Ligand bite angle-dependent palladium-catalyzed cyclization of propargylic carbonates to 2-alkynyl azacycles or cyclic dienamides

Article abstract of DOI:10.1002/anie.201309162

The regioselectivity of the palladium-catalyzed cyclization of propargylic carbonates with sulfonamide nucleophiles is critically dependent on the bite angle of the bidentate phosphine ligand. Ligands with small bite angles favor attack on the central carbon atom of an allenylpalladium intermediate to afford cyclic dienamide products, whereas the use of those with large bite angles leads to alkynyl azacycles, with high stereoselectivity. A computational analysis of the reaction pathway is also presented. Chomp! The bite angle of bidentate phosphine ligands determined the course of the palladium-catalyzed cyclization of propargylic carbonates with sulfonamide nucleophiles. A small bite angle favored attack on the central C atom of the allenylpalladium intermediate, whereas alkynyl azacycles were formed from attack on the terminal C atom using ligands with a large bite angle. Copyright

Full text of DOI:10.1002/anie.201309162

Angewandte
Chemie
DOI: 10.1002/anie.201309162
Nitrogen Heterocycles
Ligand Bite Angle-Dependent Palladium-Catalyzed Cyclization of
Propargylic Carbonates to 2-Alkynyl Azacycles or Cyclic
Dienamides**
David S. B. Daniels, Alison S. Jones, Amber L. Thompson, Robert S. Paton,* and
Edward A. Anderson*
Abstract: The regioselectivity of the palladium-catalyzed
cyclization of propargylic carbonates with sulfonamide nucle-
ophiles is critically dependent on the bite angle of the bidentate
phosphine ligand. Ligands with small bite angles favor attack
on the central carbon atom of an allenylpalladium intermediate
to afford cyclic dienamide products, whereas the use of those
with large bite angles leads to alkynyl azacycles, with high
stereoselectivity. A computational analysis of the reaction
pathway is also presented.
the reaction, but also a stereoselective entry to alkynyl
azacycles, which are important synthetic targets.[5] We report
herein the results of these studies, including our finding that
the regioselectivity of the cyclization can be tuned through
the choice of ligand to give either alkynyl azacycles or cyclic
dienamides, and computational studies that correlate this
reactivity dichotomy with the structures of the allenylpalla-
dium intermediates.
Investigations began with sulfonamide carbonate 4a,
which cyclized to pyrrolidine 5a with excellent stereoselec-
tivity under the conditions we had optimized for oxygen
nucleophiles (Table 1, entry 1).[3,6,7] However, in contrast to
our previous study, a significant amount of cyclic enamide 6a
was formed through competitive attack by the sulfonamide on
the central carbon atom of the intermediate allenylpalladium
species 3 (see mechanistic discussion below).[8,9]
Stereoselective heterocycle synthesis is of fundamental
importance in pharmaceutical and natural product
research.[1,2] We recently reported a palladium-catalyzed
asymmetric approach to 2-alkynyl oxacycles from propargylic
carbonates equipped with internal alcohol nucleophiles (1!
2, Nu = OH; Scheme 1).[3] Key to this reaction is the high
The ratio of these products was insensitive to variation of
the temperature or solvent (Table 1, entries 2–5), but highly
dependent on the nature of the bidentate phosphine ligand
(Table 1, entries 5–10). A dramatic switch in selectivity was
observed between dppe (5a/6a 19:81; Table 1, entry 6) and
DPEphos (5a/6a 89:11; entry 10),[10] with a modest increase in
stereoselectivity observed for the latter in 1,4-dioxane
(97% ee; entry 11). The role of additives, which had benefi-
cial effects on regioselectivity in our related oxacycle study,
was next examined.[3] The use of boric acid or B(OMe)3
entirely prevented the formation of 6a, but required higher
temperatures for full conversion and led to decreased
stereoselectivity (Table 1, entries 12 and 13); lowering of the
temperature improved matters, but the reaction under these
conditions was capricious on larger scales (Table 1, entry 14).
The nature of the nitrogen protecting group proved impor-
tant: whereas other sulfonamides (compounds 4b–d) were
viable substrates, albeit with slightly reduced selectivity,
amide 4e and the free amine 4 f did not cyclize (Table 1,
entries 15–19). The conclusion of this optimization was that
catalyst systems that favored the formation of either the
pyrrolidine (DPEphos) or the dienamide (dppe) had been
identified, and although dienamide formation could be
prevented through the addition of boronate additives, the
loss of enantioselectivity was viewed as too high a price to pay.
Subsequent reactions were therefore conducted in dioxane at
1008C (conditions that ensured rapid conversion) in the
absence of additives.
Scheme 1. Cyclization of propargylic carbonates to alkynyl heterocycles.
fidelity of stereochemical transfer from the carbonate to the
oxacycle, and the high regioselectivity of attack by the oxygen
nucleophile on the distal carbon atom of the intermediate
allenylpalladium complex 3, as opposed to the central carbon
atom as is usually observed in such systems.[4] We were
interested in the dependence of these selectivities on the
nature of the nucleophile and elected to explore nitrogen
nucleophiles, which might not only offer further insight into
[*] D. S. B. Daniels, A. S. Jones, Dr. A. L. Thompson, Dr. R. S. Paton,
Dr. E. A. Anderson
Chemistry Research Laboratory, University of Oxford
12 Mansfield Road, Oxford, OX1 3TA (U.K.)
E-mail: edward.anderson@chem.ox.ac.uk
Homepage: http://anderson.chem.ox.ac.uk/
[**] We thank the EPSRC (EP/E055273/1, Advanced Research Fellow-
ship to E.A.A., studentship to D.S.B.D.).
We prepared a range of mono- and disubstituted sulfon-
amide carbonates to examine the reaction scope and selec-
tivity with each of these ligands. Beginning with DPEphos
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201309162.
Angew. Chem. Int. Ed. 2014, 53, 1915 –1920
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1915

.
Angewandte
Communications
Table 1: Optimization of the azacyclization of carbonates 4a–f.[a]
quatorial disposition of the allyl
side chain reinforces the stereoelec-
tronic requirement for anti-SN2’
reductive elimination. The contrast-
ing cyclization of 4k necessitates an
axial orientation of the allyl sub-
stituent in 7k, which evidently
results in lower regioselectivity. An
equivalent trend is seen for piper-
idines 5l and 5m, albeit with less
disparity between the two diaste-
reomers, probably owing to a less
“tight” transition state with reduced
conformational influence by the
allyl substituent.
Having established a stereose-
lective cyclization to alkynyl aza-
cycles with DPEphos as the ligand,
attention turned to the complemen-
tary regioselective cyclization with
dppe. This investigation again
began with carbonate 4g (Table 2,
entry 8), which now cyclized to
enamide 6g with good regioselec-
tivity (16:84) and excellent stereo-
selectivity for the formation of
a single alkene isomer (> 95:5).
We were particularly pleased to
reverse the regioselectivity of the
cyclization of 4h, for which the
dppe system conferred excellent
regioselectivity for the formation
of seven-membered enamide 6h
Entry Substrate Solvent
T [8C] Ligand
Additive
5/6[b]
Yield [%][c] ee [%][d]
(0.5 equiv)
1
2
3
4
5
4a
4a
4a
4a
4a
4a
4a
4a
4a
4a
4a
4a
4a
4a
4b
4c
4d
4e
4 f
1,4-dioxane 100
dppf
dppf
dppf
dppf
dppf
dppe
dppp
dppb
–
–
–
–
-
–
–
–
–
–
–
71:29 63
71:29 n.d.[e]
70:30 n.d.
71:29 n.d.
71:29 48
19:81 60
32:68 n.d.
48:52 n.d.
69:31 n.d.
89:11 78
89:11 77
100:0
100:0
100:0
87:13 72
74:26 55
86:14 73
96
94
95
95
95
95
95
96
95
94
97
90
92
97
n.d.[e]
n.d.
n.d.
–
n-butanol
DME
100
85
toluene
toluene
100
60
6[f]
7
8
toluene
60
toluene
60
toluene
60
9[f]
10[g]
11
12
13
14[f]
15
16
17
18
19
toluene
60
dpppe
DPEphos
DPEphos
DPEphos B(OH)3
DPEphos B(OMe)3
DPEphos B(OMe)3
DPEphos
DPEphos
DPEphos
DPEphos
DPEphos
toluene
60
1,4-dioxane
1,4-dioxane
1,4-dioxane
1,4-dioxane
toluene
toluene
toluene
toluene
toluene
60
80
80
60
100
100
100
100
100
86
83
62[h]
–
–
–
–
–
–
–
n.r.[i]
n.r.
–
[a] Reactions were conducted with 4 (97% ee; 25 mmol), [Pd(dba)2] (5 mol%), and the ligand (10 mol%)
at a concentration of 0.1m for 15 min, unless stated otherwise. [b] The product ratio was determined by
1H NMR spectroscopic analysis of the crude reaction mixture. [c] Yield of the major product after
isolation. [d] The ee value was determined by HPLC on a chiral stationary phase (see the Supporting
Information for details). [e] n.d.=not determined. [f] The reaction was conducted for 30 min. [g] The
reaction was conducted for 45 min. [h] The reaction proceeded to 70% conversion. [i] n.r. =no reaction.
dba=dibenzylideneacetone; Ts=p-toluenesulfonyl; Ms=methanesulfonyl; 4-Ns=p-nitrobenzenesul-
fonyl; PMP=p-methoxyphenyl; DME=1,2-dimethoxyethane; dppf=1,1’-bis(diphenylphosphanyl)fer-
rocene; dppe, dppp, dppb, dpppe=1,n-bis(diphenylphosphanyl)alkane ligand (ethane, n=2; propane, (Table 2, entry 9), in spite of a pre-
n=3; butane, n=4; pentane, n=5); DPEphos=(oxybis(2,1-phenylene))bis(diphenylphosphane).
sumed thermodynamic preference
(Table 2, entries 1–7), we found that the alkyl alkyne 4g
cyclized equally efficiently as the aryl alkyne 4a to deliver
pyrrolidine 5g in excellent yield and regioselectivity (81%,
97% ee; Table 2, entry 1).[11] The cyclization could be applied
to the synthesis of alkynyl piperidines, with 5h produced in
good yield and excellent enantioselectivity (Table 2, entry 2);
however, the formation of azepane 5i was less successful, and
Scheme 2. Proposed conformations of allenylpalladium complexes 7j
to our surprise was accompanied by the formation of
and 7k in the cyclization of 4j and 4k to pyrrolidines 5j and 5k,
respectively.
significant amounts of the corresponding eight-membered
cyclic dienamide (entry 3).[12] The cyclization of enantiomer-
ically pure disubstituted sulfonamide carbonates 4j–m pro-
ceeded in good yield and diastereoselectivity in all cases
(Table 2, entries 4–7), thus enabling efficient access to both
cis- and trans-disubstituted alkynyl pyrrolidines and piper-
idines. The regioselectivity of the reactions was consistently
high,[7] with that of the cyclization of 4j (Table 2, entry 4) to
the trans-disubstituted pyrrolidine 5j being particularly
notable (single regioisomer, d.r. > 99:1). The exceptional
efficiency of this cyclization (as opposed to that of 4k;
Table 2, entry 5) can be rationalized by the possible con-
formations of the allenylpalladium(II) intermediates 7j and
7k (Scheme 2). In the case of 7j (arising from 4j), a pseudoe-
for piperidine synthesis. This success could not be maintained
for the formation of the eight-membered dienamide (Table 2,
entry 10); instead, only the elimination product 8 was formed.
This result underlines the complete disfavoring of cyclization
at the distal allene carbon atom with the ligand dppe. The
reactions of the disubstituted sulfonamide carbonates 4j–m
(Table 2, entries 11–14) afforded interesting results. In this
case, the powerful kinetic propensity for cyclization to a five-
membered ring overrode the influence of the ligand, and for
the configurationally matched substrate 4j resulted in the
formation of pyrrolidine 5j as the major product. However,
1916
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1915 –1920

Angewandte
Chemie
Table 2: Cyclization of sulfonamide carbonates 4 to alkynyl azacycles 5 and cyclic dienamides 6.[a]
for substrates 4l and 4m, the re-
duced driving force for six-mem-
bered-ring formation now restored
high regioselectivity for the forma-
tion of azepane enamides 6l and
6m, although the latter reaction
only proceeded with partial conver-
sion, even with extended reaction
times.
The contrasting outcomes of
these reactions can be explained
by a mechanism that begins with
anti-SN2’ oxidative addition[13] by
Entry Substrate
Major product
Yield [%][b] ee [%]/d.r. of
E/Z ratio of
5/6[d]
5[c]
6[d]
1[e]
81[f]
61
97
98
90:10
85:15
66:33[g]
90:10
80:20
75:25
2
the
bidentate-phosphine-com-
plexed palladium(0) catalyst to
give an h1-allenylpalladium species
9 (Scheme 3). Through ligand dis-
sociation, the metal is able to bind
to the remote double bond of the
allene to form a nonlinear cationic
h3-allenylpalladium species 10.[14]
Attack on the central allene
carbon atom of 10 (path a) leads
to a p-allylpalladium intermediate
11, which can equilibrate with 12
through a p–s–p interconversion
pathway. b-Hydride elimination
from intermediate 12, which is
probably favored as a result of
lower 1,3-allylic strain, leads to the
(E)-enamide 6a. With DPEphos,
path b is favored—namely, attack
at the terminal (distal) carbon atom
of the allenyl palladium system to
give pyrrolidine 5a through anti-
SN2’ reductive elimination.[15]
3
26
n.d.
4
5
6
76
86
>99:1
–
100:0
73:27
95:5
>95:5
68
60
96:4
97:3
80:20
75:25
77:23
85:15
7
8[e] 4g
61[f]
67
n.d.
n.d.
–
>95:5
>95:5
1:1[h]
16:84
18:82
–
9
10
11
4h
4i
To rationalize this dramatic
78
switch in regioselectivity according
to the choice of ligand,[16] we mod-
eled the structures of the cationic h3
complexes by carrying out quan-
tum-chemical calculations. First, the
structure of the parent complex
[Pd(PH3)2] bound to the allenyl
4j
71
>95:5
>95:5
65:35
12
13
4k
4l
83
64
92:8
>95:5
40:60
18:82
= =
[(HC C CH2)Pd-
cation
(i.e.
(PH3)2]+ (13); Figure 1), was opti-
mized at the wB97XD/def2-TZVPP
level of theory.[17] Overlap of the
allenyl p2 nonbonding molecular
orbital (MO) with the PdL2 2b1
MO (which is predominantly
formed from the metal dxy orbital)
leads to a bonding orbital 1a’’ and
an antibonding orbital 2a’’ (the
LUMO of the complex). As the
bonding MO 1a’’ features greater
>90:10
95:5
14
4m
13[i]
>90:10
88:12
17:83
[a] Reactions in entries 1–7 were performed with DPEphos as the ligand; reactions in entries 8–14 were
performed with dppe as the ligand. [b] Combined yield of isolated 5 and 6, unless stated otherwise.
[c] The ee values in entries 1 and 2 were determined by HPLC on a chiral stationary phase.
Diastereomeric ratios were determined by H NMR spectroscopy. [d] Determined by 1H NMR
1
spectroscopic analysis of the crude reaction mixture. [e] The reaction was performed with [Pd(dba)2] and
the ligand. [f] Yield of the major product after isolation. [g] The major alkene isomer is assumed to have electron density on the terminal
the E configuration by analogy to other enamide products. [h] The E/Z ratio of 8 is shown. [i] Remaining
material was unreacted 4m.
carbon atoms of the allene than on
the central carbon atom, the greater
Angew. Chem. Int. Ed. 2014, 53, 1915 –1920
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1917

.
Angewandte
Communications
Scheme 3. Proposed mechanisms for the formation of alkynyl aza-
cycles and dienamides.
the degree of metal–allene bonding, the more negative charge
will be located on these carbon atoms, and the higher in
energy the antibonding orbital (2a’’) will be. Both factors
disfavor nucleophilic attack at the terminal positions of the
allenylpalladium complex, and attack at the central carbon
atom becomes more favorable.
Figure 2. Calculated variation in the LUMO energy with the ligand bite
+
= =
angle for [(HC C CH2)Pd(PH3)2] (13).
suggesting that the variation in regioselectivity is not due to
electrostatic interactions with the nucleophile, but rather to
increasing orbital control of the reaction.
The structures of the complexes of the 1-phenyl-3-
methylallenyl cation with the various bidentate-phosphine-
To investigate the effect of the ligand bite angle[18] on the
electronic structure of complex 13, we systematically varied
the angle between the two phosphine ligands in 108 incre-
ments. As this angle increases, metal–allenyl bonding lessens,
and there is a progressive drop in LUMO energy (Figure 2).
This effect is caused by a lowering in energy of the PdL2 2b1
MO, as mixing between dxz and P 3s orbitals decreases as the
bite angle increases;[12] larger bite angles thus lead to
heightened reactivity at the terminal carbon atoms, at which
the LUMO coefficient is non-zero. The changes in the
computed charges at the carbon atoms are negligible, thus
complexed palladium species employed in this study (i.e.
+
= =
[(PhC C CHMe)Pd(dppx)] ) were then optimized; com-
puted structures of the dppe and DPEphos complexes are
shown in Figure 3a. Correlation of the calculated allene
bending angles with the ligand bite angles[18] (Figure 3b) and
the experimentally observed pyrrolidine/enamide ratio (Fig-
ure 3c) shows a clear trend between structure and reactiv-
ity.[15] As the ligand bite angle increases and allenyl–palladium
Figure 1. Molecular-orbital diagram of [(HC C CH2)Pd(PH3)2]+ (13) and its fragments (EH//wB97XD/def2-TZVPP).
= =
1918
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1915 –1920

Angewandte
Chemie
ꢀ
bonding weakens, there is an elongation of the Pd C bonds to
the terminal carbon atoms (Figure 3a) and a flattening of the
allenyl group; its structure begins to resemble that of the
allenyl cation, thus favoring terminal attack.[19] With the
charges on the atoms again varying little between these
structures, correlation with the MO calculations on 13
strongly suggests a lower-energy LUMO for DPEphos than
for dppe. Therefore, attack on the terminal carbon atom to
give the alkynyl azacycle product is favored.
In conclusion, we have developed a catalyst-dependent
regioselective cyclization of propargylic carbonates equipped
with sulfonamide nucleophiles. Alkynyl pyrrolidines and
piperidines are accessed through the use of bidentate
phosphine ligands with large bite angles, and piperidine/
azepane enamides using ligands with small bite angles. This
method offers an expedient entry to both ring systems and
important insight into the structure and mechanism of the
reactions of allenylpalladium complexes.
Received: October 20, 2013
Published online: January 16, 2014
Keywords: alkynes · cyclization · nitrogen heterocycles ·
.
palladium · regioselectivity
[1] a) S. Kꢀllstrçm, R. Leino, Bioorg. Med. Chem. 2008, 16, 601;
b) D. OꢁHagan, Nat. Prod. Rep. 2000, 17, 435.
[2] For reviews of stereoselective azacycle synthesis, see: a) J. P.
Wolfe, Eur. J. Org. Chem. 2007, 571; b) M. G. P. Buffat,
Tetrahedron 2004, 60, 1701; c) P. M. Weintraub, J. S. Sabol,
J. M. Kane, D. R. Borcherding, Tetrahedron 2003, 59, 2953; for
examples of azacycle synthesis from our research group, see:
d) P. R. Walker, C. D. Campbell, A. Suleman, G. Carr, E. A.
Anderson, Angew. Chem. 2013, 125, 9309; Angew. Chem. Int. Ed.
2013, 52, 9139; e) R. L. Greenaway, C. D. Campbell, H. A.
Chapman, E. A. Anderson, Adv. Synth. Catal. 2012, 354, 3187;
f) R. L. Greenaway, C. D. Campbell, O. T. Holton, C. A. Russell,
E. A. Anderson, Chem. Eur. J. 2011, 17, 14366.
[3] D. S. B. Daniels, A. L. Thompson, E. A. Anderson, Angew.
Chem. 2011, 123, 11708; Angew. Chem. Int. Ed. 2011, 50, 11506.
[4] a) J. Tsuji, H. Watanabe, I. Minami, I. Shimizu, J. Am. Chem.
Soc. 1985, 107, 2196; b) J. Tsuji, T. Mandai, Angew. Chem. 1995,
107, 2830; Angew. Chem. Int. Ed. Engl. 1995, 34, 2589; c) L.-N.
Guo, X.-H. Duan, Y.-M. Liang, Acc. Chem. Res. 2011, 44, 111;
for an isolated case of attack on the distal carbon atom, see: d) S.
Inuki, Y. Yoshimitsu, S. Oishi, N. Fujii, H. Ohno, J. Org. Chem.
2010, 75, 3831; for a review of palladium-catalyzed cyclization
reactions, see: e) T. Vlaar, E. Ruijter, R. V. A. Orru, Adv. Synth.
Catal. 2011, 353, 809.
[5] For examples of bioactive alkynyl azacycles, see: a) J. Davies, H.
Wang, T. Taylor, K. Warabi, X.-H. Huang, R. J. Andersen, Org.
Lett. 2005, 7, 5233; b) D. J. Madar et al., J. Med. Chem. 2006, 49,
6416; c) K. L. Stevens et al., Bioorg. Med. Chem. Lett. 2009, 19,
21; for recent asymmetric approaches to 2-alkynyl azacycles, see:
d) A. Sadiq, N. Sewald, Org. Lett. 2013, 15, 2720; < lit e > G.
Archibald, C.-P. Lin, P. Boyd, D. Barker, V. Caprio, J. Org.
Chem. 2012, 77, 7968; f) G. Lemonnier, A. B. Charette, J. Org.
Chem. 2012, 77, 5832; g) T. Hashimoto, M. Omote, K. Maruoka,
Angew. Chem. 2011, 123, 9114; Angew. Chem. Int. Ed. 2011, 50,
8952.
[6] The structures, including the absolute (and where appropriate,
relative) configurations, of 5a, 5j, and 5k were determined by
single-crystal X-ray diffraction. Low-temperature single-crystal
diffraction data were collected by using an Oxford Diffraction
(Agilent) SuperNova, solved by using SIR92, and refined within
the CRYSTALS suite. The Flack x parameter refined to
Figure 3. a) Optimized structures (wB97XD/6-31G(d)) of the dppe and
+
= =
DPEphos complexes [(PhC C CHMe)Pd(dppx)] . b,c) Relationship of
the calculated allene bend angle to the calculated bidentate-phosphine
bite angle (b) and the observed regioselectivity (c).
Angew. Chem. Int. Ed. 2014, 53, 1915 –1920
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1919

.
Angewandte
Communications
0.007(10), ꢀ0.002(12), and 0.00(1) for 5a, 5j, and 5k, respec-
tively. CCDC 965014, 965015, and 965016 contain the supple-
mentary 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.
[7] For the cyclization of oxygen nucleophiles to dienol ethers, see
Ref. [3] and: C. Fournier-Nguefack, P. Lhoste, D. Sinou, Synlett
1996, 553.
[8] For central carbon heteroatom attack/intermolecular nucleo-
phile addition, see Ref. [4d] and: a) A. Okano, K. Tsukamoto, S.
Kosaka, H. Maeda, S. Oishi, T. Tanaka, N. Fujii, H. Ohno, Chem.
Eur. J. 2010, 16, 8410; b) I. Ambrogio, S. Cacchi, G. Fabrizi, A.
Prastaro, Tetrahedron 2009, 65, 8916; c) Y. Kozawa, M. Mori, J.
Org. Chem. 2003, 68, 8068; d) M. Yoshida, Y. Morishita, M.
Fujita, M. Ihara, Tetrahedron 2005, 61, 4381; e) H. Ohno, H.
Hamaguchi, M. Ohata, T. Tanaka, Angew. Chem. 2003, 115,
2751; Angew. Chem. Int. Ed. 2003, 42, 2647.
[9] Use of the ligand xantphos (4,5-bis(diphenylphosphanyl)-9,9-
dimethylxanthene) further improved the ratio of 5a to 6a to
91:9, but led to a lower ee value (91%) and the formation of an
enyne derived from b-hydride elimination (see Ref. [3]).
[10] Owing to issues with the purification of the products to remove
dba (see also Table 2, entry 8), Pd(OAc)2 was used for subse-
quent reactions; no differences in reactivity or selectivity were
observed between these Pd sources.
[14] Although the exact role of the carbonate leaving group in this
catalytic process is unclear, with recent calculations suggesting
an outer-sphere-coordination role (J. O. C. Jimꢂnez-Halla, M.
Kalek, J. Stawinski, F. Himo, Chem. Eur. J. 2012, 18, 12424), it (or
methoxide) may at some point be involved in a proton-transfer
step from the nucleophile to the allene.
[15] Complementary to this study, the influence of bidentate
phosphine ligands on the regioselectivity (propargyl versus
allenyl) of allenyl palladium cross-coupling has been reported:
a) H. Ohmiya, M. Yang, Y. Yamauchi, Y. Ohtsuka, M. Sawa-
mura, Org. Lett. 2010, 12, 1796; b) S. Ma, G. Wang, Angew.
Chem. 2003, 115, 4347; Angew. Chem. Int. Ed. 2003, 42, 4215; for
a review, see: c) S. Ma, Eur. J. Org. Chem. 2004, 1175; for related
selectivity in the position of attack on an allenyl system, see
Ref. [8c]; for a review of catalytic selective synthesis, see: d) J.
Mahatthananchai, A. M. Dumas, J. W. Bode, Angew. Chem.
2012, 124, 11114; Angew. Chem. Int. Ed. 2012, 51, 10954.
[16] For discussions on the effect of the ligand bite angle on
selectivity, see: a) P. C. J. Kamer, P. W. N. M. van Leeuwen,
J. N. H. Reek, Acc. Chem. Res. 2001, 34, 895; b) P. W. N. M.
van Leeuwen, P. C. J. Kamer, J. N. H. Reek, P. Dierkes, Chem.
Rev. 2000, 100, 2741; c) P. Dierkes, P. W. N. M. van Leeuwen, J.
Chem. Soc. Dalton Trans. 1999, 1519.
[17] All calculations were performed with Gaussian09, rev. D.01. Full
details and full references may be found in the Supporting
Information.
[11] See the Supporting Information for details.
[12] a) C. J. Elsevier, P. M. Stehouwer, H. Westmijze, P. Vermeer, J.
Org. Chem. 1983, 48, 1103; b) C. J. Elsevier, H. Kleijn, J.
Boersma, P. Vermeer, Organometallics 1986, 5, 716.
[13] a) S. Ogoshi, K. Tsutsumi, H. Kurosawa, J. Organomet. Chem.
1995, 493, C19; b) K. Tsutsumi, T. Kawase, K. Kakiuchi, S.
Ogoshi, Y. Okada, H. Kurosawa, Bull. Chem. Soc. Jpn. 1999, 72,
2687; c) M. W. Baize, P. W. Blosser, V. Plantevin, D. G. Schimpff,
J. C. Gallucci, A. Wojcicki, Organometallics 1996, 15, 164; for
a general review, see: d) A. Wojcicki, Inorg. Chem. Commun.
2002, 5, 82.
[18] Although ligand bite angles have been reported in Ref. [16], for
this study we used the specific bite angles obtained through our
calculations of the structures of the allenyl palladium(II)–dppx
species.
[19] A correlation between bending angle and regioselectivity was
also seen in nucleophilic addition to arynes: P. H.-Y. Cheong,
R. S. Paton, S. M. Bronner, G.-Y. J. Im, N. K. Garg, K. N. Houk,
J. Am. Chem. Soc. 2010, 132, 1267.
1920 www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1915 –1920