anal. calcd for WO6C24H26F18: C, 30.79;
H, 2.80. Found: C, 30.97; H, 2.96.
4-F3C(C6H4)CtW(OCMe2CF3)3) (10). Complex 6 (200 mg,
0.330 mmol) was dissolved in toluene (5 mL). Then, bis(4-
trifluoromethylphenyl)acetylene (51.8 mg, 0.165 mmol, 2.0 equiv)
was dissolved in toluene (5 mL) and added to the solution of 6.
The reaction mixture was stirred for 1 h. The volatiles were removed
in vacuo. The resulting mixture was dissolved in pentane (10 mL)
and filtered. The volume of the filtrate was reduced to 3 mL, and
the solution was cooled to -35 °C. Light yellow feathers of 10
were collected via filtration and washed with 2 mL of cold pentane
(175 mg, 0.242 mmol, 73% yield). 1H NMR (400 MHz, CD2Cl2):
δ 7.62 (d, 2H, ArH, J ) 8.2 Hz), 7.13 (d, 2H, ArH, J ) 8.4 Hz),
1.65 (s, 18H, OC(CH3)2CF3). 19F NMR (400 MHz, CD2Cl2): δ
-83.18 (s, 9F, OC(CH3)2CF3), -62.99 (s, 3F, ArCF3). 13C{1H}
NMR (400 MHz, CD2Cl2): δ 265.11 (t, WtC, JC-W ) 152.0 Hz),
4-F3C(C6H4)CtW(OC(CF3)2Me)3(DME) (8-DME). 1-DME
(250 mg, 0.301 mmol) and 1-(4-trifluoromethylphenyl)-1-butyne
(59.9 mg, 0.301 mmol, 1 equiv) were slurried in toluene-d8 (1 mL)
in a J. Young tube. The reaction mixture was heated at 95 °C for
12 h. The volatiles were removed in vacuo. The resulting residue
was dissolved in 5 mL of pentane and cooled to -35 °C. The
product, a deep yellow powder, was collected via filtration (95.0
mg, 0.0975 mmol, 32%). A second crop of 71.2 mg was collected
148.95 (s, Ar), 133.06 (s, Ar), 127.16 (q, OC(CH3)2CF3, JC-F
268.2 Hz), 124.56 (q, Ar, JC-F ) 271.4 Hz), 125.32 (q, Ar, JC-F
)
)
4.0 Hz), 83.74 (q, OCCF3(CH3)2, JC-F ) 29.2 Hz), 25.71 (s,
OC(CH3)2CF3). Anal. calcd for WO3C20H22F12: C, 33.26; H, 3.07.
Found: C, 33.15; H, 3.13.
1
to give a total yield of 57% (0.171 mmol). H NMR (500 MHz,
C6D6): δ 7.33 (d, 2H, ArH, J ) 8.2 Hz), 6.70 (d, 2H, ArH, J ) 8.2
Hz), 3.57 (br s, 3H, DME), 3.78 (br s, 2H, DME), 3.44 (br s, 3H,
DME), 1.83 (s, 9H, OC(CF3)2CH3). 1H NMR (300 MHz, toluene-
d8, -20 °C): δ 7.25 (d, 2H, ArH, J ) 8.1 Hz), 6.68 (d, 2H, ArH,
J ) 8.1 Hz), 3.56 (s, 3H, DME), 3.07 (s, 3H, DME), 3.04 (t, 2H,
DME, J ) 4.3 Hz), 2.82 (t, 2H, DME, J ) 4.3 Hz), 1.89 (s, 9H,
OC(CF3)2CH3).19 F NMR (400 MHz, CD2Cl2): δ -63.21 (s, CF3)
-77.61 (s, OC(CF3)2CH3). 13C{1H} NMR (400 MHz, CD2Cl2): δ
276.41 (s, WtC), 146.44 (s, Ar), 134.13 (s, Ar), 129.13 (q, Ar,
JC-F ) 65.3 Hz), 124.50 (q, OC(CF3)2CH3, JC-F ) 289.0 Hz),
123.86 (q, Ar, JC-F ) 272.1 Hz), 124.66 (q, Ar, JC-F ) 4.4 Hz),
83.28 (m, OC(CF3)2CH3), 75.71 (s, DME), 73.55 (s, DME), 69.96
(s, DME), 59.94 (s, DME), 18.87 (s, OC(CF3)2CH3). Anal. calcd
for WO5C24H23F21: C, 29.59; H, 2.38. Found: C, 29.38; H, 2.19.
MeO(C6H4)CtW(OCMe2CF3)3 (9). Complex 6 (200 mg, 0.330
mmol) was dissolved in toluene (5 mL). To this solution, solid bis(4-
methoxyphenyl)acetylene (39.3 mg, 0.165 mmol) was added. The
solution was diluted with toluene (5 mL). The reaction mixture
was stirred for 1 h. Then the volatiles were removed in vacuo. The
resulting mixture was dissolved in pentane (15 mL) and filtered.
The volume of the filtrate was reduced to 6 mL, and the solution
was cooled to -35 °C. Deep yellow-orange crystals of 9 were
isolated via filtration and washed with 2 mL of cold pentane (188
mg, 0.274 mmol, 83% yield). 1H NMR (400 MHz, CD2Cl2): δ 7.00
(d, 2H, ArH, J ) 8.8 Hz), 6.87 (d, 2H, ArH, J ) 8.6 Hz), 3.79 (s,
OdW(OCMe2CF3)4 (11). Complex 3 (500 mg, 0.863 mmol)
was dissolved in toluene (10 mL) in a bomb flask. 3-Hexyne (294
µL, 2.59 mmol, 3 equiv) was added via syringe to the solution,
and the bomb flask was sealed and heated at 95 °C for 2 days. The
reaction mixture was filtered through Celite with Et2O (40 mL).
The filtrate was dried in vacuo and taken up in pentane (20 mL).
The mixture was again filtered, and the resulting filtrate was reduced
to 5 mL and cooled to -35 °C. Deep orange crystals of 11 were
collected via filtration and washed with 3 mL of cold pentane (111.6
mg, 0.158 mmol, 24% yield). 1H NMR (400 MHz, tol-d8): δ 1.43
(s, OC(CH3)2CF3). 19F NMR (400 MHz, tol-d8): δ -81.4 (s,
OC(CH3)2CF3). 13C{1H} NMR (400 MHz, C6D6): δ 127.67 (q,
OC(CF3)2CH3, JC-F ) 285.4 Hz), 84.60 (q, CAr, JC-F ) 29.9 Hz),
22.10 (s, OC(CF3)2CH3). Anal. calcd for WO5C16H24F12: C, 27.14;
H, 3.42. Found: C, 27.13; H, 3.32.
(Et3C3)W(OCMe2CF3)3 (12). Complex 6 (160.0 mg, 0.264
mmol) was dissolved in CD2Cl2 (500 µL). 3-Hexyne (33.0 µL, 0.290
mmol, 1.1 equiv) was introduced to the solution via syringe. The
resulting 1H NMR spectrum was observed at -60 °C. The complete
removal of volatiles resulted in conversion to 6. 1H NMR (300
MHz, tol-d8, -60 °C): δ 3.56 (q, 4H, CRCH2CH3, J ) 7.0 Hz),
2.62 (q, 2H, CꢀCH2CH3, J ) 7.0 Hz), 1.81 (s, 6H, OC(CH3)2CF3),
1.37 (t, 6H, CRCH2CH3, J ) 7.0 Hz), 1.65 (s, 18H, OC(CH3)2CF3),
0.95 (s, 18H, OC(CH3)2CF3), 0.76 (t, 3H, CꢀCH2CH3, J ) 7.0 Hz).
9
J. AM. CHEM. SOC. VOL. 130, NO. 28, 2008 8997

A R T I C L E S
Geyer et al.
1
19F NMR (300 MHz, tol-d8, -60 °C): -81.30 (s, 6F, OC(CH3)2CF3
ax), -81.55 (s, 3F, OC(CH3)2CF3 eq). 13C{1H} NMR (300 MHz,
CD2Cl2, -60 °C): δ 238.19 (t, W-CR, JW-C ) 61.7 Hz), 130.47
(s, W-Cꢀ), 127.92 (q, OC(CH3)2CF3 eq, JC-F ) 286.2 Hz), 27.60
(q, OC(CH3)2CF3 ax, JC-F ) 287.4 Hz), 81.25 (q, OCCF3(CH3)2
eq, JC-F ) 28.5 Hz), 76.34 (q, OCCF3(CH3)2 ax, JC-F ) 27.3 Hz),
29.97 (s, OC(CH3)2CF3 eq), 25.08 (s, CRCH2CH3), 25.62 (s,
OC(CH3)2CF3), 20.02 (s, CꢀCH2CH3), 14.26 (s, CH3), 12.30 (s,
CH3).
an off-white powder (161.1 mg, 0.104 mmol, 68.4%). H NMR
matched the literature values.45,52 MALDI-TOF [m/z]+
: 1542.1
(C112H140N4)
Crystal Structure Determinations. 1-DME. Colorless needles
of 1-DME were grown from a pentane/diethyl ether solution at
-35 °C. A crystal of dimensions 0.44 mm × 0.42 mm × 0.16
mm was mounted on a standard Bruker SMART 1K CCD-based
X-ray diffractometer equipped with a LT-2 low-temperature
device and normal focus Mo-target X-ray tube (λ ) 0.71073 A)
operated at 2000 W (50 kV, 40 mA). The X-ray intensities were
measured at 108(2) K; the detector was placed at a distance of
4.969 cm from the crystal. A total of 3000 frames was collected
with a scan width of 0.5° in ω and ꢁ with an exposure time of
20 s/frame. The integration of the data yielded a total of 111 329
reflections to a maximum 2θ value of 56.82°, of which 12 760
were independent and 12 059 were greater than 2σ(I). The final
cell constants (Table 1) were based on the xyz centroids of 8931
reflections above 10σ(I). Analysis of data showed negligible
decay during data collection; the data were processed with
SADABS and corrected for absorption. The compound crystal-
lized as a pseudo-merohedral twin with twin law (100, 0-10,
00-1) and twin scale factor 0.3732(8). The structure was solved
and refined with the Bruker SHELXTL (version 6.12) software
package, using the space group P2(1)/c with Z ) 8 for the
formula C16H19F18NO5W. All non-hydrogen atoms were refined
anisotropically with the hydrogen atoms placed in idealized
positions. Full matrix least-squares refinement based on F2
converged at R1 ) 0.0533 and wR2 ) 0.1384 [based on I >
2σ(I)] and R1 ) 0.0569 and wR2 ) 0.1411 for all data.
Complex 2. Colorless needles of 2 were grown from a toluene/
pentane solution at -35 °C. A crystal of dimensions 0.46 mm
× 0.26 mm × 0.22 mm was mounted on a Bruker SMART-1K
CCD-based X-ray diffractometer equipped with a low-temper-
ature device and fine focus Mo-target X-ray tube (λ ) 0.71073
A) operated at 1500 W (50 kV, 40 mA). The X-ray intensities
were measured at 108(2) K; the detector was placed at a distance
4.912 cm from the crystal. A total of 2342 frames were collected
with a scan width of 0.5° in ω and 0.45° in ꢁ with an exposure
time of 20 s/frame. The integration of the data yielded a total
of 298 848 reflections to a maximum 2θ value of 56.64° of which
37 755 were independent and 34 774 were greater than 2σ(I).
The final cell constants (Table 1) were based on the xyz centroids
of 9771 reflections above 10σ(I). Analysis of data showed
negligible decay during data collection; the data were processed
with SADABS and corrected for absorption. The structure was
solved and refined with the Bruker SHELXTL (version 6.12)
software package, using the space group P2(1)/c with Z ) 16
for the formula C24H32NLiO8F24W. There are four crystallo-
graphically independent complexes in the asymmetric unit. The
structure is a pseudo-orthorhombic twin with twin law (100,
-100, 00-1). The refined twin fraction is 0.3915(2). All non-
hydrogen atoms were refined anisotropically with the hydrogen
atoms placed in idealized positions. The dimethoxyethane ligands
are disordered on one complex. Full matrix least-squares
refinement based on F2 converged at R1 ) 0.0236 and wR2 )
0.0486 [based on I > 2σ(I)], and R1 ) 0.0292 and wR2 ) 0.0504
for all data.
Substrate Syntheses. 3,6-Dibromo-9-tetradecyl-9H-carbazole.
In a glovebox, a solution of THF/DMF (30 mL/10 mL) was placed
in a 250 mL three-necked flask fitted with a gas adapter, rubber
septum, and a solid addition funnel containing NaH (60% in mineral
oil, 0.48 g, 12.0 mmol, 1.6 equiv). Under N2 flow on a Schlenk
line, 3,6-dibromocarbazole (2.51 g, 7.72 mmol, 1.01 equiv) was
added to the flask, then 1-bromotetradecane (2.28 mL, 7.66 mmol)
was added via syringe. The resulting solution was cooled to 0 °C,
and then the NaH was added with stirring. The solution was allowed
to warm overnight with stirring, and then the volatiles were removed
in vacuo. The remaining mixture was quenched with 1 M HCl (50
mL). The product was extracted with CHCl3 (75 mL), and the
organic layer was washed with H2O (3 × 50 mL). The solvent
was removed, and the resulting powder was washed with cold
pentane (2 × 20 mL) and dried in vacuo, yielding white needles
(3.45 g, 6.62 mmol, 86.4%). 1H NMR (300 MHz, CD2Cl2): δ 8.16
3
(s, 2H, ArH), 7.57 (d, 2H, ArH, JH-H ) 8.7 Hz), 7.33 (d, 2H,
3
3
ArH, JH-H ) 8.7 Hz), 4.26 (t, 2H, NCH2, JH-H ) 7.5 Hz), 1.83
(m, 2H, NCH2CH2), 1.23-1.53 (m, 22H, (CH2)11), 0.88 (t, 3H,
CH3, 3JH-H ) 6.6 Hz). 13C{1H} NMR (CDCl3): δ 139.35, 129.07,
123.50, 123.31, 112.02, 110.46, 43.41, 32.08, 29.83, 29.80, 29.78,
29.73, 29.68, 29.60, 29.52, 29.47, 28.96, 27.34, 22.85, 14.29. Anal.
calcd for C26H35NBr2: C, 59.90; H, 6.77; N, 2.69. Found: C, 60.06;
H, 6.45; N, 2.65.
9-Tetradecyl-9H-carbazole-3,6-dicarbonitrile. In a glovebox,
dry DMF (30 mL) was measured into a 100 mL two-necked flask
equipped with a rubber septum and gas adapter. Under N2 flow on
a Schlenk line, 3,6-dibromo-9-tetradecyl-9H-carbazole (3.17 g, 6.08
mmol) and copper(I) cyanide (2.00 g, 22.3 mmol) were added to
the flask. A reflux condensor was fitted under N2 flow, and the
solution was heated to reflux with stirring for 26 h. The solution
was cooled to room temperature, and then concentrated NH4OH
(20 mL) was added. After stirring for 10 min, the mixture was
extracted with CH2Cl2 (2 × 50 mL). The combined organics were
washed with H2O (2 × 50 mL) and brine (50 mL), and then filtered
through a plug of silica to remove solid impurities. The silica plug
was washed with CH2Cl2 until the filtrate ran colorless. The
combined filtrate was concentrated to dryness. The resulting powder
was washed with acetone, then dried in vacuo, yielding a brown
powder (1.35 g, 3.26 mmol, 53.6%). 1H NMR (300 MHz, CDCl3):
3
δ 8.41 (s, 2H, ArH), 7.79 (d, 2H, ArH, JH-H ) 8.4 Hz), 7.52 (d,
3
3
2H, ArH, JH-H ) 8.4 Hz), 4.35 (t, 2H, NCH2, JH-H ) 6.9 Hz),
1.83-1.90 (m, 2H, -NCH2CH2), 1.22-1.33 (m, 22H, (CH2)11), 0.87
(t, 3H, CH3, JH-H ) 6.6 Hz). 13C{1H} NMR (CDCl3): δ 142.84,
3
130.40, 125.81, 122.24, 119.92, 110.33, 103.57, 43.90, 32.04, 29.78,
29.76, 29.74, 29.69, 29.60, 29.54, 29.48, 29.39, 28.98, 27.31, 22.81,
14.25. Anal. calcd for C28H35N3: C, 81.31; H, 8.53; N, 10.16. Found:
C, 80.87; H, 8.41; N, 10.01.
Tetramer (12). In a glovebox, 1-DME (93.4 mg, 0.112 mmol)
and 9-tetradecyl-9H-carbazole-3,6-dicarbonitrile (250.5 mg, 0.606
mmol) were weighed into a bomb flask. Bromobenzene (37 mL)
and 3-hexyne (0.14 mL, 1.23 mmol) were added, and then the bomb
flask was sealed and placed in a 95 °C oil bath. After 20 h, the
solution was cooled and a vacuum was pulled for 2 h to remove
volatile products. The bomb flask was sealed and heated for an
additional 13 h. The solution was cooled and washed through a
plug of silica with CHCl3 (150 mL). The filtrate was concentrated
to dryness, then precipitated from CHCl3/pentane at -20 °C. The
mixture was filtered, and the precipitate was dried in vacuo to give
Acknowledgment. This material is based upon work supported
by the National Science Foundation under Grant CHE-0449459.
A.M.G. thanks the University of Michigan Chemistry Department
for a Robert W. Parry Inorganic Chemistry Graduate Fellowship.
J.B.G. thanks the National Science Foundation for a Graduate
Research Fellowship.
(52) Cho, H. M.; Weissman, H.; Wilson, S. R.; Moore, J. S. J. Am. Chem.
Soc. 2006, 128, 14742.
9
8998 J. AM. CHEM. SOC. VOL. 130, NO. 28, 2008

Investigation of Nitrile-Alkyne Cross-Metathesis
A R T I C L E S
Supporting Information Available: Full crystallographic
details for 1-DME and 2. Experimental data for reversibility of
NACM and ACM, direct monitoring of catalytic reaction
progress, optimization studies, influence of AP studies, catalyst
comparison studies, functional group tolerance. Coordinates for
all atoms, bond lengths and angles of optimized structures in
DFT calculations. This material is available free of charge via
the Internet at http://pubs.acs.org.
JA800020W
9
J. AM. CHEM. SOC. VOL. 130, NO. 28, 2008 8999