Combined experimental and computational investigations of rhodium- and ruthenium-catalyzed C-H functionalization of pyrazoles with alkynes

Article abstract of DOI:10.1021/jo402592z

Detailed experimental and computational studies are reported on the mechanism of the coupling of alkynes with 3-arylpyrazoles at [Rh(MeCN) ₃Cp*][PF₆]₂ and [RuCl₂(p- cymene)]₂ catalysts. Density functi

Full text of DOI:10.1021/jo402592z

Article
pubs.acs.org/joc
Combined Experimental and Computational Investigations of
Rhodium- and Ruthenium-Catalyzed C−H Functionalization of
Pyrazoles with Alkynes
Andres G. Algarra,† Warren B. Cross,*,‡,§ David L. Davies,*,‡ Qudsia Khamker,‡ Stuart A. Macgregor,*,†
́
Claire L. McMullin,† and Kuldip Singh‡
†Institute of Chemical Sciences, Heriot-Watt University, Edinburgh EH14 4AS, U.K.
‡Department of Chemistry, University of Leicester, University Road, Leicester LE1 7RH, U.K.
S
* Supporting Information
ABSTRACT: Detailed experimental and computational studies are reported on the
mechanism of the coupling of alkynes with 3-arylpyrazoles at [Rh(MeCN)3Cp*][PF6]2
and [RuCl2(p-cymene)]2 catalysts. Density functional theory (DFT) calculations
indicate a mechanism involving sequential N−H and C−H bond activation, HOAc/
alkyne exchange, migratory insertion, and C−N reductive coupling. For rhodium, C−H
bond activation is a two-step process comprising κ2−κ1 displacement of acetate to give
an agostic intermediate which then undergoes C−H bond cleavage via proton transfer
to acetate. For the reaction of 3-phenyl-5-methylpyrazole with 4-octyne kH/kD = 2.7
0.5 indicating that C−H bond cleavage is rate limiting in this case. However, H/D
exchange studies, both with and without added alkyne, suggest that the migratory
insertion transition state is close in energy to that for C−H bond cleavage. In order to
model this result correctly, the DFT calculations must employ the full experimental
system and include a treatment of dispersion effects. A significantly higher overall barrier
to catalysis is computed at {Ru(p-cymene)} for which the rate-limiting process remains
C−H activation. However, this is now a one-step process corresponding to the κ2−κ1 displacement of acetate and so is still
consistent with the lack of a significant experimental isotope effect (kH/kD = 1.1 0.2).
It has been noted for rhodium catalysts that neutral directing
groups undergo 1:2 coupling with alkynes to give naphthalenes;
in contrast, substrates that contain a protic E−H (E = N, O)
directing group can undergo deprotonation (E−H activation)
and typically undergo 1:1 coupling with alkynes and give
heterocyclic products.6 With benzamides, both possibilities
have been observed depending on the substitution pattern (see
Scheme 1). We have become interested in directing groups that
can have an NH group through tautomerisation and can then
potentially react by either pathway.
A general catalytic cycle proposed for heterocycle formation
is shown in Scheme 2. Opening of the dimer precursor in the
presence of Cu(OAc)2 gives a [M(OAc)2(ring)] intermediate
(A), which can effect N−H and C−H bond activation of the
INTRODUCTION
■
Metal-catalyzed C−H functionalization has become a focus for
research for more atom-efficient methods for construction of
C−C and C−E bonds (E = O, N).1 These methods avoid the
requirement to prefunctionalize the C−H bond, usually to a
C−X bond (X = Cl, Br, I), and avoid the formation of salt
waste in the subsequent C−C or C−E bond formation reaction.
In particular, the recognition by us2 and others3 of the role of
carboxylates in providing facile C−H activation has led to a
huge upsurge in interest in this field. We have characterized this
as ambiphilic metal−ligand assisted (AMLA)2d,e C−H
activation, reflecting the synergic role of both metal center
and intramolecular carboxylate base in promoting this process,
and this is closely related to the concerted metalation−
deprotonation (CMD) concept proposed by Fagnou and co-
workers.3f This increased understanding of the mechanism of
this C−H activation process has enhanced its use in catalysis.
The use of Rh(III) catalysts based on [RhCl2Cp*]2 and
related derivatives for C−H functionalization was pioneered by
the groups of Miura4 and Fagnou.5 However there has been a
huge increase in interest in the past few years and the field has
recently been reviewed.6 This methodology has allowed access
to substituted naphthalenes and to a wide range of heterocycles.
More recently, related chemistry has been demonstrated using
[RuCl2(p-cymene)]2 as the catalyst precursor.1f,7
Scheme 1
Received: November 22, 2013
Published: February 24, 2014
© 2014 American Chemical Society
1954
dx.doi.org/10.1021/jo402592z | J. Org. Chem. 2014, 79, 1954−1970

The Journal of Organic Chemistry
Article
whether C−H activation is reversible. In addition, while kinetic
isotope effects (kH/kD) can in principle give information on
whether C−H activation is the rate-limiting step in catalysis,
Hartwig11 has recently highlighted the difficulties in correctly
interpreting such data.
Scheme 2
One additional means to gain mechanistic insight is the use
of density functional theory (DFT) calculations to complement
experimental observations. To date, only two DFT studies on
heterocycle formation via {RhCp*}-catalyzed C−H function-
alization have been published, both of which focused on
benzamide derivatives, ArC(O)NH(OR), in which the N-
bound OR group acts as an internal oxidant.12 In their
computational study, Guimond et al. considered isoquinolone
formation via the reaction of PhC(O)NH(OAc) with acetylene
at a simple Rh(OAc)2Cp model catalyst.13 More recently Xia
and co-workers revisited this system for dihydroquinolone
formation via coupling of PhC(O)NH(OR) (R = Me,
C(O)tBu) and ethene at Rh(OAc)2Cp*.14 Both studies
propose that the highest lying transition state in the overall
process corresponds to the C−H activation step, and Guimond
et al. interpreted this as being consistent with the significant
kH/kD kinetic isotope effect (15 1) observed experimentally
for the [RhCl2Cp*]2-catalyzed reaction of C6H5/D5-C(O)NH-
(O2CtBu) with MeCCPh. In contrast, the closely related
PhC(O)NH(OMe) substrate gave no kH/kD kinetic isotope
effect, suggesting that the nature of the rate-limiting transition
state changes with the identity of the internal oxidant. No
calculations on the PhC(O)NH(OMe) process were reported.
Here we report synthetic and mechanistic studies on the
highly efficient C−H functionalization of 3-phenylpyrazoles
with alkynes at a {RhCp*} center and show similar reactivity at
{Ru(p-cymene)}. In contrast to N-phenylpyrazole which gives
naphthalenes,15 the presence of an NH proton which can be
deprotonated leads to formation of heterocyclic pyrazoloiso-
quinoline products. In particular, for rhodium catalysis we have
extended the range of alkynes as coupling partners and shown
that [Rh(MeCN)3Cp*][PF6]2 is a more efficient catalyst
precursor than [RhCl2Cp*]2. In addition, we demonstrate
that for both rhodium and ruthenium the reactions can be run
using catalytic rather than stoichiometric copper as the
reoxidant. Deuteration and competition experiments, coupled
substrate to form a 5-membered metallacycle (C). After
HOAc/alkyne substitution, migratory insertion gives a 7-
membered metallacycle (E) from which reductive elimination
with C−E bond formation gives the heterocyclic product. The
Rh(I) or Ru(0) species formed in this step is reoxidized in the
presence of Cu(OAc)2 to regenerate the [M(OAc)2(ring)]
active species.
Direct evidence for some of the steps in Scheme 2 has been
obtained from the observation of stoichiometric C−H
activation at complexes based on {RhCp*},8 as well as in
some cases for the subsequent alkyne insertion.8a,e,9 Such
studies have usually employed neutral directing groups,
although the first examples of intermediates formed from
anionic directing groups have recently also been isolated.10
However, mechanistic information on the catalytic reactions is
hampered by the difficulty in isolating intermediates, meaning
that information on the relative accessability of the different
steps is difficult to establish. Deuteration studies can show
Table 1. Catalyst Optimization
a
catalyst
[RhCl2Cp*]2
[Rh(MeCN)3Cp*][PF6]2
[Rh(MeCN)3Cp*][PF6]2
[Rh(MeCN)3Cp*][PF6]2
[Rh(MeCN)3Cp*][PF6]2
[RuCl2(p-cymene)]2
mol %
Cu(OAc)2 (equiv)
yield of 3aa (%)
1
2
3
4
5
6
7
8
9
5
5
0
5
1
5
5
5
5
2.5
2.5
2.5
0
9
98
0
b
,c
5
8
b
2.5
2.5
0.1
0.1
0.1
d
65
[Rh(MeCN)3Cp*][PF6]2
[RuCl2(p-cymene)]2
50 + 24% of 4aa
48
70
e
[RuCl2(p-cymene)]2
a
b
c
Isolated yield after chromatography. NMR yield after 17 h, using internal standard. Yield increased to >95% after addition of 2.5 equiv of
d
e
Cu(OAc)2. In t-AmOH 120 °C. 20 mol % of AgPF6 additive.
1955
dx.doi.org/10.1021/jo402592z | J. Org. Chem. 2014, 79, 1954−1970

The Journal of Organic Chemistry
Article
a
Scheme 3. Scope of C−H Functionalization of Pyrazoles with Alkynes
a
Standard conditions: 2.5 equiv of Cu(OAc)2·H2O with 5 mol % of catalyst, either [Rh(MeCN)3Cp*][PF6]2 in DCE at 83 °C or [RuCl2(p-
cymene)]2 in t-AmOH at 120 °C.
with detailed DFT calculations, have provided a detailed
understanding of the mechanism of the catalysis and suggest
that while C−H activation is rate limiting, the subsequent
reaction with alkynes can be competitive with this process. The
computational results stress the importance of employing an
appropriate model in the calculations as well as including a
treatment of dispersion effects. In addition, the calculations
show that even if C−H activation is involved in the rate
determining process, it does not follow that a kH/kD KIE will
necessarily be observed. During the course of our work some
related chemistry using [RhCl2Cp*]2 as catalyst was reported
by Li16 for arylalkynes and similar reactions catalyzed by
[RuCl2(p-cymene)]2 were subsequently reported by Acker-
mann.17
(entry 4) we observed a 5% yield of 3aa, i.e., stoichiometric
based on the amount of [Rh(MeCN)3Cp*][PF6]2 catalyst,
indicating that reductive elimination can occur without prior
oxidation. Lowering the catalyst loading to 1 mol % led to very
low conversions. For comparison, all reactions were left
overnight to go to completion; however, detailed reaction
monitoring showed that when [Rh(MeCN)3Cp*][PF6]2 was
used as catalyst the reaction was complete in 1 h at 83 °C. We
also tested [RuCl2(p-cymene)]2 as a catalyst precursor, and this
also gave 3aa in good (65%) yield, athough under somewhat
harsher conditions than rhodium. Moreover, unlike related
reactions recently reported by Ackermann,17 for 3aa we found
no need to use Ag+ salts as additives. Finally, we tested both
catalysts with only catalytic copper, using air as the external
oxidant. Using rhodium (entry 7), the yield of 3aa was reduced
(50%) and importantly, a new isomeric product 4aa was also
formed (24%). Simlarly with ruthenium (entry 8) the yield with
catalytic copper was slightly reduced (48%). However, addition
of 20 mol % of AgPF6 to give a cationic catalyst resulted in a
70% yield even with catalytic copper, and no 4aa was observed
(entry 9). Hence, we have demonstrated that both these
catalytic systems can work with only catalytic copper using air
as the external oxidant thus improving the overall efficiency of
the process.
RESULTS AND DISCUSSION
■
Catalytic Studies. Initially, we examined the reaction of 3-
phenyl-5-methylpyrazole (1a) with 4-octyne using different
catalysts and reaction conditions (Table 1). It is clear that
[Rh(MeCN)3Cp*][PF6]2 (entry 2) is a much better catalyst
than [RhCl2Cp*]2 (entry 1), forming heterocycle 3aa in
quantitative yield. We confirmed that there was no catalysis in
the absence of rhodium (entry 3). In the absence of Cu(OAc)2
1956
dx.doi.org/10.1021/jo402592z | J. Org. Chem. 2014, 79, 1954−1970

The Journal of Organic Chemistry
Article
Figure 1. ORTEP plots of 3ac (left) and 3dc (right). Thermal ellipsoids are drawn at 50% probability and hydrogen atoms are omitted for clarity.
Since in the case of rhodium using catalytic copper led to
some isomeric products, we explored the effects of using
stoichiometric copper on the reactivity scope for rhodium and
selected examples for ruthenium catalysis (see Scheme 3).
Changing the substituent on the pyrazole had little effect, with
good yields of 3ba, 3ca, and 3da being formed in reactions with
4-octyne. Next, we tried diphenylacetylene, which gave
products 3ab−db in high yields. During the course of this
work, Li et al. reported formation of 3ab−cb in broadly similar
yields.16 However, they noted reduced yields with electron-
withdrawing CF3 on the pyrazole and difficulty in activating
heterocyclic C−H bonds. In contrast, using [Rh-
(MeCN)3Cp*][PF6]2 as catalyst with a CF3-substituted
pyrazole we obtained 3db in 90% yield. Indeed, even substrates
which require C−H activation of a thiophene work well with a
CF3-substituted pyrazole, giving 5da and 5db, in 79 and 88%
yields, respectively. The results with ruthenium were less
consistent. Thus, though the product with 4-octyne, 3aa, was
formed in good yield (65% Table 1 entry 6) the corresponding
product from diphenylacetylene, 3ab, was only formed in 14%
yield even in the presence of AgPF6 as additive, while the
thiophene substrate with a CF3-substituted pyrazole with 4-
octyne gave 5da in 40% yield; in this case, addition of AgPF6
was essential.
As both dialkyl- and diarylalkynes react to give heterocyclic
products, we were interested in investigating the regioselectivity
with a mixed aryl/alkylalkyne. Thus, PhCCMe (2c) reacted
with 1a to give 3ac as a 9:1 ratio of isomers in 50% combined
yield with 40% yield of isomer 6ac (see below). The isomers of
3ac could be separated by chromatography, and the major
isomer is that shown, with the Ph substituent next to the N
atom in the heterocycle. Interestingly, this reaction with a Ru
catalyst gave 80% yield with a slightly higher selectivity of 12:1.
This implies a preference for insertion with an aryl next to the
metal and is consistent with previous studies of catalytic
heterocycle synthesis from alkynes with RhCp* and Ru(p-
cymene) catalysts.18 Unsubstituted pyrazole 1b reacted
similarly to give 3bc in 54% combined yield19 as a 12:1 ratio
of isomers. The phenyl- and trifluoromethyl-substituted
phenylpyrazoles reacted similarly to form 3cc and 3dc,
respectively, in good yields.
Similar reactions also occur with more functionalized alkyl
groups. Thus, R1CCCH2CH2OH (2d R1 = p-C6H4NO2)
reacted with 1a and 1c to give 3ad and 3cd in high yields. In
both cases, two regioisomers were observed with the dominant
product being that with the aryl substituent adjacent to N.
However, there is significantly more of the minor isomer in
these cases than for PhCCMe, the isomer ratio being 3:1 for
3ad and 1:1 for 3cd (compared to 9:1 and 3:1 for 3ac and 3cc,
respectively). It is possible in these cases that in the minor
isomer the hydroxyethyl side chain may be able to hydrogen
bond to the uncoordinated nitrogen of the pyrazole, thereby
leading to a relative stabilization of this isomer or an
intermediate that leads to this isomer.
Several of the products have been characterized by X-ray
crystallography. The structures of the major isomers of 3ac and
3dc are shown in Figure 1 and those of 3da, 5da, and 3db are
in the Supporting Information. The three fused rings are
coplanar but are not completely delocalized. The central ring
shows clear evidence of bond localization, as expected from
Clar’s rule,20 with the original alkyne bond C(10)−C(11) now
being a double bond (CC = 1.36−1.38 Å) and the bond
joining the phenyl or thiophene to the pyrazole being 1.43−
1.44 Å. The structures of the major isomers of 3ac and 3dc
confirm that the Ph substituent is on the carbon attached to
nitrogen.
In the reaction of 1a with 2a in the presence of catalytic
copper it is noticeable that a second isomeric product 4aa is
formed in addition to the normal heterocyclic product 3aa.
Similarly in some of the reactions with alkyne 2c isomeric
products 6 are also observed. Products arising by formal
hydrogen shifts have been noticed in rhodium oxidative
couplings previously,21 and equivalent products have been
realized through the use of allene coupling partners.22 We have
tested both products and find that they do not interconvert in
either direction when exposed to the catalyst and oxidant under
catalytic conditions. In principle, compounds 6 could arise from
insertion of an allene that is itself formed by isomerization of
PhCCMe; however our DFT calculations indicate that this
route is not energetically feasible as the barrier to C(sp3)−N
reductive coupling step is prohibitively high in energy. A
1957
dx.doi.org/10.1021/jo402592z | J. Org. Chem. 2014, 79, 1954−1970

The Journal of Organic Chemistry
Article
Scheme 4
possible alternative mechanism is discussed in the computation
section below.
activation is reversible but substrate exchange is not facile then
only the bound substrate can undergo exchange (i.e., 5%,
equivalent to the catalyst loading).
In all these reactions, only heterocyclic products are formed,
and there is no trace of naphthalene-type products which were
formed from N-phenylpyrazole.15 This outcome is therefore
consistent with that shown in Scheme 2, i.e., N−H and C−H
activation followed by alkyne insertion, C−N reductive
coupling, and reoxidation of the catalyst.23
The deuteration experiment was repeated in the presence of
a slight excess of alkyne and Cu(OAc)2, i.e., similar to the
catalytic conditions, although using methanol-d4 as solvent in
place of dichloroethane (see Scheme 4, b). In this case at only
low conversion (<10%) the product showed no deuterium
incorporation by NMR spectroscopy suggesting that the
cyclometalated intermediate is being intercepted by the alkyne
rather than going back to starting material. However, if the
reaction is allowed to go to completion there is deuterium
incorporation in both positions though considerably less at the
ortho position (7%) compared to the pyrazole position (66%).
Given that the alkyne is in slight excess (1.2 equiv), at high
conversion the alkyne concentration is approximately 0.2 times
that at the start; hence, at this stage the rate of H/D exchange
in the ortho position (i.e., the reverse of C−H activation) is
competitive with alkyne insertion. These observations suggest
that the barriers to the forward (alkyne insertion) and reverse
reactions (C−H activation) are similar in energy.
Similar experiments were carried out with ruthenium. In this
case, in the absence of alkyne there was significantly less
deuterium incorporation, suggesting that C−H activation is less
reversible for ruthenium. As for rhodium, exchange at the
pyrazole position occurred as well as at the ortho positions of
the phenyl. However, in the presence of substrate ruthenium
gave a different result; even at high conversion there was only
deuterium incorporation into the pyrazole C4 position and no
H/D exchange in the phenyl ring.24 This suggests that in this
case alkyne insertion is significantly easier than the reverse of
C−H activation (see below). Note, the observation that
deuterium exchange in the pyrazole position is less affected
by alkyne than exchange in the phenyl positions is also
consistent with pyrazole exchange not occurring via a rollover
mechanism.
Experimental Mechanistic Studies. A series of experi-
ments was conducted to shed light on the mechanism of
catalysis at both Rh and Ru metal centers. With Rh, attempted
stoichiometric cyclometalation of 1a led to broad NMR spectra
suggesting the presence of exchange processes. The reversibility
of the C−H activation step was therefore assessed by a
deuteration study. Treating substrate 1a with [Rh-
(MeCN)3Cp*][PF6]2 (5 mol %) and 2 equiv of NaOAc in
methanol-d4 gave approximately 29% deuterium incorporation
into each of the ortho positions after 16 h at 60 °C (see Scheme
4, a). Surprisingly, almost 50% deuterium exchange also
occurred at the remaining pyrazole proton. We did consider
whether this exchange may occur via a rollover cyclometalation,
involving reprotonation at N and N-decoordination from
cyclometalated intermediate C, rotation about the C1−Ph
bond, and reversible C−H activation at the C4 pyrazole
position. However, BP86-D3 calculations indicate that the
double-cyclometalated species implicated in this pathway would
be too high in energy (GDCE = 22.0 kcal/mol) and so ruled out
this mechanism (see the Supporting Information for details).
Alternatively, it is possible that when the pyrazole is
coordinated the C4 proton becomes more acidic and so can
undergo exchange without direct involvement of the rhodium.
Adding Cu(OAc)2 had little effect, similar deuteration being
observed. As has been observed previously22 addition of d1-
pivalic acid increased the rate and extent of deuterium
exchange. It is worth noting that to see deuterium
incorporation into the substrate requires not only that N−H
and C−H activation are reversible but also that exchange of free
and bound pyrazole is facile under these conditions. If C−H
A value for kH/kD for rhodium was measured using 1a-d3 and
was determined to be 2.7 0.5 (see Scheme 5).25 This suggests
1958
dx.doi.org/10.1021/jo402592z | J. Org. Chem. 2014, 79, 1954−1970

The Journal of Organic Chemistry
Article
Scheme 5. Parallel KIE Experiments
Figure 2. Mechanism for coupling of alkynes with 3-phenylpyrazoles at Rh(OAc)2(C5R′5) catalysts computed with model 1 (R = R′ = R″ = H) and
n
model 2 (R = R′ = Me, R″ = Pr).
that C−H bond cleavage is likely involved in the rate-
determining step. To confirm this, the order of reaction with
respect to alkyne was measured using initial rates, and this
showed near zero order in alkyne (0.07 0.1), again consistent
with C−H bond cleavage being rate limiting. Similar
deuteration and KIE studies were carried out for ruthenium.
The kH/kD was determined to be 1.1 0.2, and this could be
interpreted as indicating that C−H bond activation is not rate
determining; however, this reaction is also zero order in alkyne
mechanism of the coupling of alkynes with 3-phenylpyrazoles
at catalysts based on {RhCp*} and {Ru(p-cymene)} fragments.
For the Rh system, two models have been employed: model 1,
which combines simplified substrates and catalyst, i.e., 3-
phenylpyrazole with HCCH and Rh(OAc)2Cp, and model 2,
which uses the actual substrates and catalyst used exper-
imentally, i.e., 3-phenyl-5-methylpyrazole (1a), reacting with
nPrCCnPr (2a) at Rh(OAc)2Cp*. All geometries have been
optimized with the BP86 functional, and we report GDCE, i.e.,
gas-phase free energies corrected for the effects of DCE solvent
through the PCM approach. In a further step, a correction for
dispersion effects (using Grimme’s D3 parameter set) was
added to the BP86 results, and a number of alternative density
functionals were also tested. The calculations indicate that the
same basic catalytic cycle is in operation for both models 1 and
(0.0
0.1). Our DFT calculations show that the rate-
determining step in fact corresponds to the κ2-κ1 displacement
of acetate that must occur as the first step of the C−H
activation process prior to actual C−H bond cleavage (see
below).
Computational Mechanistic Studies. Density functional
theory calculations have been performed to assess the
1959
dx.doi.org/10.1021/jo402592z | J. Org. Chem. 2014, 79, 1954−1970

The Journal of Organic Chemistry
Article
Figure 3. Computed reaction profiles (GDCE, kcal/mol) for the coupling of nPrCCnPr with 1a at Rh(OAc)2Cp* (model 2): (a) BP86 functional; (b)
BP86-D3.
2, but that the reaction energetics vary considerably with the
model and functional adopted. In particular, we show how the
use of an oversimplified model such as model 1 is inappropriate
if a detailed modeling of the catalysis is to be achieved. In the
following we therefore first describe the overall mechanism and
then discuss the computed energetics in the light of the
available experimental data. Computed geometries are generally
standard and are supplied in full in the Supporting Information,
with only selected key structures being highlighted for more
detailed discussion.
The catalytic cycle computed for both models 1 and 2 starts
with adduct A, in which the pyrazole substrate 1 is H-bonded to
the pendant oxygen of the κ1-acetate ligand in Rh-
(OAc)2(C5R′5) (see Figure 2, which also gives the atom
labeling scheme employed). For model 2, this H-bonded
adduct is computed to be 2.8 kcal/mol more stable than the
separated reactants. N−H activation in A proceeds in two steps,
first, formation of the Rh−N2 bond (with κ2- κ1 displacement of
the second acetate ligand) to give INT(A-B), followed by N1−
H1 bond cleavage induced by acetate dissociation via elongation
of the Rh−O2 bond. This forms B·HOAc in which acetic acid is
H-bonded to N1 and the spectator acetate has reverted to a κ2-
binding mode. Loss of HOAc then gives B, which can undergo
C−H activation via an agostic intermediate, INT(B−C),
formed via κ2−κ1 displacement of acetate by the approaching
ortho-C−H bond of the substrate. This allows for intra-
molecular H−bonding to acetate, which promotes C−H bond
cleavage via TS(B−C)2. Together, these two steps comprise an
AMLA/CMD C−H bond activation.2d,3f The cyclometalated
intermediate C1 is initially formed from which a more stable
form, C2, can be readily accessed via rotation about the Rh−O4
bond. HOAc/alkyne exchange then forms D, which undergoes
migratory insertion (D→E) followed by C−N reductive
coupling to give F in which the pyrazoloisoquinoline product,
3, is η4-bound to the Rh center. Experimentally, this last step
has been shown to occur in the absence of added Cu(OAc)2
1960
dx.doi.org/10.1021/jo402592z | J. Org. Chem. 2014, 79, 1954−1970

The Journal of Organic Chemistry
Article
Figure 4. Computed structures of intermediate E for (a) model 1 and (b) model 2. Selected distances are given in angstroms and relative free
energies in solution are in kcal/mol and are based on the BP86 functional. H atoms are omitted for clarity.
(see above) and so rules out any need for prior oxidation of the
Rh center. The catalytic cycle is completed via release of 3 and
reoxidation of an unspecified Rh(I) species, G, although we
have not attempted to model these steps here.
kcal/mol cf. −18.0 kcal/mol). This reflects the different
geometries of E computed with the two models (see Figure
4). In model 1, insertion of HCCH gives a near-planar 7-
membered metallacycle that maximizes π-delocalization around
the ring.26 With 4-octyne the propyl substituents disfavor a
planar arrangement, and instead, a boat-like conformation is
seen. This also accounts for the lower barrier to C−N reductive
coupling with model 2 (ΔG⧧DCE = 10.0 kcal/mol cf. 16.5 kcal/
mol for model 1) as less distortion is required to access the
boat-like conformation necessary for C−N bond formation in
the transition state. The overall formation of F is also far less
favorable with model 2 (ΔGDCE = −11.1 kcal/mol), and a
significant component of this arises from the reduced intrinsic
stability of the pyrazoloisoquinoline product, which is estimated
to be ca. 15 kcal/mol less stable than with model 1.27
While model 2 predicts reversible and endergonic N−H and
C−H activation, the rate-determining transition state is now
clearly identified with the alkyne insertion step. This is at odds
with the reduction in the extent of H/D exchange at the 3-
phenylpyrazole ortho positions in the presence of alkyne, as this
should not be affected if a significantly larger barrier to alkyne
insertion exists. Moreover, rate-limiting alkyne insertion is
inconsistent with the near-zero order dependency on 4-octyne
concentration and the kH/kD value of 2.7 0.5. To investigate
these issues, we considered a range of density functionals with a
particular focus on the computed difference in energy between
TS(D−E) and the highest point along the N−H and C−H
activation steps. The key outcome is that the transition states
for these processes only become close in energy when a
treatment of dispersion effects is included, either directly in the
functional (B97D, M06) or in the protocol (BP86-D3, i.e.,
adding a dispersion correction to the BP86-optimized results).
For ease of comparison we focus on the BP86-D3 results (see
Figure 3b). A number of papers have highlighted the
importance of dispersion corrections when computing ligand
dissociation or ligand-exchange energies at transition-metal
complexes.28 Given this, as expected, the effects of the
dispersion correction are most significant in processes involving
Our initial results computed with model 1 show similar
features to those computed by Guimond and co-workers for the
reaction of PhC(O)NH(OAc) with HCCH at Rh(OAc)2Cp
(see Figure C1 in the Supporting Information).13 In the present
case, N−H and C−H activation proceed with an overall barrier
of 15.7 kcal/mol (corresponding to TS(B−C)1) to give
cyclometalated C2 at −2.4 kcal/mol. HOAc/HCCH exchange
in C2 then gives D at −1.0 kcal/mol from which migratory
insertion proceeds via TS(D-E) at +9.8 kcal/mol to form E at
−19.0 kcal/mol. C−N reductive coupling has a barrier of 16.5
kcal/mol and gives F at −35.0 kcal/mol. Comparing these
computed results with experiment highlights some key
discrepancies: (i) the formation of C2 is computed to be
exergonic, suggesting that a cyclometalated species should be
observed in the absence of alkyne; however, this is not the case
in the true experimental system; (ii) once C2 is formed HOAc/
alkyne substitution and migratory insertion is predicted to be
far easier (ΔG⧧
= +12.2 kcal/mol) than the reverse C−H
DCE
bond formation (ΔG⧧
= +18.1 kcal/mol); this is
DCE
inconsistent with the persistence of H/D exchange in the
presence of added alkyne.
These discrepancies associated with the small model led us to
consider the full experimental system computationally (model
2, Figure 3a). The initial N−H and C−H activation steps are
similar to those of model 1 and again have a small overall
barrier (16.0 kcal/mol), although these processes are now
endergonic (C2: GDCE = +3.6 kcal/mol). The second half of the
profile is more model-dependent, starting with a much less
favorable HOAc/alkyne substitution (ΔGDCE = +6.8 kcal/mol
cf. +1.4 kcal/mol for model 1) that could be taken to reflect the
greater bulk of model 2 (although see the discussion of
dispersion effects below). D undergoes migratory insertion with
a barrier of 13.5 kcal/mol, somewhat higher than model 1,
although this step is now much less exergonic (ΔGDCE = −4.6
1961
dx.doi.org/10.1021/jo402592z | J. Org. Chem. 2014, 79, 1954−1970

The Journal of Organic Chemistry
Article
Figure 5. Computed structures of the AMLA/CMD C−H bond activation transition states, TS(B−C)1 and TS(B−C)2, for model 2, with selected
distances (Å) and relative free energies in solution (kcal/mol) at the BP86-D3 level. Spectator H atoms are omitted for clarity.
either ligand loss or substitution. Thus, INT(A−B) and the
preceding transition state are both stabilized reflecting the
movement of the 3-phenylpyrazole into the metal coordination
sphere. Similarly, the loss of HOAc from B·HOAc is now more
endergonic (+9.3 kcal/mol cf. +1.8 kcal/mol with the BP86
functional alone). The net result is that intermediate B and the
subsequent stationary points associated with the intramolecular
C−H activation are at similar relative energies with either BP86
or BP86-D3. A major effect is computed for the HOAc/4-
octyne substitution (C2 to D, BP86: ΔGDCE = +6.8 kcal/mol;
BP86-D3: ΔGDCE = −1.5 kcal/mol). As the barrier to migratory
insertion is similar with both approaches (ca. 14 kcal/mol), a
significant stabilization of TS(D-E) is computed at the BP86-
D3 level (GDCE = +14.6 kcal/mol).29
In comparing the BP86-D3 profile with experiment, the
stabilization of INT(A−B) means that the subsequent
energetics must be quoted relative to this species. The
formation of C2 is now endergonic by +8.9 kcal/mol and has
an increased barrier of 21.4 kcal/mol. H/D exchange is
therefore reversible, with the equilibrium lying to the
uncyclometalated reactants. Most importantly TS(D−E)
(GDCE = +14.6 kcal/mol) for alkyne insertion29 is now at a
similar energy to the N−H/C−H activation high point
(TS(B−C)2, GDCE = +14.6 kcal/mol): the two processes are
therefore now competitive, more consistent with reduced ortho-
H/D exchange in the presence of alkyne.30 The overall barrier
for the catalytic cycle is 21.4 kcal/mol. The BP86-D3 profile
therefore captures much of the experimental behavior, with the
transition states for C−H activation and alkyne insertion being
very close in energy. We therefore propose the BP86-D3
protocol provides a robust approach to modeling the current
system. Very similar behavior was computed with the M06 and
B97D functionals, which also include a treatment of dispersion
effects, although interestingly with M06L the alkyne insertion
transition state remains 6.2 kcal/mol above that for C−H
activation; at present we have no explanation for this behavior,
although it appears to be linked to a less favorable energy for
HOAc/4-octyne exchange (C2→D, ΔG = +10.5 kcal/mol)
computed with this functional. A range of nondispersion
corrected functionals (PBE, PBE0, BLYP, B3LYP) gave similar
results to BP86, i.e., clearly (and incorrectly) suggesting rate-
limiting alkyne insertion. As with BP86, the addition of the D3
dispersion correction to the PBE and PBE0 results again
brought the free energies of TS(B−C)2 and TS(D−E) to
within 1 kcal/mol of each other. For BLYP and B3LYP, TS(D−
E) is computed to lie approximately 14 kcal/mol above TS(B−
C)2 and although the dispersion correction reduces this to ca. 5
kcal/mol, alkyne insertion is still clearly identified as the rate-
determining step. This, along with the rather high overall
barriers computed with the BLYP-D3 and B3LYP-D3
approaches (in excess of 28 kcal/mol), suggest these func-
tionals are a poor choice for the current system. Full details of
all functional testing are given in the Supporting Information.
To interpret the observed kH/kD value of 2.7 0.5 we have
compared the computed geometries of the two transition states
associated with the C−H bond activation process (see Figure
5). TS(B−C)1 corresponds to the initial κ2-κ1 displacement of
acetate by the incoming substrate, and as such, the C3−H2
bond is remote from the Rh center and shows little bond
elongation (1.11 Å). In contrast, the transition state for C−H
bond cleavage, TS(B−C)2, exhibits significant bond elongation
(C3···H2 = 1.30 Å). In this case, as TS(B−C)2 is higher in
energy than TS(B−C)1 it corresponds to the rate-determining
transition state and the elongation of the C3−H2 bond is
therefore consistent with the significant kH/kD value.
Computation of the kH/kD KIE gives a value of 5.48, somewhat
higher than the experimental figure, although it should be borne
in mind that the latter will also include a contribution from the
reverse reaction. However, more generally, our studies on
AMLA/CMD C−H bond activation have shown the presence
of both transition states is rather system-dependent and that in
several cases only the initial κ2−κ1 acetate displacement
transition state (equivalent to TS(B−C)1) is located.2b,e In
such cases, the kH/kD value would be expected to be much
closer to 1. Thus the nonobservation of a significant kinetic
isotope effect may not necessarily mean that C−H bond
activation can be ruled out as the rate-limiting step. Indeed, this
appears to be the case for the {Ru(p-cymene)} system
discussed below.
Reaction profiles have also been computed for the coupling
of MeCCPh with pyrazole 1a at Rh(OAc)2Cp* (i.e., model 2)
to probe the regioselectivity of this process. Similar profiles are
1962
dx.doi.org/10.1021/jo402592z | J. Org. Chem. 2014, 79, 1954−1970

The Journal of Organic Chemistry
Article
obtained to that computed for 4-octyne (see Table C3,
Supporting Information), with the lower energy migratory
insertion transition state, TS(D−E), being computed when the
Ph group is adjacent to the metal (GDCE = 14.5 kcal/mol, 2.4
kcal/mol more stable than the alternative when Me is adjacent
to the metal). Thus, the calculations correctly model the
regioselectivity observed experimentally. A similar preference of
1.9 kcal/mol is also computed with the BP86 functional alone,
indicating that dispersion effects and, hence, steric effects are
not important in determining the regioselectivity. Instead,
electronic effects dominate, and indeed a similar regioselectivity
has been observed for alkyne insertion into 5-membered
phosphanickelacycles.31 In that case, DFT calculations linked
this preference to the asymmetric distribution of the frontier
molecular orbitals of MeCCPh, where both the π-bonding
HOMO and π-antibonding LUMO have a greater contribution
on the Me-substituted carbon. As both these orbitals participate
in C−C bond fomation in the migratory insertion transition
state this structure is therefore stabilized when the C(Me)
center is involved in this process, i.e., when the Ph substitutent
carbon is adjacent to the metal, as is the preference seen
experimentally.
proportion of this species increased when catalytic Cu(OAc)2
was employed. This implies that side-product formation is
enhanced when reoxidation of the Rh(I) species based on F is
slow and, hence, that reorganization of F may be linked to side
product formation. One such possible rearrangement is shown
in Figure 6 and starts with slippage of the η4-bound species
FMePh (GDCE = −13.2 kcal/mol) formed upon C−N reductive
coupling, to an η2-bound species F2MePh (GDCE = −12.2 kcal/
mol) where Rh interacts with the C4−C5 bond. F2MePh also
features an agostic interaction with one C6−H bond and readily
undergoes β-H transfer to give a hydrido allyl intermediate
F3MePh at −20.9 kcal/mol. H transfer back onto C5 would
generate F4MePh from which 6ac could dissociate; however, the
high energy of F4MePh (−4.9 kcal/mol) suggests this is unlikely.
An alternative is that oxidation of F3MePh can induce C−H
reductive elimination, either with reformation of the C6−H
bond (to give 3ac), or with C5−H bond formation (to give
6ac).
Having good agreement between experiment and computa-
tion for rhodium, we applied the same approach to ruthenium.
The reaction profile computed at the BP86-D3 level for the
coupling of pyrazole 1a with 4-octyne at Ru(OAc)2(p-cymene)
is shown in Figure 7, where the solvent corrcetion has been
performed with MeOH, reflecting the t-AmOH solvent used
experimentally. The {Ru(p-cymeme)} system is observed to
undergo slower H/D exchange in the absence of alkyne, and
this is reflected in the computed barriers to both the forward
C−H activation (25.4 kcal/mol) and the reverse process (16.9
kcal/mol), respectively 4.0 and 4.4 kcal/mol higher than for the
{Cp*Rh} system. The major difference is the higher energy of
TS(B−C)1 (GDCE = +16.8 kcal/mol). This corresponds to the
κ2−κ1 acetate dissociation step and presumably reflects stronger
Ru−OAc bonding in this case. Similarly, the energies of TS(A-
B)1 and TS(A-B)2 are both ca. 4 kcal/mol higher than their
{Cp*Rh} counterparts, and both of these also involve cleavage
of an M−O bond. Once the agostic/H-bonded intermediate
INT(B−C) is formed the subsequent C−H bond cleavage is
barrierless on the GDCE surface.32
Another feature observed with MeCCPh is the formation of
6ac as a significant side product. One possible route to this
species would be via C(sp3)−N reductive elimination from an
intermediate such as E2MePh (see Scheme 6), itself formed by
Scheme 6
isomerization earlier in the cycle, for example from EMePh
.
From C2, HOAc/nPrCCnPr substitution is slightly ender-
gonic giving D at +2.3 kcal/mol. The subsequent migratory
insertion has a somewhat lower barrier than at {Cp*Rh} (9.6
kcal/mol cf. 14.0 kcal/mol), and this, along with the facile
reductive coupling, means that the C−H activation is clearly the
However, the computed transition state barrier for this
C(sp3)−N bond formation is extremely high (+47.3 kcal/
mol) and effectively rules out this possibility.
An analogous side product, 4aa, was also seen with 4-octyne,
and during the reaction screening it was noted that the
Figure 6. Possible mechanism of formation of isomerized heterocyclic product 6ac. Relative free energies (GDCE, kcal/mol) are computed with the
BP86-D3 protocol.
1963
dx.doi.org/10.1021/jo402592z | J. Org. Chem. 2014, 79, 1954−1970

The Journal of Organic Chemistry
Article
n
Figure 7. Computed reaction profiles (GMeOH, kcal/mol) for the coupling of PrCCnPr with 1a at Ru(OAc)2(p-cymene) (BP86-D3 protocol).
rate-limiting process in the Ru catalysis. The overall barrier for
the coupling process is 25.4 kcal/mol, 4.0 kcal/mol higher than
for Rh(OAc)2Cp* and reflecting the harsher conditions
required experimentally.33 Moreover, the geometry of the
rate-determining transition state, TS(B−C)1, is very similar to
that computed at {RhCp*} in Figure 5 and so features minimal
C−H bond elongation. Thus, despite the fact that the C−H
bond activation process is rate-limiting, it does not follow that a
significant kH/kD should be expected. Indeed, in agreement
with this, the experimentally determined kH/kD is only 1.1
0.2, while the calculated value is 1.23.
of the computed energy profiles with the experimental data
highlights the importance of using the full experimental system
in the calculations, as well as including a treatment of dispersion
effects in order to model correctly the observed competition
between C−H activation and alkyne migratory insertion. The
use of oversimplified models is particularly unrealistic in this
case, while calculations on the full system that omit a dispersion
correction severely overestimate the energy of the alkyne
migratory insertion transition state. This is linked to the
computed energy of the HOAc/alkyne substitution step, which
is particularly sensitive to dispersion effects and overly
endergonic in their absence. A significantly higher overall
barrier to catalysis is computed for ruthenium, in which the
rate-limiting process also corresponds to C−H activation. This
is consistent with the harsher reaction conditions employed
experimentally and reflects the need to dissociate a stronger
Ru−O bond in the κ2−κ1-acetate displacement step that in this
case leads directly to C−H bond cleavage.
The BP86-D3 calculations correctly reproduce the close
competition between C−H bond activation and alkyne
migratory insertion seen in the Rh system, as well as identifying
C−H bond activation as being rate limiting for Ru. The
calculations show that the observed deuterium isotope effects
(Rh: kH/kD = 2.7 0.5; Ru: 1.1 0.2) can both be consistent
with rate-limiting C−H activation due to the one- or two-step
nature of the C−H activation process. For Rh the higher lying
transition state corresponds to C−H bond cleavage with
significant bond elongation. However, for Ru the highest lying
transition state involves the κ2−κ1 displacement of acetate
which occurs without any significant lengthening of the C−H
bond. Thus, for AMLA/CMD C−H activation reactions the
non-observation of a significant kinetic isotope effect does not
CONCLUSIONS
■
The Rh- and Ru-catalyzed formation of a range of
pyrazoloisoquinolines has been demonstrated using a C−H
functionalization strategy based on the coupling of 3-phenyl-
pyrazoles with aryl- and alkylalkynes. For rhodium, [Rh-
(MeCN)3Cp*][PF6]2 is shown to be a more effective precursor
than [RhCl2Cp*]2 with efficient catalysis achieved at 80 °C.
While catalysis with [RuCl2(p-cymene)]2 requires higher
temperatures, for both metals the reactions can be run using
catalytic rather than stoichiometric copper as the reoxidant.
Deuteration and competition experiments on the rhodium
catalysis suggest that C−H activation is reversible and rate
limiting, but that the subsequent reaction with alkynes can be
competitive with this process. For ruthenium, rate-limiting C−
H bond activation is much less reversible and in the presence of
alkyne is essentially irreversible.
DFT calculations on both rhodium and ruthenium catalysis
indicate a mechanism involving sequential N−H and AMLA/
CMD C−H bond activation, HOAc/alkyne exchange,
migratory insertion, and C−N reductive coupling. Comparison
1964
dx.doi.org/10.1021/jo402592z | J. Org. Chem. 2014, 79, 1954−1970

The Journal of Organic Chemistry
Article
methyl-1H-pyrazole (1a, 108 mg, 0.68 mmol), Cu(OAc)2·H2O (10
mg, 10 mol %, 0.05 mmol), 4-octyne (2a, 90 mg, 0.43 mmol), and
DCE (10 mL). The crude 1H NMR spectrum showed the presence of
two products in a 3:1 ratio. The products were purified by column
chromatography eluting from 100% dichloromethane to 10% ethyl
acetate in dichloromethane to give 3aa (133 mg, 50%, 0.50 mmol) and
4aa as a yellow oil (65 mg, 24%, 0.24 mmol).
provide sufficient grounds to rule out C−H bond activation
process as the rate-determining step.
EXPERIMENTAL SECTION
■
Unless stated otherwise, all reactions were carried out under an inert
atmosphere of nitrogen, and workup was carried out in air.
Electrospray (ESI) mass spectra, including high resolution, were
recorded in acetonitrile or methanol. FAB mass spectra (including
high resolution) were recorded with NBA as the matrix. NMR spectra
were recorded on spectrometers operating at 400.13 MHz (1H),
[376.50 (19F) and 100.61 MHz (13C)] or at 500 MHz (1H) [125 MHz
(13C)] at ambient temperature; chemical shifts (ppm) are referred to
the residual protic solvent peaks and coupling constants are expressed
1
4aa. H NMR (400 MHz, CDCl3): δ 0.82 (t, J = 7.4 Hz, 3H,
CH2CH2Me), 1.10 (t, J = 7.4 Hz, 3H, CH2Me), 1.23−1.31 (m, 2H,
CH2CH2Me), 1.51−1.61 (m, 1H, CH2CH2Me), 1.67−1.76 (m, 1H,
CH2CH2Me), 2.23 (sext, J = 7.4 Hz, 1H, CH2Me), 2.32 (s, 3H, Me),
2.38 (sext, J = 7.4 Hz, 1H, CH2Me), 5.34 (dd, J = 5.9, 7.8 Hz, 1H, N−
CH), 6.02 (t, J = 7.4 Hz, 1H, CCH), 6.29 (s, 1H, Pz-H), 7.23−7.30
(m, 2H, H2, H3), 7.46−7.49 (m, 1H, H1), 7.51−7.54 (m, 1H, H4). 13
C
{1H} NMR (125 MHz, CDCl3): δ 13.7 (CH2CH2Me), 14.0 (CH2Me),
14.0 (Me), 18.7 (CH2CH2Me), 21.2 (CH2Me), 38.8 (CH2CH2Me),
57.9 (N-CH), 99.7 (Pz), 123.5 (C1), 125.0 (C4), 125.2, 127.8 (C2/C3),
128.2 (C2/C3), 131.9, 132.6, 132.7 (CCH), 138.0, 148.3. ESIMS:
m/z 267 [M + H]+. HRMS (ES): calcd for C18H23N2 [M + H]+
267.1861, found 267.1849
1
in hertz (Hz). Assignments of H NMR and 13C NMR signals were
made where possible, using appropriate standard 2-D NOESY,
COSYDQF, DEPT135 or APT, HMQC(BIRD sequence) and/or
HSQC experiments. All chemicals were obtained from commercial
suppliers and used without further purification except [Cp*Rh-
(MeCN)3][PF6]2,34 3-phenyl-1H-pyrazole,35 3-phenyl-5-methyl-1H-
pyrazole and 3,5-diphenyl-1H-pyrazole,36 3-phenyl-5-(trifluorometh-
yl)-1H-pyrazole,37 3-(2-thiophene)-5-(trifluoromethyl)-1H-pyrazole,37
and (4-nitrophenyl)but-3-yn-1-ol,38 which were prepared using
literature procedures.
General Procedure for Catalysis Reactions with Rh. [Rh-
(MeCN)3Cp*][PF6]2 (33 mg, 5 mol %), the appropriate pyrazole (1
equiv), Cu(OAc)2·H2O (2.5 equiv), alkyne (1.2 equiv), and DCE (10
mL) were added to a Schlenk flask. The flask was sealed with a screw-
cap and then transferred to a preheated oil bath and stirred at 83 °C
for 16 h. The reaction mixture was cooled to room temperature with
continuous stirring and diluted with Et2O (10 mL).The mixture was
transferred to separating funnel, and ammonium hydroxide solution
(10 mL, 2 M) was added. The aqueous layer was extracted with Et2O
(3 × 10 mL), and the organic layers were combined and dried over
MgSO4. The drying agent was removed by filtration, solvent was
removed under reduced pressure, and the product was purified by
column chromatography.
Synthesis of 3ba. Following the general procedure, a Schlenk flask
was loaded with [Rh(MeCN)3Cp*][PF6]2 (33 mg, 5 mol %), 3-
phenyl-1H-pyrazole (1b, 144 mg, 1 mmol), Cu(OAc)2·H2O (500 mg,
2.5 mmol), 4-octyne (2a, 132 mg, 1.2 mmol), and DCE (10 mL). The
product was purified by column chromatography eluting with 50%
ethyl acetate in hexane to give 3ba as a white solid (219 mg, 87%, 0.87
1
mmol). H NMR (400 MHz, CDCl3): δ 1.11 (t, J = 7.0, 7.4 Hz, 6H,
H9, H12), 1.70 (sext, J = 7.0, 7.8 Hz, 2H, H8), 1.83 (sext, J = 7.4, 8.2
Hz, 2H, H11), 2.93−2.97 (m, 2H, H7), 3.26−3.30 (m, 2H, H10), 6.99
(d, J = 2.3 Hz, 1H, H2), 7.50 (td, J = 1.6, 7.8 Hz, 1H, H4), 7.55 (td, J =
1.2, 8.2 Hz, 1H, H5), 7.84 (br dd, J = 0.8, 8.2 Hz, 1H, H6), 7.95 (d, J =
2.3 Hz, 1H, H1), 8.11 (dd, J = 1.6, 7.8 Hz, 1H, H3), 13C {1H} NMR
(100 MHz, CDCl3): δ 14.4 (C9), 14.5 (C12), 21.3 (C11), 23.8 (C8),
29.7 (C7), 30.1 (C10), 97.1 (C2), 119.1, 123.9, 124.0 (C3,C6), 126.3
(C5), 127.6 (C4), 129.0, 136.8, 137.8, 139.9 (C1). ESIMS: m/z 253 [M
+ H]+. HRMS (ES): calcd for C17H21N2 [M + H]+ 253.1705, found
253.1705.
Synthesis of 3ca. Following the general procedure, a Schlenk flask
was loaded with [Rh(MeCN)3Cp*][PF6]2 (33 mg, 5 mol %), 3,5-
diphenyl-1H-pyrazole (1c, 220 mg, 1 mmol), Cu(OAc)2·H2O (500
mg, 2.5 mmol), 4-octyne (2a, 132 mg, 1.2 mmol), and DCE (10 mL).
The product was purified by column chromatography eluting from
40% dichloromethane in petroleum spirit (40−60 °C) to give 3ca as a
white solid (298 mg, 91%, 0.91 mmol). Mp: 85−87 °C. 1H NMR (300
MHz, CDCl3): δ 1.11 (t, J = 5.6, 7.3 Hz, 3H, H10), 1.14 (t, J = 5.6, 7.3
Hz, 3H, H13), 1.65−1.78 (m, 2H, H9), 1.83−1.96 (m, 2H, H12), 2.93−
2.98 (m, 2H, H8), 3.31−3.36 (m, 2H, H11), 7.28 (s, 2H, Pz-H), 7.35
(tt, J = 1.2, 7.4, 8.6 Hz, 1H, H1), 7.46 (tt, J = 1.2, 7.4 Hz, 2H, H2), 7.50
(td, J = 1.6, 7.4, 8.6 Hz, 1H, H6), 7.55 (br td, J = 1.6, 7.0, 8.6 Hz, 1H,
H5), 7.84 (dd, J = 1.6, 7.4 Hz, 1H, H7), 8.06 (dd, J = 1.2, 8.2 Hz, 2H,
H3), 8.13 (dd, J = 2.0, 7.0 Hz, 1H, H4). 13C {1H} NMR (100 MHz,
CDCl3): δ 14.5 (C10, C13), 21.3 (C12), 23.8 (C9), 29.8 (C8), 30.0 (C11),
94.1 (Pz), 119.0, 123.8, 123.9 (C4), 124.0 (C7), 126.2, 126.3 (C3),
127.6 (C5), 127.9 (C6), 128.7 (C2), 129.1 (C1), 133.9, 137.0, 139.0,
151.3. ESIMS: m/z 329 [M + H]+. FAB MS: m/z 329 [M + H]+, 299
[M-2(Me)]+, 271 [M-2(Me)-2(CH2)]+. HRMS (ES): calcd for
C23H25N2 [M + H]+ 329.2018, found 329.2012.
Synthesis of 3da. Following the general procedure, a Schlenk flask
was loaded with [Rh(MeCN)3Cp*][PF6]2 (33 mg, 5 mol %), 3-
phenyl-5-(trifluoromethyl)-1H-pyrazole (1d, 212 mg, 1.0 mmol),
Cu(OAc)2·H2O (500 mg, 2.5 mmol), 4-octyne (2a, 110 mg, 1.2
mmol), and DCE (10 mL). The crude 1H NMR spectrum showed the
presence of two products in a 17:1 ratio. The products were purified
by column chromatography eluting from 50% dichloromethane in
petroleum ether (40−60 °C) to 100% dichloromethane to give 3da as
a white powder (255 mg, 80%, 0.80 mmol). Mp: 63−65 °C. 1H NMR
(400 MHz, CDCl3): δ 1.09 (t, J = 7.4 Hz, 3H, H10), 1.11 (t, J = 7.0, 7.4
Hz, 3H, H7), 1.65−1.75 (m, 2H, H6), 1.77−1.86 (m, 2H, H9), 2.93−
2.98 (m, 2H, H5), 3.25−3.29 (m, 2H, H8), 7.21 (s, 1H, Pz-H), 7.54
(td, J = 7.0, 7.4 Hz, 1H, H2), 7.60 (td, J = 7.0, 7.4 Hz, 1H, H3), 7.87
General Procedure for Catalysis Reactions with Ru. [RuCl2(p-
cymene)]2 (31 mg, 5 mol %), the appropriate pyrazole (1 equiv),
AgPF6 (0.2 equiv) if added, Cu(OAc)2·H2O (2.5 equiv), alkyne (1.2
equiv), and t-AmOH (3 mL) were added to a Schlenk flask. The
Schlenk flask was sealed with a screw-cap and then transferred to a
preheated oil bath and stirred at 120 °C for 16 h. The reaction mixture
was cooled to room temperature with continuous stirring and diluted
with Et2O (10 mL).The mixture was transferred to separating funnel,
and ammonium hydroxide solution (10 mL, 2 M) was added. The
aqueous layer was extracted with Et2O (3 × 10 mL), and the organic
layers were combined and dried over MgSO4. The drying agent was
removed by filtration, solvent was removed under reduced pressure,
and the product was purified by column chromatography.
Synthesis of 3aa. Following the general procedure, a Schlenk flask
was loaded with [Rh(MeCN)3Cp*][PF6]2 (22 mg, 5 mol %), 3-
phenyl-5-methyl-1H-pyrazole (1a, 108 mg, 0.68 mmol), Cu(OAc)2·
H2O (341 mg, 1.71 mmol), 4-octyne (2a, 90 mg, 0.43 mmol), and
DCE (10 mL). The product was purified by column chromatography
eluting with 70% dichloromethane in hexane to give 3aa as a white
solid (108 mg, 60%, 0.41 mmol). Mp: 62−64 °C. 3aa was also
obtained with [RuCl2(p-cymene)]2 as catalyst (173 mg, 65%, 0.65
1
mmol). H NMR (400 MHz, CDCl3): δ 1.08 (t, J = 7.4 Hz, 6H,
CH2CH2Me), 1.63−1.72 (m, 2H, H6), 1.77−1.87 (m, 2H, H8), 2.52 (s,
3H, Me), 2.88−2.92 (m, 2H, H5), 3.22−3.26 (m, 2H, H7), 6.74 (s, 1H,
Pz-H), 7.44 (td, J = 1.2, 7.8, 8.2 Hz, 1H, H2), 7.50 (td, J = 1.6, 7.0, 8.2
Hz, 1H, H3), 7.79 (brd, J = 8.6 Hz, 1H, H4), 8.01 (dd, J = 1.2, 7.8 Hz,
1H, H1). 13C {1H} NMR (100 MHz, CDCl3): δ 14.2 (CH2CH2Me),
14.3 (CH2CH2Me), 14.4 (Me), 21.3 (C8), 23.8 (C6), 29.6 (C5), 29.9
(C7), 96.7 (Pz), 117.8, 123.5, 123.8 (C4), 123.9 (C1), 125.9 (C2/C3),
127.3 (C2/C3), 129.0, 136.7, 138.6, 149.4. ESIMS: m/z 267 [M + H]+.
HRMS (ES): calcd for C18H23N2 [M + H]+ 267.1861, found 267.1857.
Synthesis of 3aa with Catalytic Copper. A Schlenk flask was
loaded with [Rh(MeCN)3Cp*][PF6]2 (22 mg, 5 mol %), 3-phenyl-5-
1965
dx.doi.org/10.1021/jo402592z | J. Org. Chem. 2014, 79, 1954−1970

The Journal of Organic Chemistry
Article
(br d, J = 2.0 Hz, 1H, H4), 8.08 (dt, J = 2.0 Hz, 1H, H1). 13C {1H}
NMR (100 MHz, CDCl3): δ 14.2 (C7/C10), 14.4 (C7/C10), 21.2 (C9),
23.7 (C6), 29.8 (C5, C8), 95.6 (Pz), 121.4, 121.9 (q, J = 273.6 Hz,CF3),
123.5, 123.2, 123.9 (C1), 124.2 (C4), 126.9 (C2), 128.4 (C3), 129.1,
136.8, 138.5, 142.4 (q, J = 39.2 Hz,C-CF3). 19F {1H} NMR (376 MHz,
CDCl3): δ −61.4 (CF3). HRMS (ES): calcd for C18H20N2F3 [M +
H]+321.1579, found 321.1566.The product was recrystallized from
dichloromethane/hexane to give 3da as clear needles.
123.4 (C1), 123.6, 126.0, 126.9 (C4), 127.3 (Ar), 127.6 (Ar), 127.8
(C2), 128.0 (Ar), 128.3 (C3), 128.4 (Ar), 129.9, 131.0 (Ar), 131.2
(C5), 131.8, 135.5, 136.2, 139.2, 143.3 (q, J = 38.0 Hz, CF3). 19F {1H}
NMR (376 MHz, CDCl3): δ −61.3 (CF3). ESIMS: m/z 389 [M +
H]+. HRMS (ES): calcd for C24H16N2F3 [M + H]+389.1266, found
389.1263. The product was recrystallized from dichloromethane/
hexane to give 3db as clear crystals.
Synthesis of 3ac and 3ac′. Following the general procedure, a
Schlenk flask was loaded with [Rh(MeCN)3Cp*][PF6]2 (33 mg, 5 mol
%), 3-phenyl-5-methyl-1H-pyrazole (1a, 158 mg, 1.00 mmol),
Cu(OAc)2·H2O (500 mg, 2.50 mmol), 1-phenyl-1-propyne (2c, 139
Synthesis of 3ab. Following the general procedure, a Schlenk flask
was loaded with [Rh(MeCN)3Cp*][PF6]2 (10 mg, 5 mol %), 3-
phenyl-5-methyl-1H-pyrazole (1a, 50 mg, 0.32 mmol), Cu(OAc)2·
H2O (161 mg, 0.81 mmol), diphenylacetylene (2b, 68 mg, 0.38
mmol), and DCE (10 mL). The product was purified by column
chromatography eluting from 70% dichloromethane in hexane to give
3ab as an orange solid (97 mg, 91%, 0.29 mmol). Compound 3ab was
also obtained with [RuCl2(p-cymene)]2 as catalyst with AgPF6 (25 mg,
0.1 mmol) as additive (yield 14%, based on NMR integration against
1
mg, 1.20 mmol), and DCE (10 mL). The crude H NMR spectrum
showed the presence of three products in a 12:9:1 ratio. The products
were purified by column chromatography eluting with 30% ethyl
acetate in hexane to give 3ac as an orange powder (122 mg, 45%, 0.45
mmol) (mp 98−101 °C), 3ac′ as orange oil (impure), and isomer 6ac
as an orange powder (110 mg, 40%, 0.40 mmol). Compound 3ac′ was
purified further by preparative TLC eluting with 30% ethyl acetate in
hexane to give 3ac′ as an orange oil (13 mg, 5%, 0.05 mmol).
Compounds 3ac and 3ac′ (218 mg, 80%, 0.80 mmol) were obtained
with [RuCl2(p-cymene)]2 as catalyst in a 12:1 ratio and purified by
column chromatography to give 3ac (182 mg, 67%, 0.67 mmol).
1
an internal standard). H NMR (400 MHz, CDCl3): δ 3.71 (s, 3H,
Me), 6.89 (s, 1H, Pz-H), 7.15−7.18 (dd, J = 1.2, 7.4 Hz, 2H, Ar-H),
7.21−7.27 (m, 4H, Ar−H), 7.31−7.33 (m, 3H, Ar-H), 7.37−7.39 (m,
2H, Ar-H), 7.50−7.54 (m, 2H, H2, Ar-H), 8.11 (d, J = 7.4 Hz, 1H, H1).
13C {1H} NMR (100 MHz, CDCl3): δ 43.4 (Me), 97.4 (Pz), 122.9
(Ar), 123.3 (Ar), 123.5 (Ar), 126.6 (Ar), 127.0 (Ar), 127.1 (Ar), 127.5
(Ar), 127.7 (Ar), 127.9 (Ar), 128.2 (Ar), 128.2 (Ar), 128.3 (Ar),
128.7, 130.0 (Ar), 131.1 (Ar), 131.6 (Ar), 131.7 (Ar), 133.2 (Ar),
136.2 (Ar), 136.5, 139.4, 150.7. ESIMS: m/z 335 [M + H]+. HRMS
(ES): calcd for C24H19N2 [M + H]+ 335.1548, found 335.1541.
Synthesis of 3bb. Following the general procedure, a Schlenk
flask was loaded with [Rh(MeCN)3Cp*][PF6]2 (33 mg, 5 mol %), 3-
phenyl-1H-pyrazole (1b, 144 mg, 1 mmol), Cu(OAc)2·H2O (500 mg,
2.5 mmol), diphenylacetylene (2b, 214 mg, 1.2 mmol), and DCE (10
mL). The product was purified by column chromatography with ethyl
acetate/dichloromethane as eluant to give 3bb as a white solid (250
mg, 78%, 0.78 mmol). 1H NMR (400 MHz, CDCl3): δ 7.13 (d, J = 2.0
Hz, 1H, H2), 7.19−7.22 (m, 2H, Ar-H), 7.26−7.35 (m, 8H, Ar-H),
7.40−7.43 (m, 1H, Ar-H), 7.44 (td, J = 1.2, 6.7, 7.0 Hz, 1H, H5), 7.58
(td, J = 2.0, 6.7, 8.2 Hz, 1H, H4), 7.98 (d, J = 2.0 Hz, 1H, H1), 8.21 (m,
1H, H3), 13C {1H} NMR (100 MHz, CDCl3): δ 97.6 (C2), 123.6 (C3),
124.1, 126.7 (C5), 127.2 (C4), 127.4, 127.7, 127.9, 128.01, 128.4,
130.9, 131.6, 141.1 (C1). ESIMS: m/z 321 [M + H]+. HRMS (ES):
calcd for C23H17N2 [M + H]+321.1392, found 321.1394.
1
3ac. H NMR (400 MHz, CDCl3): δ 2.29 (s, 3H, CCMe), 2.41
(s, 3H, Me), 6.79 (s, 1H, Pz-H), 7.45−7.50 (m, 5H, H5, H6, H7), 7.51−
7.56 (m, 2H, H2, H3), 7.84−7.86 (m, 1H, H4), 8.06−8.09 (m, 1H, H1).
13C {1H} NMR (100 MHz, CDCl3): δ 14.3 (Me), 15.0 (CCMe),
97.0 (Pz), 115.0, 123.8 (C1), 123.9, 124.3 (C4), 127.0 (C2/C3), 127.6
(C2/C3), 128.5 (C5/C6), 128.7 (C7), 129.9, 130.6 (C5/C6), 134.0,
135.8, 138.9, 150.0. ESIMS: m/z 273 [M + H]+. HRMS (ES): calcd
for C19H17N2 [M + H]+ 273.1392, found 273.1392.
1
3ac′. H NMR (400 MHz, CDCl3): δ 2.56 (s, 3H, CCMe), 2.59
(s, 3H, Me), 6.88 (s, 1H, Pz-H), 7.24 (br d, J = 7.8 Hz, 1H, H4), 7.34
(dt, J = 2.0, 6.7, 8.6 Hz, 2H, H5), 7.35−7.39 (m, 1H, H3), 7.44−7.53
(m, 4H, H2, H6, H7), 8.08 (dd, J = 0.8, 7.8 Hz, 1H, H1). 13C {1H}
NMR (100 MHz, CDCl3): δ 14.2 (Me), 15.9 (CCMe), 97.4 (Pz),
121.6, 122.8, 123.4 (C1), 126.0 (C4), 126.2 (Ar), 127.4 (C3), 127.6
(Ar), 128.6 (C6), 129.9, 131.0 (C5), 133.1, 137.1, 139.0, 150.3. ESIMS:
m/z 273 [M + H]+. HRMS (ES): calcd for C19H17N2 [M + H]+
273.1392, found 273.1409.
1
6ac. H NMR (400 MHz, CDCl3): δ 2.32 (s, 3H,Me), 5.49 (s, 1H,
CCH2), 5.67 (s, 1H, CCH2), 6.15 (s, 1H, CH(Ph)), 6.42 (s, 1H,
Pz-H), 6.94 (dd, J = 2.3, 8.2 Hz, 2H, H5), 7.14−7.20 (m, 3H, H6, H7),
7.22 (td, J = 1.2, 7.4, 7.8 Hz, 1H, H3), 7.32 (td, J = 1.2, 7.4, 7.8 Hz, 1H,
H2), 7.52 (dd, J = 1.2, 7.8 Hz, 1H, H4), 7.54 (dd, J = 1.2, 7.8 Hz, 1H,
H1). 13C {1H} NMR (100 MHz, CDCl3): δ 13.8 (Me), 66.4
(CH(Ph)), 100.5 (Pz), 115.3 (CCH2), 123.7 (C4), 125.5 (C5),
125.5, 127.6 (C7), 128.4 (C3), 128.5 (C2), 128.6 (C6), 128.8 (C1),
129.1, 129.8, 139.0, 140.6, 142.3, 149.6. ESIMS: m/z 273 [M + H]+.
HRMS (ES): calcd for C19H17N2 [M + H]+ 273.1392, found 273.1391.
Synthesis of 3bc and 3bc′. Following the general procedure, a
Schlenk flask was loaded with [Rh(MeCN)3Cp*][PF6]2 (33 mg, 5 mol
%), 3-phenyl-1H-pyrazole (1b, 144 mg, 1 mmol), Cu(OAc)2·H2O
(500 mg, 2.5 mmol), 1-phenyl-1-propyne (2c, 139 mg, 1.2 mmol), and
DCE (10 mL). The crude 1H NMR spectrum showed the presence of
free pyrazole and two products in a 12:1 ratio. The products were
purified by column chromatography eluting from 100% dichloro-
methane to 20% ethyl acetate in petroleum ether (40−60 °C) to give
3bc′ as orange oil (impure) and 3bc as an orange powder (111 mg,
79% (based on 54% conversion), 0.43 mmol).
Synthesis of 3cb. Following the general procedure, a Schlenk flask
was loaded with [RhCp*(MeCN)3][PF6]2 (33 mg, 5 mol %), 3,5-
diphenyl-1H-pyrazole (1c, 220 mg, 1.0 mmol), Cu(OAc)2·H2O (500
mg, 2.5 mmol), diphenylacetylene (2b, 214 mg, 1.2 mmol), and DCE
(10 mL). The product was purified by column chromatography eluting
from 100% dichloromethane to give 3cb as an orange solid (390 mg,
1
98%, 0.98 mmol). H NMR (400 MHz, CDCl3): δ 7.21 (dd, J = 1.6,
7.4 Hz, 2H, Ar−H), 7.26−7.34 (m, 6H), 7.36−7.44 (m, 6H, Ar-H),
7.42 (s, 1H, Pz-H), 7.56−7.60 (m, 1H, Ar-H), 7.92 (dd, J = 1.2, 7.0 Hz,
2H, Ar-H), 8.23 (d, J = 7.8 Hz, 1H, Ar-H), 13C {1H} NMR (100 MHz,
CDCl3): δ 94.6 (Pz), 123.6, 124.0, 126.8, 127.1, 127.3, 127.5, 127.7,
128.0, 128.1, 128.4, 128.6, 130.0, 131.4, 131.8, 133.0,133.5,
136.4,139.9, 152.3. ESIMS: m/z 397[M + H]+. HRMS (ES): calcd
for C29H21N2 [M + H]+397.1705, found 397.1711.The product was
recrystallized from standing in chloroform to give 3cb as orange
blocks. Anal. Calcd for (C29H21N2): C, 87.85; H, 5.08; N, 7.07. Found:
C, 87.99; H, 4.95; N, 6.95%.
Synthesis of 3db. Following the general procedure, a Schlenk
flask was loaded with [Rh(MeCN)3Cp*][PF6]2 (16 mg, 5 mol %), 3-
phenyl-5-(trifluoromethyl)-1H-pyrazole (1d, 106 mg, 0.50 mmol),
Cu(OAc)2·H2O (250 mg, 1.25 mmol), diphenylacetylene (2b, 107 mg,
0.60 mmol), and DCE (5 mL). The product was purified by column
chromatography eluting from 50% dichloromethane in petroleum
ether (40−60 °C) to give 3db as a white powder (175 mg, 90%, 0.45
mmol). 1H NMR (400 MHz, CDCl3): δ 7.18 (dd, J = 2.3, 7.8 Hz, 2H,
H5), 7.24−7.33 (m, 8H, Ar-H), 7.36 (s, 1H, Pz-H), 7.45 (dd, J = 1.6,
8.2 Hz, 1H, H4), 7.49 (td, J = 1.2, 7.0, 8.2 Hz, 1H, H3), 7.62 (td, J =
1.2, 7.0, 8.2 Hz, 1H, H2), 8.20 (dd, J = 1.2, 8.2 Hz, 1H, H1). 13C {1H}
NMR (125 MHz, CDCl3): δ 95.9 (Pz), 121.6 (q, J = 270.1 Hz, CF3),
3bc′. This was obtained in a mixture with 3bc. 1H NMR (400 MHz,
CDCl3): δ 2.59 (s, 3H, Me), 7.09 (d,J = 2.3 Hz, 1H, H2), 7.24−7.27
(m, 1H, Ar-H), 7.33 (dd, J = 1.6, 8.2 Hz, 1H, H3/H6), 7.39 (td, J = 1.6,
7.0, 8.2 Hz, 1H, H4/H5), 7.45−7.62 (m, 6H, Ar-H), 8.04 (d, J = 2.0
Hz, 1H, H1).
3bc. 1H NMR (400 MHz, CDCl3): δ 2.35 (s, 3H, Me), 7.03 (d, J =
2.3 Hz, 1H, H2), 7.46−7.57 (m, 5H, Ar−H), 7.58−7.61 (m, 2H, H4,
H5), 7.88 (d, J = 2.0 Hz, 1H, H1), 7.89−7.92 (m, 1H, H6), 8.15−8.17
(m, 1H, H3). 13C {1H} NMR (100 MHz, CDCl3): δ 14.8 (Me), 97.1
(C2), 116.1, 123.7 (C3), 124.0, 124.3 (C6), 127.0 (C4/C5), 127.7 (C4/
C5), 128.5, 128.7 (Ar), 129.7 (Ar), 130.2 (Ar), 133.7, 135.8, 137.8,
1966
dx.doi.org/10.1021/jo402592z | J. Org. Chem. 2014, 79, 1954−1970

The Journal of Organic Chemistry
Article
140.3 (C1). HRMS (ES): calcd for C18H14N2 [M + H]+ 259.1341,
found 259.1336.
H1, H4). 13C {1H} NMR (100 MHz, CDCl3): δ 67.0 (N-CH), 99.4
(Pz), 116.8 (CCH2), 121.4 (q, J = 268.7 Hz,CF3), 124.0 (C1/C4),
124.3, 125.5 (C7), 125.7 (C1/C4), 128.0 (C8/C9), 128.5 (C8/C9), 129.0
(C2/C3), 129.5 (C2/C3), 129.6, 139.1, 139.7, 141.1, 142.8 (q, J = 40.9
Hz, C-CF3). 19F {1H} NMR (376 MHz, CDCl3): δ −61.9 (CF3).
HRMS (ES): calcd for C19H14N2F3 [M + H]+ 327.1109, found
327.1123.
Synthesis of 3cc and 3cc′. Following the general procedure, a
Schlenk flask was loaded with [Rh(MeCN)3Cp*][PF6]2 (33 mg, 5 mol
%), 3,5-diphenyl-1H-pyrazole (1c, 220 mg, 1 mmol), Cu(OAc)2·H2O
(500 mg, 2.50 mmol), 1-phenyl-1-propyne (2c, 139 mg, 1.20 mmol),
and DCE (10 mL). The crude 1H NMR spectrum showed the
presence of two products in a 3:1 ratio. The products were purified by
column chromatography eluting from 50% dichloromethane in
petroleum ether (40−60 °C) to 100% dichloromethane to give 3cc
and 3cc′ (219 mg, 66% combined yield, 0.66 mmol). Compound 3cc′
was obtained pure as a brown solid (35 mg, 10%, 0.10 mmol) and 3cc
as an orange solid (147 mg, 44%, 0.44 mmol). Mp: 145−148 °C.
3cc′. 1H NMR (400 MHz, CDCl3): δ 2.64 (s, 3H, Me), 7.26 (d, J =
8.2 Hz, 1H, H7), 7.34−7.42 (m, 4H, H1, H6, H8), 7.39 (s, 1H, Pz-H),
7.46−7.55 (m, 6H, H2, H5, H9, H10), 8.07 (dd, J = 1.6, 8.6 Hz, 2H, H3),
8.18 (br dt, J = 1.2, 2.0, 7.8 Hz, 1H, H4). 13C {1H} NMR (100 MHz,
CDCl3): δ 15.8 (Me), 94.8 (Pz), 122.5, 123.2, 123.5 (C4), 126.2 (C7),
126.4 (C3), 127.6 (Ar), 127.7 (Ar), 128.2 (Ar), 128.7 (C2/C9), 128.7
(C2/C9), 130.0, 131.0 (C8), 133.6, 133.7, 137.1, 139.5, 152.2. ESIMS:
m/z 334 [M + H]+. HRMS (ES): calcd for C24H19N2 [M + H]+
335.1548, found 335.1557.
Synthesis of 3ad and 3ad′. Following the general procedure, a
Schlenk flask was loaded with [Rh(MeCN)3Cp*][PF6]2 (26 mg, 5 mol
%), 3-phenyl-5-methyl-1H-pyrazole (1a, 125 mg, 0.79 mmol),
Cu(OAc)2·H2O (395 mg, 1.98 mmol), (4-nitrophenyl)but-3-yn-1-ol
1
(2d, 182 mg, 0.95 mmol), and DCE (10 mL). The crude H NMR
spectrum showed the presence of two products in a 3:1 ratio. Column
chromatography eluting from 1% ethyl acetate in dichloromethane
gave a mixture of regioisomers (264 mg, 96%, 0.76 mmol).
Recrystallization from dichloromethane/hexane gave 3ad as yellow
crystals (200 mg, 73%, 0.73 mmol). Mp: 170−172 °C.
3ad. 1H NMR (400 MHz, CDCl3): δ 1.36 (br s, 1H, OH), 2.39 (s,
3H, Me), 3.00 (t, J = 7.0 Hz, 2H, CH2CH2OH), 3.83 (br t, J = 5.9 Hz,
2H, CH2CH2OH), 6.83 (s, 1H, Pz-H), 7.60 (t, J = 3.1 Hz, 1H, H2/
H3), 7.62 (t, J = 3.5 Hz, 2H, H2/H3), 7.69 (d, J = 9.0 Hz, 2H, H5), 7.93
(dd, J = 3.1 Hz, 1H, H4), 8.11 (dd, J = 3.1 Hz, 1H, H1), 8.42 (d, J = 9.0
Hz, 2H, H6). 13C {1H} NMR (100 MHz, CDCl3): δ 14.2 (Me), 31.3
(CH2CH2OH), 62.3 (CH2CH2OH), 97.5 (Pz), 115.9, 124.0 (C6),
124.3 (C1), 124.3 (C4), 124.4, 127.7 (C2/C3), 128.0 (C2/C3), 128.4,
131.8 (C5), 135.0, 139.1, 140.4, 148.1, 150.8. ESIMS: m/z 348 [M +
H]+. HRMS (ES): calcd for C20H18N3O3 [M + H]+ 348.1348, found
348.1348. The product was recrystallized from dichloromethane/
hexane to give 3ad as orange needles. Anal. Calcd for (C20H17N3O3):
C, 69.15; H, 4.93; N, 12.10. Found: C, 68.99; H, 4.80; N, 11.98.
3ad′. Recrystallization with dichloromethane/hexane left a filtrate of
3ad′ in a mixture with 3ad. 1H NMR (400 MHz, CDCl3): δ 1.36 (br s,
1H, OH), 2.39 (s, 3H, Me), 3.00 (t, J = 7.0 Hz, 2H, CH2CH2OH),
3.83 (br t, J = 5.9 Hz, 2H, CH2CH2OH), 6.83 (s, 1H, Pz-H), 7.04 (d, J
= 8.2 Hz, 1H, H1/H4), 7.93−7.43 (m, 1H, H2/H3), 7.53−7.56 (m, 1H,
H2/H3), 7.54 (d, J = 9.0 Hz, 2H, H5), 8.11 (d, J = 7.4 Hz, 1H, H1/H4),
8.41 (d, J = 9.0 Hz, 2H, H6).
3cc. 1H NMR (400 MHz, CDCl3): δ 2.37 (s, 3H, Me), 7.27 (tt, J =
1.2, 2.7, 6.7, 8.2 Hz, 1H, H1), 7.32 (s, 1H, Pz-H), 7.35−7.37 (m, 2H,
H2), 7.50−7.55 (m, 5H, H8, H9, H10), 7.57−7.60 (m, 1H, H5/H6),
7.60−7.62 (m, 1H, H5/H6), 7.87 (dd, J = 1.2, 8.2 Hz, 2H, H3), 7.89−
7.91 (m, 1H, H7), 8.16−8.20 (m, 1H, H4). 13C {1H} NMR (125 MHz,
CDCl3): δ 15.2 (Me), 94.3 (Pz), 116.1, 123.7 (C4), 123.8, 124.1 (C7),
124.5 (C3), 127.2 (C1/C5/C6), 127.8 (C1/C5/C6), 127.9 (C1/C5/C6),
128.2 (C2/C9), 128.5 (C2/C9), 128.6 (C10), 130.0, 130.9 (C8), 133.6,
133.7, 136.1, 139.3, 151.6. ESIMS: m/z 334 [M + H]+. HRMS (ES):
calcd for C24H19N2 [M + H]+335.1548, found 335.1560. The product
was recrystallized from dichloromethane/hexane to give 3cc as orange
blocks.
Synthesis of 3dc and 3dc′. Following the general procedure, a
Schlenk flask was loaded with [Rh(MeCN)3Cp*][PF6]2 (33 mg, 5 mol
%), 3-phenyl-5-(trifluoromethyl)-1H-pyrazole (1d, 212 mg, 1.0
mmol), Cu(OAc)2·H2O (500 mg, 2.5 mmol), 1-phenyl-1-propyne
Synthesis of 3cd and 3cd′. Following the general procedure, a
Schlenk flask was loaded with [Rh(MeCN)3Cp*][PF6]2 (16 mg, 5 mol
%), 3,5-diphenyl-1H-pyrazole (1c, 110 mg, 0.5 mmol), Cu(OAc)2·
H2O (250 mg, 1.25 mmol), (4-nitrophenyl)but-3-yn-1-ol (2d, 115 mg,
1
(2c, 139 mg, 1.2 mmol), and DCE (10 mL). The crude H NMR
spectrum showed the presence of three products in a 7:2:1 ratio. The
products were purified by column chromatography eluting from 50%
dichloromethane in petroleum ether (40−60 °C) to give 3dc′ as an
orange solid (30 mg, 9%, 0.09 mmol), 3dc as an orange powder (178
mg, 55%, 0.55 mmol) (mp 135−137 °C), and isomer 6dc as an orange
oil (53 mg, 16%, 0.16 mmol) which was purified further by preparative
TLC eluting with 50% dichloromethane in petroleum ether (40−60
°C).
1
0.6 mmol), and DCE (5 mL). The crude H NMR spectrum showed
the presence of two products in a 3:1 ratio. The products were purified
by column chromatography eluting with 50% ethyl acetate in hexane
to give 3cd as a yellow solid (90 mg, 44%, 0.22 mmol) (mp 200−202
°C), and 3cd′ as a yellow solid (80 mg, 39%, 0.20 mmol). Mp: 175−
178 °C.
3dc′. H NMR (400 MHz, CDCl3): δ 2.60 (s, 3H, Me), 7.28−7.34
3cd. 1H NMR (400 MHz, CDCl3): δ 1.54 (br s,1H, OH), 3.08 (t, J
= 7.0 Hz, 2H, CH2CH2OH), 3.87 (br q, J = 6.7 Hz, 2H, CH2CH2OH),
7.30 (tt, J = 1.2, 7.4, 8.6 Hz, 1H, H7), 7.35 (s, 1H, Pz-H), 7.37 (tt, J =
1.2, 1.6, 7.0, 7.4 Hz, 2H, H6), 7.62−7.68 (m, 2H, H2, H3), 7.76 (d, J =
8.6 Hz, 2H, H8), 7.83 (dd, J = 1.6, 7.0 Hz, 2H, H5), 7.97−8.00 (m, 1H,
H4), 8.21−8.24 (m, 1H, H1), 8.42 (d, J = 8.6 Hz, 2H, H9). 13C {1H}
NMR (100 MHz, CDCl3): δ 31.2 (CH2CH2OH), 62.0 (CH2CH2OH),
94.5 (Pz), 116.8, 123.5 (C9), 124.1 (C1), 124.3 (C4), 124.5, 126.1 (C5),
127.7 (C2/C3), 128.0 (C2/C3), 128.1 (C7), 128.3, 128.4 (C6), 131.9
(C8), 132.8, 135.0, 139.4, 139.9, 147.9, 152.2. ESIMS: m/z 410 [M +
H]+. HRMS (ES): calcd for C25H20N3O3 [M + H]+410.1505, found
410.1502.
1
(m, 4H, Pz-H, H4,H5), 7.44−7.58 (m, 5H, H2,H3, H6,H7), 8.16 (d, J =
7.4 Hz, 1H, H1). 13C {1H} NMR (100 MHz, CDCl3): δ 15.7 (Me),
96.2 (Pz), 119.3 (q, J = 299.8 Hz, CF3), 123.0, 123.5 (C1), 124.8, 126.5
(Ar), 127.2 (Ar), 128.1 (Ar), 128.4 (Ar), 128.8 (Ar), 130.0, 130.7
(Ar), 133.5, 136.3, 138.9, 143.1 (q, J = 35.9 Hz, C-CF3), 19F {1H}
NMR (376 MHz, CDCl3): δ −61.4 (CF3). HRMS (ES): calcd for
C19H14N2F3 [M + H]+ 327.1109, found 327.1109.
3dc. 1H NMR (400 MHz, CDCl3): δ 2.37 (s, 3H, Me), 7.25 (s, 1H,
Pz-H), 7.45 (dd, J = 2.0, 7.8 Hz, 2H, H5), 7.49−7.55 (m, 3H, H6,H7),
7.60−7.67 (m, 2H, H2,H3), 7.91−7.94 (m, 1H, H4), 8.12−8.14 (m,
2H, H1). 13C {1H} NMR (100 MHz, CDCl3): δ 15.3 (Me), 95.9 (Pz),
118.7, 121.7 (q, J = 268.9 Hz, CF3), 123.8, 123.8 (C1), 124.7 (C4),
127.8 (C2/C3), 128.5 (C6), 128.7 (C2/C3), 129.0 (C7), 129.9, 130.8
(C5), 132.7, 135.9, 138.8, 143.1 (q, J = 39.5 Hz, C-CF3). 19F {1H}
NMR (376 MHz, CDCl3): δ −61.2 (CF3). HRMS (ES): calcd for
C19H14N2F3 [M + H]+ 327.1109, found 327.1114. The product was
recrystallized from dichloromethane/hexane to give 3dc as orange
blocks.
1
3cd′. H NMR (400 MHz, CDCl3): δ 3.32 (t, J = 5.5 Hz, 2H,
CH2CH2OH), 4.02 (q, J = 5.5 Hz, 2H,CH2CH2OH), 4.16 (t, J = 5.5
Hz, 1H, OH), 7.09 (d, J = 7.8 Hz, 1H, H4), 7.42 (s, 1H, Pz-H), 7.41−
7.51 (m, 4H, H3, H6, H7), 7.56−7.61 (m, 1H, H2), 7.57 (dd, J = 6.7,
9.0 Hz, 2H, H8), 8.02 (dd, J = 1.6, 7.0 Hz, 2H, H5), 8.21 (d, J = 8.2 Hz,
1H, H1), 8.41 (dd, J = 6.7, 8.6 Hz, 2H, H9). 13C {1H} NMR (100
MHz, CDCl3): δ 31.6 (CH2CH2OH), 61.0 (CH2CH2OH), 94.3 (Pz),
120.9, 122.3, 122.8 (C1), 123.0 (C9), 124.8 (C4), 125.4 (C5), 126.5
(Ar), 127.3 (Ar), 127.7 (Ar), 127.9 (C6), 128.1, 131.2 (C8), 131.6,
134.3, 139.0, 142.6, 146.7, 151.7. ESIMS: m/z 410 [M + H]+. HRMS
(ES): calcd for C25H20N3O3 [M + H]+410.1505, found 410.1494.
1
6dc. H NMR (400 MHz, CDCl3): δ 5.56 (d, J = 0.8 Hz, 1H, H6),
5.76 (s, 1H, H5), 6.27 (s, 1H, N−CH), 6.87 (s, 3H, Pz-H), 6.94−6.96
(m, 2H, H7), 7.17−7.21 (m, 3H, H8,H9), 7.32 (td, J = 1.6, 7.4 Hz, 1H,
H2/H3), 7.38 (td, J = 1.6, 7.4, 7.8 Hz, 1H, H2/H3), 7.56−7.60 (m, 2H,
1967
dx.doi.org/10.1021/jo402592z | J. Org. Chem. 2014, 79, 1954−1970

The Journal of Organic Chemistry
Article
Synthesis of 3dd and 3dd′. Following the general procedure, a
Schlenk flask was loaded with [Rh(MeCN)3Cp*][PF6]2 (16 mg, 5 mol
%), 3-phenyl-5-(trifluoromethyl)-1H-pyrazole (1d, 212 mg, 1.0
mmol), Cu(OAc)2·H2O (500 mg, 2.5 mmol), 2d, 4-(nitrophenyl)-
Initial Rates Experiments. Reaction vials equipped with a stirrer
bar were loaded with 5-methyl-3-phenyl-1H-pyrazole (1a, 52 mg, 0.33
mmol), 4-octyne (2a, 0.3−2.4 equiv), Cu(OAc)2·H2O (165 mg, 0.83
mmol, 2.5 equiv), 1,3,5-trimethoxybenzene (3.0 mg, 0.017 mmol, 5
mol %) as internal standard, and DCE (reaction made up to 5 mL
volume). The vials were sealed with a screw cap and transferred to a
preheated heating block at 50 °C to stir for 5 min. [Rh(MeCN)3Cp*]-
[PF6]2 (11 mg, 0.017 mmol, 5 mol %) was added, and timing was
started. Each reaction was repeated to calculate the average rate. After
a suitable time, a vial was removed and placed in an ice bath to stop
the reaction. The mixture from each vial was diluted with diethyl ether
(10 mL), and 2 M ammonium solution (10 mL) was added. The blue
aqueous layer was extracted with diethyl ether (3 × 10 mL), and the
combined organic layers were dried over MgSO4. The solvent was
removed on the rotary evaporator. The 1H NMR spectra were
recorded, and yield was found by integrating the CH2 2H peak of the
product. The order in alkyne was found to be 0.07 0.1.
A similar procedure was used with [RuCl2(p-cymene)]2 as catalyst,
except the reactions were carried out a smaller scale, in tert-amyl
alcohol at 100 °C. The amounts used were 5-methyl-3-phenyl-1H-
pyrazole (1a, 32 mg, 0.20 mmol), 4-octyne (2a, 0.3−2.4 equiv),
Cu(OAc)2·H2O (100 mg, 0.50 mmol, 2.5 equiv), 1,3,5-trimethox-
ybenzene (1.7 mg, 0.010 mmol, 5 mol %) as internal standard, and
tert-amyl alcohol (reaction made up to 5 mL volume). The vials were
sealed with a screw cap and transferred to a preheated heating block at
100 °C to stir for 5 min. [RuCl2(p-cymene)]2 (6.1 mg, 0.01 mmol, 5
mol %) was added, and timing was started. The reactions were worked
up as for Rh above.
1
but-3-yn-1-ol (115 mg, 0.6 mmol), and DCE (5 mL). The crude H
NMR spectrum showed the presence of two products in a 1:1 ratio.
The products were purified by column chromatography eluting from
10% ethyl acetate in dichloromethane to give and 3dd and 3dd′ (163
mg, 90% combined yield, 0.45 mmol). Compound 3dd′ was obtained
as an orange solid (18 mg, 9%, 0.05 mmol) and 3dd as an orange solid
(impure).
3dd′. 1H NMR (400 MHz, CDCl3): δ 2.40 (br s, 1H, OH), 3.28 (t,
J = 5.9 Hz, 2H, CH2CH2OH), 4.02 (br q, J = 5.9 Hz, 2H,
CH2CH2OH), 7.11 (d, J = 8.2 Hz, 1H, H4), 7.36 (s, 1H, Pz-H), 7.48−
7.53 (m, 1H, H3), 7.59 (dt, J = 2.3, 9.0 Hz, 2H, H5), 7.62−7.66 (m,
1H, H2), 8.19 (d, J = 7.8 Hz, 1H, H1), 8.42 (dt, J = 2.3, 8.6 Hz, 2H,
H6). 13C {1H} NMR (100 MHz, CDCl3): δ 32.4 (CH2CH2OH), 60.9
(CH2CH2OH), 96.7 (Pz), 123.3, 123.8 (C1), 124.1 (C6), 124.4, 126.1
(C4), 128.2 (C2), 129.1 (C3), 129.2, 132.1 (C5), 134.5, 139.5, 142.9,
148.0. 19F {1H} NMR (376 MHz, CDCl3): δ −61.6 (CF3). ESIMS: m/
z 402 [M + H]+. HRMS (ES): calcd for C20H15N3F3O3 [M +
H]+402.1066, found 402.1067.
1
3dd. This was obtained in a mixture with 3dd′. H NMR (400
MHz, CDCl3): δ 3.07 (t, J = 5.9 Hz, 2H, CH2CH2OH), 3.88 (t, J = 5.9
Hz, 2H, CH2CH2OH), 6.76 (s, 1H, Pz-H), 7.40−8.19 (multiplets, 7H,
Ar-H), 8.34 (dt, J = 2.3, 8.6 Hz, 2H, Ar-H). ESIMS: m/z 402 [M +
H]+, 386 [M − O]+.
Synthesis of 5da. Following the general procedure, a Schlenk flask
was loaded with [Rh(MeCN)3Cp*][PF6]2 (33 mg, 5 mol %), 3-
(thiophene-2-yl)-5-(trifluoromethyl)-1H-pyrazole (222 mg, 1.0
mmol), Cu(OAc)2·H2O (500 mg, 2.5 mmol), 4-octyne (2a, 132 mg,
1.2 mmol), and DCE (10 mL). The crude 1H NMR spectrum showed
the presence of two products in a 4:1 ratio. The products were purified
by column chromatography eluting from 50% dichloromethane in
petroleum ether (40−60 °C) to 100% dichloromethane to give 5da as
a white solid (261 mg, 79%, 0.79 mmol) (mp 79−82 °C), and a small
amount of a vinyl product was also formed. Compound 5da was also
obtained with [RuCl2(p-cymene)]2 as catalyst with AgPF6 (25 mg, 0.1
mmol) as additive (yield 40%, based on NMR integration against an
internal standard). A vinyl product was also formed.
KIE Experiment with Rh. The same general procedure as above
was followed using 5-methyl-3-phenyl-1H-pyrazole (1a, 52 mg, 0.33
mmol) and d3-1a (53 mg, 0.33 mmol) with 4-octyne (2a, 44 mg, 0.39
mmol, 1.2 equiv), Cu(OAc)2·H2O (165 mg, 0.83 mmol, 2.5 equiv),
and the catalyst (5 mol % [Rh(MeCN)3Cp*][PF6]2) 1,3,5-
trimethoxybenzene (3.0 mg, 0.017 mmol, 5 mol %) as internal
standard, and DCE (reaction made up to 5 mL volume). Each reaction
was done twice. The H NMR spectra were recorded, and yield was
found by integrating the CH2 2H peak of the product. After rates for
each reaction were calculated, the kinetic isotope effects were
1
calculated as kH/kD = 2.7
0.5 for Rh. A similar procedure was
followed for Ru and kH/kD = 1.1 0.2 for Ru.
KIE Experiment with Ru. The general procedure for the initial
rates experiment was followed using 5-methyl-3-phenyl-1H-pyrazole
(1a, 32 mg, 0.20 mmol) and d3-1a (32 mg, 0.20 mmol) with 4-octyne
(2a, 26 mg, 0.24 mmol, 1.2 equiv), Cu(OAc)2·H2O (100 mg, 0.50
mmol, 2.5 equiv), 1,3,5-trimethoxybenzene (1.7 mg, 0.01 mmol, 5 mol
%), and DCE (reaction made up to 3 mL volume). The vials were
sealed with a screw cap and transferred to a preheated heating block at
83 °C to stir for 5 min. [RuCl2(p-cymene)]2 (6.1 mg, 0.01 mmol, 5
mol %) was added, and timing was started. All reactions were left on
for 60 min. Each reaction had a repeat reaction within the same
1
5da. H NMR (400 MHz, CDCl3): δ 1.06 (t, J = 7.4 Hz, 3H, H5),
1.07 (t, J = 7.0, 7.4 Hz, 3H, H8), 1.65−1.74 (m, 2H, H4), 1.76−1.85
(m, 2H, H7), 2.89−2.93 (m, 2H, H3), 3.22−3.26 (m, 2H, H6), 6.90
(s,1H, Pz-H), 7.38 (d, J = 5.5 Hz, 1H, H2), 7.52 (d, J = 5.1 Hz, 1H,
H1). 13C {1H} NMR (100 MHz, CDCl3): δ 14.2 (C5/C8), 14.3 (C5/
C8), 20.8 (C7), 24.0 (C4), 29.4 (C6), 31.6 (C3), 93.8 (Pz), 120.5, 121.8
(q, J = 270.4 Hz,CF3), 123.2 (C2), 125.9, 126.7 (C1), 135.7, 135.9,
136.6, 142.8 (q, J = 39.9 Hz,C-CF3). 19F {1H} NMR (376 MHz,
CDCl3): δ −61.2 (CF3). ESIMS: m/z 327 [M + H]+. HRMS (ES):
calcd for C16H18N2F3S1 [M + H]+327.1143, found 327.1158.The
product was recrystallized from dichloromethane/hexane to give 5da
as clear needles.
1
heating block in order to calculate averages. The H NMR spectra
were recorded and yield was found by integrating the CH2 2H peak of
the product. This led to a kH/kD = 1.1 0.2.
Synthesis of 5db. Following the general procedure, a Schlenk
flask was loaded with [Rh(MeCN)3Cp*][PF6]2 (16 mg, 5 mol %), 3-
(thiophene-2-yl)-5-(trifluoromethyl)-1H-pyrazole (222 mg, 1.0
mmol), Cu(OAc)2·H2O (250 mg, 1.25 mmol), diphenylacetylene
(2b, 214 mg, 1.2 mmol), and DCE (10 mL). The product was purified
by washing with hexane to give 5db as an orange powder (348 mg,
88%, 0.88 mmol) (mp 230−232 °C), and a vinyl byproduct was also
formed. 1H NMR (500 MHz, CDCl3): δ 7.04 (s, 1H, Pz-H), 7.09 (d, J
= 5.1 Hz, 1H, H3), 7.19 (dd, J = 2.7, 7.0 Hz, 2H, H3), 7.24−7.30 (m,
6H, H4, H5, H7, H8), 7.35 (dd, J = 2.7, 7.0 Hz, 2H, H6), 7.48 (d, J = 5.5
Hz, 1H, H1). 13C {1H} NMR (125 MHz, CDCl3): δ 94.2 (Pz), 121.5
(q, J = 270.1 Hz, CF3), 123.9, 125.3 (C2), 126.8 (C1), 127.0, 127.5
(C5/C8), 127.9 (C4/C7), 128.2 (C4/C7), 128.7 (C5/C8), 130.7 (C3),
131.5 (C6), 131.8, 135.4, 136.5, 136.6, 136.9, 143.8 (q, J = 39.7 Hz,
CF3), 19F {1H} NMR (376 MHz, CDCl3): δ −61.3 (CF3). ESIMS: m/
z 395 [M + H]+. HRMS (ES): calcd for C22H14N2F3S1 [M +
H]+395.0830, found 395.0839.
General Procedure for Deuteration of Pyrazoles. A Schlenk
flask equipped with a stirrer bar was evacuated and backfilled with
nitrogen. It was loaded with the relevant pyrazole (1.0 mmol), the
relevant metal catalyst (0.05 equiv of metal), NaOAc (164 mg, 2.0
mmol, 2 equiv), and CD3OD (2.5 mL). The flask was sealed with a
screw cap and transferred to a preheated oil bath at 60 °C to stir
overnight. Monitoring by 1H NMR showed some deuteration had
occurred. PivOD (16 mg, 0.2 mmol, 0.2 equiv) was added, and the
reaction was stirred at 60 °C with further monitoring. Changing the
CD3OD on one or two occasions led to high (>97%) deuterium
incorporation.
Deuteration in the Presence of Alkyne. Following the general
procedures for catalysis reactions (see above), a Schlenk flask was
loaded with [Rh(MeCN)3Cp*][PF6]2 (16 mg, 5 mol %), 3-phenyl-5-
methyl-1H-pyrazole (1a, 79 mg, 0.50 mmol), Cu(OAc)2·H2O (250
mg, 1.25 mmol), 4-octyne (66 mg, 0.60 mmol), and CD3OD (3 mL)
and transferred to a preheated oil bath at 70 °C. After 16 h, this gave
1968
dx.doi.org/10.1021/jo402592z | J. Org. Chem. 2014, 79, 1954−1970

The Journal of Organic Chemistry
Article
d2-3aa as a brown solid (124 mg, 93%, 0.47 mmol). Reaction on the
same scale with the catalyst [RuCl2(p-cymene)]2 (15 mg, 0.025 mmol,
2.5 mol %) gave after 16 h d1-3aa as a brown solid (130 mg, 97%, 0.49
mmol).
optimized geometries were reoptimised with Gaussian 09 using the
above basis set combinations, with frequency calculations again used to
confirm the nature of all stationary points. Single-point dispersion
corrections to the BP86 results employed Grimme’s D3 parameter set
as implemented in Gaussian 09.48
X-ray Crystal Structure Determination of 3ac, 3dc, 3da, 5da,
and 3db. Data were collected on a Bruker Apex 2000 CCD
diffractometer using graphite-monochromated Mo Kα radiation, λ =
0.7107 Å. The data were corrected for Lorentz and polarization effects,
and empirical absorption corrections were applied. The structure was
solved by direct methods and with structure refinement on F2
employed in SHELXTL version 6.10.39 Hydrogen atoms were
included in calculated positions (C−H = 0.93−1.00 Å, O−H = 0.84
Å) riding on the bonded atom with isotropic displacement parameters
set to 1.5Ueq (O) for hydroxyl H atoms, 1.5Ueq (C) for methyl
hydrogen atoms and 1.2Ueq (C) for all other H atoms. All non-
hydrogen atoms were refined with anisotropic displacement
parameters without positional restraints. Disordered solvent was
removed with the Squeeze option in PLATON.40 Figures were drawn
using the program ORTEP.41 Coordinates have been deposited with
the Cambridge crystallographic database CCDC nos. CCDC950944−
950948.
ASSOCIATED CONTENT
* Supporting Information
■
S
Crystallographic data (CIF), NMR spectra, details of kinetics
and KIE measurements, and details of all computed structures
and associated energies, as well as functional testing and full
references 42 and 43. This material is available free of charge via
the Internet at http://pubs.acs.org. CCDC 950944-950948
contain complete crystallographic data for this paper. These
data can be obtained free of charge from The Cambridge
Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_
request/cif.
AUTHOR INFORMATION
Corresponding Authors
*E-mail: warren.cross@ntu.ac.uk.
*E-mail: dld3@le.ac.uk.
*E-mail: s.a.macgregor@hw.ac.uk.
■
Crystal Data for the Complexes. 3ac: C19H19N2O1.5, M =
299.36, tetragonal, a = 22.783(3) Å, b = 22.783(3) Å, c = 5.7864(11)
Å, α = β = γ = 90°, V = 3003.5(8) Å3, T = 150(2) K, space group I4, Z
̅
= 8, 10897 reflections measured, 2623 independent reflections (Rint
=
0.0682). The final R1 values were 0.0597 (I > 2σ(I)), 0.0699 (all data).
The final wR(F2) values were 0.1485 (I > 2σ(I)), 0.1522 (all data).
GOF = 1.103.
3da: C18H19F3N2, M = 320.35, triclinic, a = 16.927(16) Å, b =
4.693(5) Å, c = 21.28(2) Å, α = 90°, β = 110.54(2)°, γ = 90°, V =
1583(3) Å3, T = 150(2) K, space group P2(1)/n, Z = 4, 10780
reflections measured, 2788 independent reflections (Rint = 0.2392).
The final R1 values were 0.0722 (I > 2σ(I)) 0.2318 (all data). The
final wR(F2) values were 0.1216 (I > 2σ(I)), 0.1655 (all data). GOF =
0.782.
Present Address
§School of Science and Technology, Nottingham Trent
University, Clifton campus, Nottingham NG11 8NS, U.K.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported through EPSRC awards EP/J021709/
1, EP/J002917/1 (D.L.D.), and EP/J021911/1 (S.A.M.,
C.L.M.). We also thank the University of Leicester for funding
(Q.K.), the Spanish Government for a Postdoctoral Fellowship
(EX2009-0398, A.G.A.), and Johnson Matthey for a loan of
rhodium trichloride.
3db: C24H15F3N2·C2H3N, M = 429.43, triclinic, a = 9.5265(15) Å, b
= 9.5816(15) Å, c = 12.880(2) Å, α = 70.575(3)°, β = 70.337(4)°, γ =
84.964(3)°, V = 1043.7(3) Å3, T = 150(2) K, space group P-1, Z = 2,
7678 reflections measured, 3671 independent reflections (Rint
=
0.0990). The final R1 values were 0.0723 (I > 2σ(I)) 0.1219 (all data).
The final wR(F2) values were 0.1530 (I > 2σ(I)), 0.1780 (all data).
GOF = 0.896.
3dc: C19H13F3N2, M = 326.31, monoclinic, a = 10.826(4) Å, b =
16.541(7) Å, c = 17.451(7) Å, α = 90°, β = 96.945(9)°, γ = 90°, V =
3102(2) Å3, T = 150(2) K, space group P2(1)/n, Z = 8, 22238
reflections measured, 5462 independent reflections (Rint = 0.1020).
The final R1 values were 0.0511 (I > 2σ(I)) 0.0991 (all data). The
final wR(F2) values were 0.0896 (I > 2σ(I)), 0.1027 (all data). GOF =
0.877.
REFERENCES
■
(1) (a) Alberico, D.; Scott, M. E.; Lautens, M. Chem. Rev. 2007, 107,
174. (b) Lyons, T. W.; Sanford, M. S. Chem. Rev. 2010, 110, 1147.
(c) Chen, X.; Engle, K. M.; Wang, D.-H.; Yu, J.-Q. Angew. Chem., Int.
Ed. 2009, 48, 5094. (d) Colby, D. A.; Bergman, R. G.; Ellman, J. A.
Chem. Rev. 2009, 110, 624. (e) Bellina, F.; Rossi, R. Chem. Rev. 2009,
110, 1082. (f) Ackermann, L. Chem. Rev. 2011, 111, 1315.
5da: C16H17F3N2S, M = 326.38, triclinic, a = 8.548(2) Å, b =
9.543(2) Å, c = 19.894(5) Å, α = 88.270(5)°, β = 81.218(6)°, γ =
82.774(5)°, V = 1591.0(7) Å3, T = 150(2) K, space group P-1, Z = 4,
(2) (a) Davies, D. L.; Donald, S. M. A.; Macgregor, S. A. J. Am. Chem.
Soc. 2005, 127, 13754. (b) Davies, D. L.; Donald, S. M. A.; Al-Duaij,
O.; Fawcett, J.; Little, C.; Macgregor, S. A. Organometallics 2006, 25,
5976. (c) Davies, D. L.; Donald, S. M. A.; Al-Duaij, O.; Macgregor, S.
A.; Polleth, M. J. Am. Chem. Soc. 2006, 128, 4210. (d) Boutadla, Y.;
Davies, D. L.; Macgregor, S. A.; Poblador-Bahamonde, A. I. Dalton
Trans. 2009, 5820. (e) Boutadla, Y.; Davies, D. L.; Macgregor, S. A.;
Poblador-Bahamonde, A. I. Dalton Trans. 2009, 5887.
(3) (a) Musashi, Y.; Sakaki, S. J. Am. Chem. Soc. 2000, 122, 3867.
(b) Garcia-Cuadrado, D.; Braga, A. A. C.; Maseras, F.; Echavarren, A.
M. J. Am. Chem. Soc. 2006, 128, 1066. (c) Lafrance, M.; Fagnou, K. J.
Am. Chem. Soc. 2006, 128, 16496. (d) Ackermann, L.; Vicente, R.;
Althammer, A. Org. Lett. 2008, 10, 2299. (e) Lafrance, M.; Rowley, C.
N.; Woo, T. K.; Fagnou, K. J. Am. Chem. Soc. 2006, 128, 8754.
(f) Gorelsky, S. I.; Lapointe, D.; Fagnou, K. J. Am. Chem. Soc. 2008,
130, 10848.
11669 reflections measured, 5558 independent reflections (Rint
=
0.0973). The final R1 values were 0.0645 (I > 2σ(I)) 0.1188 (all data).
The final wR(F2) values were 0.1217 (I > 2σ(I)), 0.1436 (all data).
GOF = 0.896.
COMPUTATIONAL DETAILS
■
DFT calculations were run with Gaussian 03 (Revision D.01)42 and
Gaussian 09 (Revision A.02).43 Rh and Ru centers were described with
the Stuttgart RECPs and associated basis sets44, and 6-31G** basis
sets were used for all other atoms.45 Initial BP8646 optimizations were
performed with Gaussian 03 using the ‘grid=ultrafine’ option, with all
stationary points being fully characterized via analytical frequency
calculations as either minima (all positive eigenvalues) or transition
states (one negative eigenvalue). IRC calculations and subsequent
geometry optimizations were used to confirm the minima linked by
each transition state. Corrections for the effects of dichloroethane
solvent (ε = 10.125) were run with Gaussian 09 and used the
polarizable continuum model.47 For the functional testing, the BP86-
(4) (a) Ueura, K.; Satoh, T.; Miura, M. Org. Lett. 2007, 9, 1407.
(b) Ueura, K.; Satoh, T.; Miura, M. J. Org. Chem. 2007, 72, 5362.
(c) Shimizu, M.; Hirano, K.; Satoh, T.; Miura, M. J. Org. Chem. 2009,
74, 3478. (d) Mochida, S.; Hirano, K.; Satoh, T.; Miura, M. J. Org.
Chem. 2009, 74, 6295. (e) Fukutani, T.; Umeda, N.; Hirano, K.; Satoh,
1969
dx.doi.org/10.1021/jo402592z | J. Org. Chem. 2014, 79, 1954−1970

The Journal of Organic Chemistry
Article
T.; Miura, M. Chem. Commun. 2009, 5141. (f) Umeda, N.; Hirano, K.;
Satoh, T.; Miura, M. J. Org. Chem. 2009, 74, 7094. (g) Umeda, N.;
Tsurugi, H.; Satoh, T.; Miura, M. Angew. Chem., Int. Ed. 2008, 47,
4019.
(5) (a) Stuart, D. R.; Bertrand-Laperle, M.; Burgess, K. M. N.;
Fagnou, K. J. Am. Chem. Soc. 2008, 130, 16474. (b) Guimond, N.;
Fagnou, K. J. Am. Chem. Soc. 2009, 131, 12050.
(6) Song, G.; Wang, F.; Li, X. Chem. Soc. Rev. 2012, 41, 3651.
(7) Arockiam, P. B.; Bruneau, C.; Dixneuf, P. H. Chem. Rev. 2012,
112, 5879.
(8) (a) Li, L.; Brennessel, W. W.; Jones, W. D. J. Am. Chem. Soc.
2008, 130, 12414. (b) Li, L.; Brennessel, W. W.; Jones, W. D.
Organometallics 2009, 28, 3492. (c) Sortais, J. B.; Pannetier, N.;
Holuigue, A.; Barloy, L.; Sirlin, C.; Pfeffer, M.; Kyritsakas, N.
Organometallics 2007, 26, 1856. (d) Boutadla, Y.; Al-Duaij, O.;
Davies, D. L.; Griffith, G. A.; Singh, K. Organometallics 2009, 28, 433.
(e) Boutadla, Y.; Davies, D. L.; Al-Duaij, O.; Fawcett, J.; Jones, R. C.;
Singh, K. Dalton Trans. 2010, 39, 10447. (f) Davies, D. L.; Al-Duaij,
O.; Fawcett, J.; Giardiello, M.; Hilton, S. T.; Russell, D. R. Dalton
Trans. 2003, 4132.
(9) (a) Li, L.; Jiao, Y.; Brennessel, W. W.; Jones, W. D.
Organometallics 2010, 29, 4593. (b) Han, Y.-F.; Li, H.; Hu, P.; Jin,
G.-X. Organometallics 2011, 30, 905.
(10) (a) Li, B.; Feng, H. L.; Wang, N. A. C.; Ma, J. F.; Song, H. B.;
Xu, S. S.; Wang, B. Q. Chem.−Eur. J. 2012, 18, 12873. (b) Wang, N.;
Li, B.; Song, H.; Xu, S.; Wang, B. Chem.−Eur. J. 2013, 19, 358.
(11) Simmons, E. M.; Hartwig, J. F. Angew. Chem., Int. Ed. 2012, 51,
3066.
(28) (a) Ahlquist, M. S. G.; Norrby, P. O. Angew. Chem., Int. Ed.
2011, 50, 11794. (b) Minenkov, Y.; Occhipinti, G.; Jensen, V. R. J.
Phys. Chem. A 2009, 113, 11833. (c) Ryde, U.; Mata, R. A.; Grimme, S.
Dalton Trans. 2011, 40, 11176. (d) Sieffert, N.; Buhl, M. Inorg. Chem.
̈
2009, 48, 4622. (e) Zhao, Y.; Truhlar, D. G. Org. Lett. 2007, 9, 1967.
(29) The alternative possibility of insertion into the Rh−N bond
entails a TS at +22.0 kcal/mol, 8 kcal/mol above that for insertion into
the Rh−C bond, and therefore is highly unlikely.
(30) This conclusion is not affected by the choice of solvent used in
the calculations: in MeOH (the solvent used for the experimental H/D
exchange studies), the computed energies of TS(B−C)2 and TS(D−
E) are +10.6 kcal/mol and +11.0 kcal./mol
(31) (a) Bennett, M. A.; Macgregor, S. A.; Wenger, E. Helv. Chim.
Acta 2001, 84, 3084. (b) Edwards, A. J.; Macgregor, S. A.; Rae, A. D.;
Wenger, E.; Willis, A. C. Organometallics 2001, 20, 2864.
(c) Macgregor, S. A.; Wenger, E. Organometallics 2002, 21, 1278.
(32) Stationary points corresponding to INT(A-B) and TS(B-C)2
were located on the electronic energy surface but the free energy of
TS(B-C)2 falls below that of INT(A-B) suggesting C−H activation
proceeds directly from B to C1 via TS(B-C)1.
(33) We cannot rule out the possibility that catalyst reoxidation is
rate limiting as this would also be consistent with a KIE of 1.1. We
note, however, that the computed barriers based on rate limiting C−H
activation provide barriers that are consistent with the experimental
conditions and reproduce the lower catalytic activity of the Ru system.
We also repeated the KIE measurements starting with preformed
[Ru(OAc)2(p-cymene)] and obtained the same result as before, i.e., a
kH/kD KIE of ca. 1.
(34) White, C.; Thompson, S. J.; Maitlis, P. M. J. Chem. Soc., Dalton
Trans. 1977, 1654.
(12) A DFT study of Rh(III)-catalyzed oxidative Heck coupling of
phenol carbamates with alkenes has recently appeared; see: Zhang, Q.;
Yu, H.-Z.; Li, Y.-T.; Liu, L.; Huang, Y.; Fu, Y. Dalton Trans. 2013, 42,
4175. Very recently, a computational study of the intramolecular
annulation of benzamides with tethered alkynes has been reported;
(35) Pleier, A. K.; Glas, H.; Grosche, M.; Sirsch, P.; Thiel, W. R.
Synthesis 2001, 55.
(36) Texierboullet, F.; Klein, B.; Hamelin, J. Synthesis 1986, 409.
(37) Guillou, S.; Bonhomme, F. J.; Ermolenko, M. S.; Janin, Y. L.
Tetrahedron 2011, 67, 8451.
(38) Li, J.-H.; Liang, Y.; Xie, Y.-X. J. Org. Chem. 2005, 70, 4393.
(39) Bruker; Version 6.10 ed.; Bruker Inc., Madison, WI, 1998−2000.
(40) Vandersluis, P.; Spek, A. L. Acta Crystallogr., Sect. A 1990, 46,
194.
(41) Farrugia, L. J. J. Appl. Crystallogr. 1997, 30, 565.
(42) Frisch, M. J. Gaussian, Inc., Wallingford, CT, 2004.
(43) Frisch, M. J. Gaussian, Inc., Wallingford, CT, 2009.
see: Quinones, N.; Seoane, A.; García-Fandino, R.; Mascarenas, J. L.;
̃
̃
̃
Gulías, M. Chem. Sci. 2013, 4, 2874.
(13) Guimond, N.; Gorelsky, S. I.; Fagnou, K. J. Am. Chem. Soc.
2011, 133, 6449.
(14) Xu, L.; Zhu, Q.; Huang, G.; Cheng, B.; Xia, Y. J. Org. Chem.
2012, 77, 3017.
(15) Umeda, N.; Hirano, K.; Satoh, T.; Shibata, N.; Sato, H.; Miura,
M. J. Org. Chem. 2011, 76, 13.
(44) Andrae, D.; Haußermann, U.; Dolg, M.; Stoll, H.; Preuß, H.
Theor. Chim. Acta 1990, 77, 123.
(45) (a) Hariharan, P. C.; Pople, J. A. Theor. Chim. Acta 1973, 28,
213. (b) Hehre, W. J.; Ditchfield, R.; Pople, J. A. J. Chem. Phys. 1972,
56, 2257.
̈
(16) Li, X.; Zhao, M. J. Org. Chem. 2011, 76, 8530.
(17) Ma, W.; Graczyk, K.; Ackermann, L. Org. Lett. 2012, 14, 6318.
(18) See for examples refs 4−7, 16 and 17 and references cited
therein.
(19) In this reaction there was a lot of unreacted 1b, the yield based
(46) (a) Becke, A. D. Phys. Rev. A 1988, 38, 3098. (b) Perdew, J. P.
Phys. Rev. B 1986, 33, 8822.
(47) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. Rev. 2005, 105,
2999.
on 1b consumed is 98%.
(20) Clar, E. Polycyclic Hydrocarbons; Academic Press: London, 1964.
(21) Morimoto, K.; Hirano, K.; Satoh, T.; Miura, M. Org. Lett. 2010,
12, 2068.
(48) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. J. Chem. Phys.
(22) Wang, H.; Glorius, F. Angew. Chem., Int. Ed. 2012, 51, 7318.
(23) A similar mechanism has been proposed by Li; see ref 16.
(24) Note that in similar studies (ref 17) with nitrophenyl-
substituted pyrazoles Ackermann found that CH activation was
reversible and that deuterium exchange occurred even in the presence
of diphenylacetylene, suggesting that the relative barriers to CH
activation and alkyne insertion are substrate dependent as expected.
(25) This value is likely to arise from a combination of a kinetic
isotope effect arising from the proportion of the substrate that reacts
straight through to the coupled products after C−H bond activation
and an equilibrium isotope effect due to the proportion of C−H
activated intermediates that undergo the reverse reprotonation.
(26) All optimizations aiming to locate an equivalent planar form of
E reverted back to boat structures such as that shown in Figure 4.
(27) Based on the free energy change associated with the reaction of
5-phenyl-3-methylpyrazole + RCCH (R = H, nPr) to give the
corresponding pyrazoloisoquinoline products + H2. ΔGDCE = −43.3
2010, 132.
n
kcal/mol (R = H); ΔGDCE = −28.0 kcal/mol (R = Pr).
1970
dx.doi.org/10.1021/jo402592z | J. Org. Chem. 2014, 79, 1954−1970