Origins of Regioselectivity in Iridium Catalyzed Allylic Substitution

Article abstract of DOI:10.1021/jacs.5b08911

Detailed studies on the origin of the regioselectivity for formation of branched products over linear products have been conducted with complexes containing the achiral triphenylphosphite ligand. The combination of iridium and P(OPh)₃ was the first catalytic system shown to give high regioselectivity for the branched product with iridium and among the most selective for forming branched products among any combination of metal and ligand. We have shown the active catalyst to be generated from [Ir(COD)Cl]₂ and P(OPh)₃ by cyclometalation of the phenyl group on the ligand and have shown such species to be the resting state of the catalyst. A series of allyliridium complexes ligated by the resulting P,C ligand have been generated and shown to be competent intermediates in the catalytic system. We have assessed the potential impact of charge, metal-iridium bond length, and stability of terminal vs internal alkenes generated by attack at the branched and terminal positions of the allyl ligand, respectively. These factors do not distinguish the regioselectivity for attack on allyliridium complexes from that for attack on allylpalladium complexes. Instead, detailed computational studies suggest that a series of weak, attractive, noncovalent interactions, including interactions of H-bond acceptors with a vinyl C - H bond of the alkene ligand, favor formation of the branched product with the iridium catalyst. This conclusion underscores the importance of considering attractive interactions, as well as repulsive steric interactions, when seeking to rationalize selectivities.

Full text of DOI:10.1021/jacs.5b08911

Subscriber access provided by NEW YORK MED COLL
Article
Origins of Regioselectivity in Iridium Catalyzed Allylic Substitution
Sherzod T. Madrahimov, Qian Li, Ankit Sharma, and John F. Hartwig
J. Am. Chem. Soc., Just Accepted Manuscript • Publication Date (Web): 26 Oct 2015
Downloaded from http://pubs.acs.org on October 27, 2015
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the
course of their duties.

Page 1 of 30
Journal of the American Chemical Society
1
2
3
4
5
Origins of Regioselectivity in Iridium Catalyzed Allylic Substitution
6
7
8
Sherzod T. Madrahimov,‡ Qian Li,§ Ankit Sharma,§ and John F. Hartwig‡§
9
‡Department of Chemistry, University of Illinois, 600 South Mathews, Urbana, Illinois 61801
10
11
12
13
14
15
16
§Department of Chemistry, University of California, Berkeley, CA 94708
Abstract. Detailed studies on the origin of the regioselectivity for formation of branched products
over linear products have been conducted with complexes containing the achiral triphenylphosphite
ligand. The combination of iridium and P(OPh)3 was the first catalytic system shown to give high
regioselectivity for the branched product with iridium and among the most selective for forming
branched products among any combination of metal and ligand. We have shown the active catalyst to
be generated from [Ir(COD)Cl]2 and P(OPh)3 by cyclometalation of the phenyl group on the ligand,
and have shown such species to be the resting state of the catalyst. We have generated a series of
allyliridium complexes ligated by the resulting P,C ligand and shown them to be competent
intermediates in the catalytic system. We have assessed the potential impact of charge, metal-iridium
bond length, stability of terminal vs internal alkenes generated by attack at the branched and terminal
positions of the allyl ligand, respectively. These factors do not distinguish the regioselectivity for
attack on allyliridium complexes from that for attack on allylpalladium complexes. Instead, detailed
computational studies suggest that a series of weak, attractive, non-covalent interactions, including
interactions of H-bond acceptors with a vinyl C-H bond of the alkene ligand, favor formation of the
branched product with the iridium catalyst. This conclusion underscores the importance of
considering attractive interactions, as well as repulsive steric interactions, when seeking to
rationalize selectivities.
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 30
1
2
3
4
5
Introduction
6
7
8
9
Allylic substitutions catalyzed by cyclometalated iridium phosphoramidite complexes1-4 have
become processes studied by many researchers5-13 and used in synthetic applications.14 Such
cyclometalated systems catalyze allylic substitution reactions with high selectivity for the formation
of branched allylic substitution products.3,4 Mechanistic studies on reactions catalyzed by these
iridium complexes have revealed the resting state of the catalyst, the origins of enantioselection, and
the importance of cyclometalation to generate the active catalyst.15-18 However, mechanistic studies
of reactions catalyzed by iridium complexes of triphenylphosphite, the first system shown by
Takeuchi to give branched products from linear allylic carbonates,19-21 are limited. A cyclometalated
species was proposed to be the active catalyst, but this species was never observed directly and
characterized.22 Most important, the origin of the high selectivity for the formation of the branched
product has not been revealed for reactions catalyzed by complexes of either phosphite or
phosphoramidite ligands.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Palladium catalyzed allylic substitution reactions typically give products from allylic substitution
in which the nucleophile adds to the least hindered terminus.23,24 Thus, palladium-catalyzed allylic
substitution with mono-substituted allylic electrophiles gives linear allylic substitution products.
Many theories have been proposed to account for the regioselectivity of allylic substitution
reactions.24,25 These theories include reaction at the terminus with the longer metal-carbon bond,26
reaction at the site with the greater positive charge,27 reaction trans to the softer donor of a ligand
with phosphorus and nitrogen donors,26,28-34 reaction at the alkyl terminus of an eneyl structure,35 and
reaction to form the more stable metal-alkene complex as product.36,37 In addition, studies have
shown that the isomer formed as product can be derived from a thermodynamic selectivity, rather
than a kinetic selectivity. For example, Yudin and co-workers showed that the branched substitution
product was kinetically favored for reactions of amines with an allylpalladium species, but the
initially formed branched product isomerized to the thermodynamically more stable linear product
during the course of the reaction.38,39
Here we describe the preparation of Ir(I) and Ir(III) complexes containing cyclometalated
triphenylphosphite ligands and detailed studies to provide insight into the origin of the high
regioselectivity for formation of the branched product from iridium-catalyzed allylic substitutions.
2
ACS Paragon Plus Environment

Page 3 of 30
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
The kinetic competence of the corresponding allyliridium complexes to be intermediates in allylic
substitution reactions catalyzed by iridium-triphenylphosphite complexes was confirmed by their
reactions with stabilized carbon and heteroatom nucleophiles. The isolated Ir(I) and Ir(III) complexes
then serve as a tool to reveal the origins of regioselectivity of iridium-catalyzed allylic substitution
reactions. The ligands trans to the two allyl termini are distinct and fortuitously lead to nearly
identical Ir-C distances to the two allyl termini. A combination of experimental studies with these
complexes and detailed computation show that the major factor favoring the formation of the
branched product is not the presence of the phosphorus ligand trans to the substituted allylic
terminus, metal-carbon bond distance, or the relative binding affinities of terminal olefins over
branched olefins to the different metal centers. Instead, the computational data suggest that a series
of weak, attractive, non-covalent interactions, including interactions of H-bond acceptors with a
vinyl C-H bond of the alkene ligand, favor formation of the branched product with the iridium
catalyst.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Results and Discussion
Preparation and characterization of complexes of iridium and palladium. To reveal the
differences in the regioselectivity of allylic substitution reactions catalyzed by iridium and palladium
complexes, we prepared mono-substituted allyliridium complexes of the achiral triphenylphosphite
ligand and mono-substituted allylpalladium complexes of the achiral DPPF ligand. As will be seen
from their structures, these two types of complexes allowed us to assess the importance of the donor
group trans to the two ends of the allyl unit and the relative metal-carbon bond lengths in the
ground-state structure on the regioselectivity for attack of the nuclei on the two termini of the allyl
group.
Preparation of Ir(III) and Ir(I) complexes containing cyclometalated triphenylphosphite
ligand. Allyliridium complexes containing a cyclometalated triphenylphosphite ligand formed from
the reaction of [Ir(COD)Cl]2, triphenylphosphite, AgBF4 and an allylic carbonate (Scheme 1). This
reaction is similar to that reported by Helmchen and co-workers for the preparation of allyliridium
complexes containing cyclometalated phosphoramidite ligands.40 To prepare the allyl complexes
containing an iridacycle derived from triphenylphosphite, [Ir(COD)Cl]2 was combined with P(OPh)3
in THF to form P(OPh)3Ir(COD)Cl. A solution of AgBF4 in THF was added to the solution of
[P(OPh)3]Ir(COD)Cl, followed by an allylic carbonate. Allyliridium complexes 3a and 3b were
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 30
1
2
3
4
5
6
formed in one hour and were isolated in 80% and 85% yields, respectively. Carbon dioxide, ethanol
or methanol and AgCl were the byproducts of this reaction.
7
8
9
O
+ CO2
Et
O
THF
1 h
Ph
O
O
O
[Ir(COD)Cl]2
+ 2 P(OPh)3
+ 2 AgBF4
+ EtOH
/MeOH
CODIr
P(OPh)
2 BF4
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
or
O
+
Me
R
Me
O
3a: R = Ph
+ AgCl
2
3b: R = Me
Scheme 1. Preparation of allyliridium complexes 3a and 3b
Preparation of Ir(I) complexes containing a single cyclometalated triphenylphosphite ligand
proved to be more challenging than the preparation of allyliridium(III) complexes containing this
ligand. Initially we attempted to prepare Ir(I) complexes by base-assisted cyclometalation of
coordinated triphenylphosphite, followed by trapping of the formed cyclometalated complex with an
alkene ligand, as we reported for complexes of cyclometalated phosphoramidite ligands.16 However,
the reaction shown in Scheme 2a formed an iridium(III) hydride ligated by two cyclometalated
phosphites and one κ1-phosphite, as evidenced by three doublets of doublets in the 31P NMR
spectrum, and a doublet of triplets at -8.9 ppm in the 1H NMR spectrum. The same complex formed
quantitatively from the reaction of [Ir(COD)Cl]2 with 6 equiv of triphenylphosphite in the presence
of an amine base (Scheme 2b).
a
H
P(OPh)3
P(OPh)
2 equiv.
[Ir(COD)Cl]2
1 equiv.
+
3
O
Ir
+ byproducts
+ PrNH3Cl
PrNH2
P(OPh)2
2 O
P
(OPh)
+ C2H4 / COE / octene
2 equiv.
4
H
b
P(OPh)3
P(OPh)2
2 O
Ir
P
O
PrNH2
+ PrNH3Cl
PH(OPh)
6 equiv.
[Ir(COD)Cl]2
1 equiv.
+
3
(OPh)
4
Scheme 2. Attempts to prepare cyclometalated Ir(I) complex through base assisted cyclometalation.
O
O
NHPr
+
CODIr
PrNClH
+
P(OPh)
2 BF4
COD
Ir
P(OPh)2
CH2CH2
+ PrNH2
Me
3b
5
Scheme 3. Preparation of the ethylene-ligated cyclometalated Ir(I) complex by nucleophilic attack on allyliridium
complex 3b, followed by a ligand exchange.
Instead, Ir(I) complex containing a single cyclometalated triphenylphosphite 5 was prepared
4
ACS Paragon Plus Environment

Page 5 of 30
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
bynucleophilic attack on cinnamyl or crotyl iridium complexes 3a or 3b containing cyclometalated
triphenylphosphite. Complex 3b was chosen as the most suitable allyliridium complex for this
procedure because the addition of propylamine would form a volatile allylic amine product.
According to this procedure, crotyliridium complex 3b was treated with an excess of propylamine.
After nucleophilic attack, the reaction vessel was exposed to 1 atm of ethylene to form the
ethylene-ligated complex 5 in 85% isolated yield (Scheme 3).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Characterization of the cyclometallated Ir(III), Ir(I) complexes. Cinnamyliridium complex 3a
and crotyliridium complex 3b containing iridacyclic centers both form as a mixture of two
diastereomers, as determined by 31P-NMR spectroscopy. The 31P-NMR spectra of the two
diastereomers of cinnamyl complex 3a in CH2Cl2 consisted of resonances at 97.0 ppm and 95.7 ppm
in a 70:30 ratio, respectively. The 31P-NMR spectra of the two diastereomers of crotyliridium
complex 3b formed in CH2Cl2 consisted of resonances at 99.4 ppm and 97.1 ppm in a 43:57 ratio.
Both complexes are sparingly soluble in THF. Single crystals were obtained of 3a that were suitable
for X-ray crystallographic analysis.
Both cinnamyl and crotyl complexes 3a and 3b crystallize predominantly as one of the two
diastereomers, as determined by NMR spectroscopy at low temperature. However, dissolution of the
crystals in CH2Cl2 at room temperature led to the rapid establishment of the ratios of diastereomers
observed in the solutions prior to recrystallization (70:30 for 3a and 43:57 for 3b). In both cases, the
major diastereomer present at equilibrium is the isomer that was isolated by crystallization. The
isomerization was sufficiently slow at -78 ̊C to obtain NMR spectra of the single isomers. Thus
solutions of complexes 3a and 3b prepared at -78 °C containing predominantly one isomer were
characterized by 1H-NMR, 31P-NMR, and 2D-gCOSY-NMR spectroscopies at low temperature. The
31P-NMR spectrum at -60 °C of the isolated isomer of the cinnamyliridium complex 3a consisted of
1
one resonance at 97.3 ppm. The H and 2D-gCOSY NMR spectrum of the isolated isomer of 3a
contained resonances for the allyl group at 4.90, 4.63, 4.34 and 3.98 ppm. A solution of complex 3b
containing predominantly one isomer generated by dissolving the isolated sample of 3b in CD2Cl2 at
1
-78 °C was also characterized by 31P-NMR, H-NMR spectroscopies. The 31P-NMR spectrum
consisted of a single peak at 99.7 ppm at -30 ̊C. The allyliridium complexes 3a and 3b also were
characterized by high resolution ESI-MS. M+ values of 3a and 3b obtained by ESI-MS matched the
calculated values.
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 30
1
2
3
4
5
6
7
8
9
PhO
PhO
PhO
PhO
P
Ir
P
Ir
Ph
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
O
Ph
H
H
a: major isomer
a: minor isomer
distances (Å)
angles (deg)
2.256
2.285
2.220
2.281
2.092
2.239
2.356
P–Ir–C3
92.2
104.1
99.1
65.5
97.7
94.1
77.7
78.0
90.3
93.3
97.3
Ir–P
Ir–C1
Ir–C2
Ir–C3
Ir–C4
Ir–cent1
Ir–cent2
cent2–Ir–C1
cent2–Ir–cent1
C4–Ir–P
C4–Ir–cent1
C4–Ir–C1
P–Ir–cent1
cent1–Ir–C1
C1–Ir–C3
cent2–Ir–P
cent2–Ir–C3
91.5
C4–Ir–C3
Figure 1. ORTEP diagram of 3a (counterion and hydrogens are omitted for clarity; ellipsoids are drawn to 35%
probability).
The solid-state structure of the major diastereomer of 3a is shown in Figure 1. The complex
adopts a structure that is close to a six-coordinate octrahedral geometry with the allyl group,
phosphorus atom, and an olefin coordinated in one plane. The covalently bound carbon atom and one
of the π-bound olefins occupy the two additional, mutually trans positions of the structure.
The relative lengths of the Ir-C1 and Ir-C3 bonds have been proposed to influence the
regioselectivity of allylic substitution. In complex 3a, these bond lengths are both 2.28 Ǻ. The
similarity between these bond lengths likely results from the binding of the unsubstituted allylic
terminus trans to the phosphorus atom and binding of the substituted allylic terminus trans to the
equatorial olefin. The stronger trans influence of the phosphine would lengthen the Ir-C bond trans
to the phosphorus and the substituent on the other allyl terminus would lengthen the Ir-C bond trans
to the olefin. Apparently, the influence of the trans ligand on the Ir-C distance is similar to the
influence of the phenyl substituent on the iridium-carbon length.
The ethylene-ligated iridium (I) complex 5 forms by the displacement of the olefinic unit of the
allylic substitution product generated by nucleophilic attack on the allyliridium complex 3b. This
6
ACS Paragon Plus Environment

Page 7 of 30
Journal of the American Chemical Society
1
2
3
1
complex was characterized by H, 31P and 13C-NMR spectroscopies. The 31P-NMR spectrum of
4
5
6
ethylene-ligated Ir(I) complex 5 in THF consists of a singlet at 128.4 ppm.
7
8
9
Stoichiometric reactions of allyliridium complexes with nucleophiles and reactions catalyzed by
cyclometalated allyliridium triphenylphosphite complexes
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Cyclometalated allyliridium triphenylphosphite complexes have been proposed as intermediates in
allylic substitution reactions catalyzed by iridium triphenylphosphite complexes.22 Although,
cyclometalated iridium triphenylphosphite complexes have been isolated previously,41-43 they had
not been evaluated as potential intermediates in iridium catalyzed allylic substitution reactions. We
performed a series of catalytic and stoichiometric reactions to assess the competence of allyliridium
complexes 3a and 3b to be intermediates in iridium catalyzed allylic substitution reactions.
Stoichiometric reactions of allyliridium complexes 3a and 3b with nucleophiles
Allyliridium complexes 3a and 3b were allowed to react with stabilized enolate, alkoxide and
amine nucleophiles to assess the competence of 3a and 3b to be intermediates in iridium catalyzed
allylic substitution reactions catalyzed by complexes of iridium containing cyclometallated
triphenylphosphite ligands. Suspensions of allyliridium complexes 3a or 3b in THF-d8 were allowed
to react with the nucleophiles. These reactions form adducts between the iridium(I) fragment and the
alkene unit of the substitution product, as described for addition of aniline and propylamine
previously.16 A slight excess of triphenylphosphine was added to the reaction solutions after full
conversion to form the known PPh3-ligated Ir(I) product and the free organic product.2 The yield of
1
the reaction was determined by H-NMR spectroscopy by integrating the peak corresponding to the
product vs. the internal standard (Si2Me6 or mesitylene). The outcome of these stoichiometric
reactions is summarized in Table 1.
The yields of the reactions of cinnamyliridium complex 3a depended on the strength of the
nucleophile. Anionic nucleophiles (Table 1, entries 1, 2, 4) or alkylamine nucleophiles (Table 1,
entry 5) formed the products of nucleophilic attack in quantitative yields. However, the reaction of
aniline with triethylamine as proton acceptor (Table 1, entry 7) gave a lower yield of the organic
product (60%). In this reaction a small amount of product from double allylation of aniline also
formed. Very high regioselectivities for the formation of the branched products were observed for all
of the reactions (6:7 > 94:6). Both secondary and tertiary, stabilized carbon nucleophiles formed the
branched products with high yield and selectivity (Table 1, entries 1 and 2). Because Takeuchi
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 30
1
2
3
4
5
6
7
8
9
reported allylic substitutions in ethanol solvent in some cases, we compared the selectivity and yield
in EtOH to those in THF. The reactions of octylamine and malonate nucleophiles in EtOH (entries 3
and 6) occurred to give the products in yields and selectivities that are similar to those on THF.
Table 1. Reactions of allyliridium complexes with nucleophiles
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Nu
PPh3
THF
O
+
+
R
Nu
O
(COD)Ir
P(OPh)
2 BF4
R
(COD)Ir
P(OPh)2
7
6
R
+ Nucleophile
3a: R = Ph
PPh3
3b: R = Me
Allyliridium
complex
Nucleophile
Solvent
6:7
yield
(%)
100
96
1
3a
3a
3a
3a
3a
3a
3a
3b
3b
3b
3b
NaCH(COOMe)2
NaCMe(COOMe)2
NaCH(COOMe)2
KOPh
THF
THF
EtOH
THF
THF
EtOH
THF
THF
THF
THF
THF
99:1
97:3
95:5
94:6
97:3
99:1
97:3
99:1
99:1
99:1
99:1
2
3
70
4
100
100
100
60
5
OctylNH2
6
OctylNH2
7
PhNH2/TEA
NaCH(COOMe)2
NaCMe(COOMe)2
KOPh
8
80
9
96
10
11
70
OctylNH2
70
The yields of organic products from the reactions of crotyliridium complex 3b with sodium malonate,
potassium phenoxide and octylamine were slightly lower than those of the reactions of
cinnamyliridium complex 3a with the same nucleophiles (Table 1, entries 8, 10 and 11). However,
the selectivities for the formation of branched organic products were universally high (6:7 = 99:1).
The yields of the reactions of crotyliridium complex 3b with stabilized carbon nucleophiles were
similar to the yields of the reactions of cinnamyliridium complex 3a with the same nucleophiles
(Table 1, entries 6 and 7).
Allylic substitution reactions catalyzed by allyliridium complex 3a
8
ACS Paragon Plus Environment

Page 9 of 30
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
The activity of the cyclometalated allyliridium complex 3a as a catalyst for allylic
substitution reactions was assessed with a series of nucleophiles. Ethyl cinnamyl carbonate 1 was
allowed to react with stabilized carbon and heteroatom nucleophiles in the presence of 8%
cinnamyliridium complex 3a in THF at 23 °C and EtOH at 50 °C (Table 2, entry 1 - 2).
Table 2. Allylic substitution reactions catalyzed by cinnamyliridium complex 3
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Nu
Nu
3a (8 mol%)
+
+
NuX
Ph
OCO2Et
Ph
Ph
Solvent, Temp.
4h
6
7
NuX
NaCH(CO2Me)2
octylamine
Solvent Temp Yield 6:7
85% 95:5
87% >99:1
THF
EtOH
25 °C
50 °C
The reaction of sodium dimethylmalonate with allylic electrophile 1 in THF catalyzed by 3a
formed the allylic substitution product in yields and selectivities that are comparable to those
observed for the reactions catalyzed by a mixture of [Ir(COD)Cl]2 and triphenylphosphite in the
same medium.21 The reaction of alkylamines with cinnamyl carbonates catalyzed by [Ir(COD)Cl]2
and P(OPh)3 is reported to occur in refluxing ethanol,19 and the reaction of octylamine with
carbonate 1 catalyzed by isolated 3a occurred in good yield with high branched-to-linear selectivity
in this medium.44 Moreover, monitoring of the reaction of octylamine with ethyl cinnamyl carbonate
catalyzed by the combination of [Ir(COD)Cl]2 and P(OPh)3 in a 1:1 or 1:2 ratio in EtOH by 31P NMR
spectroscopy showed that a cyclometalated complex is the catalyst resting state. A single complex
was observed by 31P NMR spectroscopy, and this complex was identified by independent synthesis45
to be the cyclometalated Ir(I) bisposphite complex shown in Figure 2.46 The formation of this
complex from a 1:1 or 1:2 ratio of metal to ligand indicates that the five–coordinate structure with an
iridium-carbon bond is more stable than the analogous square-planar complex lacking the
κ1-phosphite; the same type of structure was observed to form with a 1:1 or 1:2 ratio of [Ir(COD)Cl]2
and phosphoramidite ligand previously.2
O
(COD)Ir
P(OPh)2
P(OPh)3
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 30
1
2
3
4
5
Figure 2. Identity of the resting state of the catalyst in reactions conducted the combination of excess
P(OPh)3, [Ir(COD)Cl]2 in EtOH, as determined by 31P NMR spectroscopy.
6
7
8
9
These stoichiometric and catalytic reactions showed that the allyliridium complexes with
cyclometalated triphenylphosphite ligand 3a and 3b are chemically and kinetically competent to be
intermediates in allylic substitution reactions catalyzed by iridium triphenylphosphite complexes.
Moreover, these data, in combination with the structural data in Figure 1, speak to the potential
origins of regioselectivity proposed for allylic substitution reactions. In the solid-state structure
(Figure 1), the substituted allylic terminus is located trans to an olefin, and the unsubstituted
terminus is located trans to a phosphorus ligand. The iridium-carbon distance for both allylic termini
is the same 2.28 Ǻ. Thus, the observed high selectivity of nucleophilic attack for the formation of
branched allylic substitution product shows that neither the iridium-carbon bond length nor the
presence of a soft phosphorus atom trans to the allyl terminus are responsible for the high selectivity
of nucleophilic attack at the more substituted position.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Kinetic studies on oxidative addition of branched and linear allylic acetates to ethylene-ligated
iridium (I) complex 5 containing a cyclometalated triphenylphosphite ligand.
The selectivity for nucleophilic attack at the more substituted position of the allyl group was
sufficiently high to prevent a direct assessment of the relative rate for nucleophilic attack at the
substituted allylic terminus (to form branched product) and the unsubstituted allylic terminus (to
form linear product) of allyliridium complexes; the amounts of linear product were too small to
measure precisely. However, insight into the relative rates for nucleophilic attack can be gained by
studying the kinetics of oxidative addition of branched and linear allylic electrophiles to the
cyclometalated Ir(I) triphenylphosphite complex 5. Studies on the oxidative addition of linear and
branched allylic electrophiles can address this issue because this reaction is the reverse of the formal
reductive elimination by nucleophilic attack on the allyl complex and the rates can be measured with
these two reactants independently.
The oxidative addition of allylic esters to Ir(I) systems is not strongly favored
thermodynamically.16 Thus, we studied the oxidative addition of allylic trifluoroacetates, which
contain a good leaving group and were shown to add to Ir(I) complexes of a cyclometallated
phosphoramidite.18 In particular, the reactions of the linear and branched isomers of the
phenalkyl-substituted allylic trifluoroacetates 8 and 9 with the Ir(I) ethylene adduct 5 were studied.
10
ACS Paragon Plus Environment

Page 11 of 30
Journal of the American Chemical Society
1
2
3
4
5
These reactions occurred to form an equilibrium mixture of unreacted 5 and two diastereoisomers of
allyliridium complexes 3c.
6
7
8
9
Due to the reversibility of the reaction, we measured the rate constants for the approach to
equilibrium. These measurements were conducted with excess amounts of allylic electrophiles (25
equiv) 8 and 9 and with an excess of ethylene (20 equiv) to render the reaction pseudo-first order in
the concentration of ethylene-ligated Ir(I) complex 5.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. Change of concentration over time graph for the reaction of branched allylic electrophile 9
with ethylene-ligated complex 5.
The oxidative additions of linear and branched allylic trifluoroacetates 8 and 9 to 5 form
allyliridium complex 3c-TFA as a mixture of two isomers, as determined by 31P-NMR spectroscopy
31
(eq 1). These isomers formed in a 65:35 ratio, as determined from the intensities of the P-NMR
signals at 100.2 ppm and 98.6 ppm corresponding to the two complexes. The reaction was monitored
31
by P-NMR spectroscopy, and a plot of concentration vs. time for the reaction of branched allylic
electrophile 9 with ethylene-ligated complex 5 is provided in Figure 3. The rate constants for
oxidative addition were derived from the rate constant of approach to equilibrium and equilibrium
constant of the reaction. The rate constants for oxidative addition of linear and branched allylic
electrophiles are provided in Table 3, and details of calculation of rate constants are provided in the
experimental section.
11
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 30
1
2
3
4
O
k1
5
6
7
8
Ph
O
CF3
k-1
8 25 equiv
O
(1)
O
O
THF, T
(COD)Ir
or
+
(COD)Ir
P(OPh)2
P(OPh)
2 OCOCF3
O
CF3
k2
5
(CH2)2Ph
3c-TFA
Ph
k-2
9
9 25 equiv
THF, T
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 3. Rate constants for oxidative addition of linear and branched allylic electrophiles 8 and 9
with the ethylene ligated Ir(I) complex 5 (eq 1)
8 or 9
T (
̊
k1 or k2 (s-)
8·10-4
1
2
3
4
5
9
9
9
9
8
-10
-15
-20
-25
20
3.5·10-4
1.3·10-4
7.5·10-5
2.7·10-4
The rate constants k1 for the oxidative addition of branched electrophile 9 to ethylene-ligated
complex 5 were measured at -10, -15, -20 and -25 C. The rate constant k2 for oxidative addition of
̊
linear electrophile 8, was measured at 20 °C. To make a direct comparison between the rate constant
k1 and k2 for addition of the linear and branched allylic esters, respectively, at the same temperature,
we extrapolated the rate constants for oxidative addition of branched electrophile 9 to
ethylene-ligated complex 5 at the lower temperatures to a rate constant at 20 °C. Activation
parameters were derived from an Eyring plot for the rate constants k2 at -10 to -25 °C (Figure 3). The
enthalpy of activation for the reaction of branched electrophile 9 with ethylene ligated complex 5
was 20.5 kcal/mol, and the entropy of activation for the same reaction involving dissociation of
ethylene and addition of the allylic ester was 4.4 eu. The rate constant for oxidative addition of 9 to 5
at 20 °C was calculated to be 0.046 s-1 from these activation parameters.
12
ACS Paragon Plus Environment

Page 13 of 30
Journal of the American Chemical Society
1
2
3
-12.5
4
5
6
7
8
-13
-13.5
-14
ΔHǂ = 20.5 ± 2.2 kcal/mol
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
-14.5
-15
-15.5
0.00379 0.00384 0.00389 0.00394 0.00399 0.00404
1/T
Figure 4. Eyring plot for the oxidative addition of the branched electrophile 9 to ethylene-ligated
complex 5.
These kinetic studies showed that the rate constant for oxidative addition of the branched
electrophile 9 to ethylene-ligated complex 5 is approximately 170 times faster than the rate constant
of oxidative addition of the linear electrophile 8 to Ir(I) complex 5. This ratio of 170 corresponds to a
difference in the free energy of activation of 2.99 kcal. This selectivity is consistent with the
observed selectivity for nucleophilic attack to form the branched substitution product over attack to
form the linear product.
The difference of 2.99 kcal/mol between the activation energies for oxidative addition of
linear and branched electrophiles likely arises from the combination of two factors. The first factor is
the difference between the ground state energies of linear allylic trifluoroacetate 8 and branched
allylic trifluoroacetate 9. When an excess of linear allylic trifluoroacetate 8 is allowed to equilibrate
in the presence of ethylene ligated complex 5 and excess ethylene, a 3:1 mixture of linear 8 and
branched 9 forms. This equilibrium ratio shows that the ground state of linear 8 is 0.64 kcal/mol
more stable than the ground state of branched 9. A second factor that could influence the difference
between the two activation energies is the binding energies of the branched and linear alkenes to the
cyclometalated iridium center. The branched alkene is a monosubstituted alkene and should have a
greater bonding affinity than the linear disubstituted alkene.47 The contributions from the difference
in ground state energies of the allylic esters and the difference in bonding affinity of the two olefins
are summarized in Figure 5.
13
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 30
1
2
3
4
5
O
6
7
O
O
CF3
(COD)Ir
P(OPh)2
TSlinear
G‡ = ca. 2.99 kcal/mol
8
9
Ph
Ph
OC(O)CF3
10a
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
TSbranched
10a
10b
O
(COD)Ir
P(OPh)2
O
5
9 + 5
PhO
PhO
(COD)Ir
P(OPh)2
P
Ir
C
G = 0.64 kcal/mol
O
C1
3C2
O
8 + 5
[3c]TFA
(CH2)2Ph
F3C(O)CO
Ph
O
CF3
10b
8
[3c]
Figure 5. Energy diagram for the reaction of linear and branched allylic electrophiles 8 and 9 with
ethylene-ligated complex 5.
The products from substitution at the least hindered allylic terminus are typically formed by
palladium catalysts.48,49 The difference in selectivity between the palladium-catalyzed allylic
substitution and iridium-catalyzed allylic substitution could stem from a smaller difference in
bonding affinities of terminal and internal olefins for palladium because the metal center contains
fewer ligands and would be less sterically crowded. However, as will be discussed in detail later in
this paper, this difference in affinities of the terminal and internal alkenes to iridium is computed to
be similar to the difference in affinities of these alkenes to palladium. Thus, the difference in energy
of the terminal and internal alkene complexes could contribute to the formation of branched over
linear substitution products but this difference in energy does not distinguish between the selectivity
of iridium and palladium complexes. In addition, nucleophilic attack by amine nucleophiles on
allylpalladium complexes is usually reversible, and this reversibility allows for the formation of
thermodynamically more stable linear products and masks the kinetic selectivity for formation of
branched products.38,50
Computational Study of the Regioselectivity.
We conducted DFT calculations to provide additional insight into the origin of the regioselectivity
of the iridium catalyzed allylic substitution. We computed the energies of the allyl-iridium
14
ACS Paragon Plus Environment

Page 15 of 30
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
complexes that would react with nucleophiles and the energies of the alkene complexes that would
precede the oxidative addition to form the allyliridium species. We also computed the activation
energies for reactions to form and to consume these complexes. We did so for the complexes with
the coordination sphere of the active catalyst containing the unsymmetrical environment created by
the cyclometallation process. In addition, we computed these energies for a prototypical palladium
complex to assess the origin of the difference in regioselectivity between iridium-catalyzed and
palladium-catalyzed allylic substitution. The calculations of the iridium and palladium systems were
conducted on reactions involving the formation or cleavage of C-O bonds and C-C bonds to
determine the effect of the leaving group and nucleophile on the energetics to form and cleave bonds
at the more and less substituted termini of the allyl group.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
To begin this portion of our studies, we computed the energies of eight different stereoisomers of
the cyclometallated iridium fragment bound to a terminal alkene (Figure 6). The computed energies
of these complexes vary by 5.1 kcal/mol. We also computed the energies of four different
stereoisomers of the same metal fragment bound to the E internal alkene of the linear allylic
trifluoroacetate (Figure 7). The energies of the four stereoisomers of the internal alkene complexes
are similar to each other.
15
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 30
1
2
3
4
Figure 6. Intermediates of the terminal alkene coordinated iridium complexes.
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 7. Intermediates of the internal alkene coordinated iridium complexes.
We deduced experimentally the structure of the major isomer of the oxidative addition product 3a
(cinnamyl-coordinated iridium complex) by X-ray diffraction (Figure 1). In complex 3a, the
phenyl-substituted terminus of the π-allyl moiety is cis to the phosphorus atom and is trans to the
olefinic unit of the cod ligand. In the structure of the related iridium complex (also deduced by X-ray
diffraction) generated from a phosphoramide ligand,17,18 the substituted terminus of the π-allyl
moiety is cis to the olefin unit of the cod ligand and is trans to the phosphorus atom. On the basis of
these structures, we propose the structure of the minor isomer of the allyliridium complex in the
current work to be 3a’ shown in Figure 1. We presume that the alkene complexes which generate the
π-allyl-Ir complex have the same orientation of the alkyl groups as in the products of the oxidative
addition 3a and 3a’. Thus, we located the transition states for the oxidative addition from
branched-C, linear A, branched-B’, and linear-C. These transition states lead to the π-allyl-Ir
complexes Int-π-allyl-1 and Int-π-allyl-2, which have the same orientation of the allyl moiety as in
complexes 3a and 3a ’.
16
ACS Paragon Plus Environment

Page 17 of 30
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
LG = OC(O)CF3
ΔG
THF 298 K
(ΔG gas 298 K
[ΔH gas 298 K
in kcal/mol
)
]
TS-linear-minor
PhO
PhO
25.6
TS-linear-major
P
Ir
(29.6)
[26.8]
PhO
F3C(O)CO
O
23.4
P
Ir
PhO
(27.7)
[25.0]
Ph
TS-branched-minor
OC(O)CF3
O
TS-branched-major
Ph
20.5
(25.5)
[23.0]
12.4
(9.3)
[6.9]
linear-C
12.6
(9.9)
[6.6]
18.9
14.3
(19.5)
[17.9]
13.6
(18.3)
[16.9]
(22.7)
[20.2]
12.3
(9.1)
[6.5]
11.1
(6.8)
[2.1]
linear-A
Int-π-
allyl-1
Int-π-allyl-2
branched-B'
branched-C
PhO
LG
LG
PhO
PhO
Ph
PhO
TFAO
PhO
P
Ir
PhO
PhO
Ph
PhO
P
Ir
P
Ir OTFA
O
P
Ir
PhO
(R)
P
Ir
PhO
O
Ph
O
O
O
(S)
Ph
0.0
+ internal and
terminal alkene
R next to the olefinic portion of cod
R next to P atom
Figure 8. Intermediates and transition states for the oxidative addition of a linear and branched
allylic trifluoroacetate to the iridium catalyst.
As shown in Figure 8, the computed transition states for the reactions of branched-C (via
TS-branched-major) and branched-B’ (via TS-branched-minor) are 18.9 kcal/mol and 20.5
kcal/mol in activation free energy above the combination of the ethylene-ligated iridium catalyst and
the free alkenes. In contrast, the computed transition states TS-linear-major and TS-linear-minor
involving oxidative addition of the linear allylic ester linear-A and linear-C lie 23.4 kcal/mol and
25.6 kcal/mol higher in activation free energy than the combination of uncoordinated alkene and
ethylene-coordinated Ir-catalyst. These data indicate that the iridium complexes containing the
terminal alkene are much more reactive than the iridium complexes containing the internal alkene.
This finding is consistent with the experimental observation that the rate for oxidative addition of the
branched allylic ester containing a terminal alkene is much faster than that of the linear allylic ester
containing an internal alkene.
The principle of microscopic reversibility leads to information on the reductive elimination by
nucleophilic attack of a trifluroacetate nucleophile on the π-allyl iridium complex. On the basis of the
energetics calculated for oxidative addition, reductive elimination from iridium complex
17
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 18 of 30
1
2
3
4
5
6
7
8
9
Int-π-allyl-1 (or Int-π-allyl-2) via TS-branched-major (or TS-branched-minor) to form the
branched isomer should be much faster than reductive elimination via TS-linear-major (or
TS-linear-minor) to form the linear isomer of the substitution product (Figure 9). Thus, the energies
that we computed by DFT methods are consistent with the experimental observation.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 9. Reductive elimination by nucleophilic attack of trifluoroacetate on the allyliridium catalyst
in which the allyl is substituted by a CH2CH2Ph group.
Origin of regio-selectivity of the Ir-catalyzed allylation. We sought to use these computations
to reveal the factors creating the high selectivity of the iridium catalyst for the formation of branched
allylic substitution products. Many allyl complexes containing two different donors trans to the allyl
ligand have been studied, and the trans effect of these ligands has been proposed to control the
regioselectivity with palladium catalysts,26,28-31 From the above computational data, we could
conclude that the ligand located trans to a terminus of the allyl unit does not strongly influence the
regioselectivity observed for the reactions catalyzed by the cyclometallated iridium complex.
Although the orientation of the allylic termini relative to the phosphite and olefin ligands are
different in TS-branched-major and TS-branched-minor (major and minor refer to the major and
minor diastereomers), these two transition states lie at similar free energies. The same result is also
observed for the two transition states TS-linear-major and TS-linear-minor that form the two
complexes of the linear allylic substitution products. Because the transition state energies for reaction
18
ACS Paragon Plus Environment

Page 19 of 30
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
of the major and minor π-allyliridium complexes are similar to each other, we focused on the
reaction pathway for formation of the major diastereoisomer to conserve computational resources.
In the experimental structure of the major diastereoisomer of the π-allyl complex 3a in Figure 10
determined by single-crystal X-ray diffraction, one of the hydrogen atoms in the cod moiety is close
to one of the fluorine atoms in the tetrafluoroborate anion. The distance between these two atoms is
only 2.43 Å, which is much shorter than the sum of the Van der Waal radii of hydrogen and fluorine
atoms (2.67 Å). This short distance suggests the presence of a C-H···F attractive interaction between
the cod ligand and the tetrafluoroborate anion in this π-allyl iridium complex. For our computations
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
-
on the allylic substitution reaction, we used trifluoroacetate as the nucleophile. Like BF4 ,
-
trifluoroacetate is anionic; therefore, it could interact with the cod ligand in a similar fashion to BF4 .
We considered that an attractive C-H···O interaction between the same hydrogen atom in the cod
ligand and one of the oxygen atoms in the trifluoroacetate anion could stabilize this structure. The
differences in the C-H···O attractive interactions in the transition states to form the branched and
linear products might contribute to the regioselectivity of the Ir-catalyzed allylic substitution
reactions. With this hypothesis in mind, we analyzed the differences in the computed transition states
for the formation of the branched and the linear products.
Figure 10. C-H···F attractive interaction in π-allyl complex 3a.
In these transition states, the two oxygen atoms in the trifluoroacetate nucleophiles are both close
to the same C-H bond in the cod ligand (Figure 11). Several additional pairs of C-H···O interactions
between the C-H bond in the cod ligand or the C-H bond in the π-allyl moiety and the oxygen atoms
in the trifluoroacetate anion appear to be present in this structure. The distances between the
hydrogen and oxygen atoms listed in Figure 11 are all shorter than the sum of their van der Waals
radii (2.72 Å). These C-H···O interactions appear to serve as bridges between the trifluoroacetate
19
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 20 of 30
1
2
3
4
5
6
nucleophile and the allyliridium complex, facilitating approach of the nucleophile to the allyl
electrophile.
7
8
9
Using NBO second-order perturbation analysis, we computed the stabilization energy provided by
these C-H···O interactions in the transition states for attack on the allyliridium species. The C-H···O
interactions in the transition state for the formation of the branched product (TS-branched-major)
are 4.4 kcal/mol stronger than those in the transition state for the formation of the linear product
(TS-linear-minor). This 4.4 kcal/mol difference in the stabilization energies from the C-H···O
interactions in the two transition states is similar to the 4.5 kcal/mol difference in activation free
energies of the same transition states. As a result, we propose that the C-H···O interactions in the
transition states for oxidative addition are a major factor contributing to the formation of the
branched product for the iridium-catalyzed allylic substitution reaction.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 11. C-H···O interactions in the transition states TS-branched-major and TS-linear-major.
To assess further the effect of the C-H···O interactions on regioselectivity, we computed the
structure and reaction of the analogous complexes containing a C-F bond in place of the C-H bond of
the cod ligand that we propose participates in C-H···O interactions (Figure 12). In the transition state
for reaction of the fluoro-COD complex, one of the strongest C-H···O interactions is absent and two
pairs of C-H···F interactions are now present. The transition state leading to the branched product
from the fluoro-COD complex is more favorable than the transition state leading to the linear product,
but by only 1.6 kcal/mol. Using NBO second-order perturbation analysis, we obtained the
20
ACS Paragon Plus Environment

Page 21 of 30
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
stabilization energy provided by the C-H···O and C-H···F interactions in the nucleophilic
substitution transition states. The C-H···O and C-H···F interactions in the transition state forming the
branched product (TS-branched-F) are 1.8 kcal/mol stronger than those in the transition state
leading to the linear product (TS-linear-F). The difference in energies of the C-H···X interactions is,
again, similar to the difference in energies of the transition states. The smaller difference in transition
state energies for the system lacking the strongest C-H···O interaction between the C-H bond in the
cod ligand and the oxygen atom in the trifluoroacetate anion is consistent with the importance of this
interaction in controlling the relative energies of the transition states. The remaining weak
interactions also favor the formation of the branched product, suggesting the importance of a series
of weak interactions in controlling the selectivity of these reactions.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 12. C-H···O and C-H···F interactions in the transition states TS-branched-F and
TS-linear-F.
In addition to computing the Ir-catalyzed allylic substitution reaction with trifluoroacetate as the
nucleophile, we computed the reaction with the anion of dimethylmalonate as nucleophile (The
branched : linear selectivity was 10:1 when the dimethylmalonate anion was used in the experiment.).
The computed energies of the two transition states for addition of dimethylmalonate to form the
branched and linear products were closer than those for addition of trifluoroacetate. The computed
transition state for the formation of the branched products is only 0.7 kcal/mol lower than that for the
21
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 22 of 30
1
2
3
4
5
6
7
8
9
formation of the linear product. It is clear from the computational work that the difference in energy
is much smaller for the reaction of the more delocalized and more hindered carbon nucleophile,
and this result is in accord with the experimental data. However, the NBO second-order perturbation
analysis shows that the difference in the stabilization energies contributed by the C-H···O
interactions in the transition state is much larger than 0.7 kcal/mol. The difference in the sum of the
interaction energies is 3.9 kcal/mol, favoring the formation of the branched product.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 13. C-H···O interactions in transition states involving dimethylmalonate as the nucleophile.
The difference between the energetics of the transition states and the sum of the computed weak
interactions could stem from the absence of the cation in our calculations for the addition of the
malonate anion. In the experimental system, the electrophile is allyl trifluoroacetate and the
nucleophile is sodium dimethylmalonate. One byproduct is sodium trifluoroacetate, and the sodium
trifluoroacetate could function as a bridge between the allyliridum moiety and the dimethylmalonate
anion during the nucleophilic substitution. Thus, we computed the transition states containing
sodium trifluoroacetate as a bridge, and these structures are shown in Figure 14. Two tetrahydrofuran
molecules were used as ligands for sodium to fill two of the six coordination sites around the sodium
cation. The other four coordination sites were occupied by the trifluoroacetate and the malonate. The
computed transition state for the formation of the branched products was 1.4 kcal/mol lower than that
for the formation of the linear product. This value fits the experimentally observed regioselectivity
22
ACS Paragon Plus Environment

Page 23 of 30
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
more closely than the value computed in the absence of the sodium trifluoroacetate. Moreover, the
stabilization energies contributed by the C-H···O interactions in the transition state were computed
by NBO second-order perturbation analysis to be 1.3 kcal/mol in favor of formation of the branched
product. This value matches closely with the calculated difference in activation free energy.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 14. C-H···O interactions in transition states involving dimethylmalonate as the nucleophile
with NaOTFA as a bridge between the malonate and the iridium complex. The energies are
referenced to that for TS-branched in Figure 13.
Computational studies of analogous Pd-catalyzed allylation. It is well established for
palladium-catalyzed allylic substitution that reactions of soft nucleophiles, such as malonates and
amines, tend to form higher ratios of linear to branched products than do reactions of hard
nucleophiles and that they form smaller amounts of the branched products than do reactions
catalyzed by complexes of iridium (Figure 15).23,51,52 To compare the origin of the regioselectivity
for reactions of the iridium system to the origin of the regioselectivity for reactions of palladium
complexes, we computed the relative energies of the transition states for the formation of linear and
23
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 24 of 30
1
2
3
4
5
6
7
8
branched products from attack of hard and soft nucleophiles on the allylpalladium complex and
interactions between the nucleophiles and the allylpalladium species that could control
regioselectivity..
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 15. Selectivity in Ir-catalyzed and Pd-catalyzed allylic substitution.
We computed the reaction of allylpalladium complexes ligated by PPh3 with trifluoroacetate and
the anion of dimethyl malonate. The allyl group studied contains one phenethyl substituent to match
the allyl group on the iridium system we computed. According to our DFT calculations, the reaction
of this allylpalladium complex with trifluoroacetate (Figure 16) forms a higher ratio of branched over
linear product than does the reaction of the same complex with dimethylmalonate (Figure 17). These
computational results are in accord with the experimentally observed selectivity.27,53
NBO second-order perturbation analysis was used to reveal the weak interactions that are present
in the transition states to form the branched and linear products. This analysis indicates that the
stabilization energies contributed by the C-H···O interactions in the transition state to form the
branched product from reaction of trifluoroacetate are larger than those in the transition state to form
the linear product by 1.8 kcal/mol. In contrast, the stabilization energies contributed by the C–H···O
interactions in the transition state for nucleophilic attack by dimethylmalonate are larger in the
transition state to form the linear product than they are in the transition state to form the branched
product by 1.2 kcal/mol.
Because the reaction of the anion of a malonate with an allylic ester would contain an alkali metal
counterion and the reaction would contain alkali metal carboxylate as a byproduct, we also computed
activation free energies for reaction of dimethylmalonate with the allylpalladium species in the
presence of sodium trifluoroacetate as a bridge between the allyl moiety and the dimethylmalonate
24
ACS Paragon Plus Environment

Page 25 of 30
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
anion (Figure 18). The activation free energy computed for this reaction is much lower than that
computed for this reaction without sodium trifluoroacetate. For the structures that include sodium
trifluoroacetate, the difference between the free energies of the transition states for formation of the
linear and branched products from addition of the anion of dimethyl malonate is 4.5 kcal/mol,
favoring the formation of the linear product.54 This value is also similar to the difference in weak
interactions between the two transition states (3.1 kcal/mol favoring the formation of the linear
product).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 16. C-H···O interactions in transition states involving trifluoroacetate as the nucleophile in
Pd-catalyzed allylation.
These two examples of the Pd-catalyzed allylic substitution reactions and the NBO second-order
perturbation analysis of the structures of the transition states suggest that the C-H···O interactions,
again, are a major factor controlling the regioselectivities for the allylic substitution. However, the
regioselectivity for the reactions of allylpalladium complexes changes when different nucleophiles
are used. Unlike the C–H···O interactions in the iridium system, which favor the formation of the
branched product for all the nucleophiles studied, the C-H···O interactions in the palladium system
are computed to be stronger for formation of the linear product than for formation of the branched
product with the soft nucleophile, but are computed to be stronger for formation of the branched
product than for formation of the linear product with the hard nucleophile.
25
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 26 of 30
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 17. C-H···O interactions in transition states involving dimethylmalonate as the nucleophile in
Pd-catalyzed allylation.
26
ACS Paragon Plus Environment

Page 27 of 30
Journal of the American Chemical Society
1
2
3
4
5
Figure 18. C-H···O interactions in transition states involving dimethylmalonate as the nucleophile
with NaOTFA as a bridge between the malonate nucleophile and the palladium catalyst. The
energies are referenced to the combination of TS linear in Figure 17 and NaOTFA
6
7
8
9
Conclusion
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Allyliridium complexes 3a and 3b with cyclometalated triphenylphosphite were prepared. Through a
series of stoichiometric and catalytic reactions these allyliridium complexes were found to be
competent to be intermediates in allylic substitution reactions catalyzed by a combination of iridium
and triphenylphosphite. The selectivity of nucleophilic attack along with solid state structural data
showed that iridium-carbon bond lengths or presence of a trans-phosphorus do not influence the
position of nucleophilic attack, which is also confirmed by the computational study. Ethylene-ligated
Ir(I) complex 5 containing a cyclometalated triphenylphosphite ligand was prepared from
allyliridium complexes 3a and 3b. Kinetic measurements on oxidative addition of branched and
linear allylic electrophiles to ethylene-ligated complex 5 showed that branched electrophiles have
significantly favorable kinetics for oxidative addition over the linear electrophiles. Computational
studies on the origin of the regioselectivity of the allylic substitution reactions catalyzed by iridium
triphenylphosphite system revealed the relative strength of the C-H···O interactions in transition
state for the formation of the branched and linear products appears to be a major factor controlling
the regioselectivity. The C-H···O interactions are stronger in the transition state leading to the
branched products for reactions of both heteroatom and carbon nucleophiles. Besides, calculations on
the regioselectivity of the reactions of allylpalladium complexes ligated by triphenylphosphine
support the proposal that the difference in the C-H···O interactions in the transition state for
nucleophilic attack, again, are a major factor controlling the site of attack.
ASSOCIATED CONTENT
Experimental methods, characterization of complexes and reaction products, supplementary kinetic
data, and computational output. This material is available free of charge via the Internet at
http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
jhartwig@berkeley.edu
27
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 28 of 30
1
2
3
4
5
Notes
6
7
The authors declare no competing financial interests.
8
9
ACKNOWLEDGMENT
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
This work was supported by NIGMS in the National Institutes of Health (GM-55382). We thank
Johnson-Matthey for a gift of [Ir(COD)Cl]2. We thank Dr. Konstantin Troshin for obtaining NMR
spectra of the products of catalytic reactions.
References
(1) Ohmura, T.; Hartwig, J. F. J. Am. Chem. Soc. 2002, 124, 15164.
(2) Kiener, C. A.; Shu, C.; Incarvito, C.; Hartwig, J. F. J. Am. Chem. Soc. 2003, 125, 14272.
(3) Hartwig, J. F.; Stanley, L. M. Acc. Chem. Res. 2010, 43, 1461.
(4) Hartwig, J. F.; Pouy, M. J. Top. Organomet. Chem. 2011, 34, 169.
(5) Alexakis, A.; Polet, D. Org. Lett. 2004, 6, 3529.
(6) Tissot-Croset, K.; Polet, D.; Alexakis, A. Angew. Chem. Int. Ed. 2004, 43, 2426.
(7) Helmchen, G.; Dahnz, A.; Dubon, P.; Schelwies, M.; Weihofen, R. Chem. Commun. 2007,
675.
(8) Lyothier, I.; Defieber, C.; Carreira, E. M. Angew. Chem. Int. Ed. 2006, 45, 6204.
(9) Singh, O. V.; Han, H. J. Am. Chem. Soc. 2007, 129, 774.
(10) Singh, O. V.; Han, H. Org. Lett. 2007, 9, 4801.
(11) Zhuo, C.-X.; Zheng, C.; You, S.-L. Acc. Chem. Res. 2014, 47, 2558.
(12) Liu, W.-B.; Xia, J.-B.; You, S.-L. Top. Organomet. Chem. 2012, 38, 155.
(13) Liu, W.-B.; Reeves, C. M.; Stoltz, B. M. J. Am. Chem. Soc. 2013, 135, 17298.
(14) Tosatti, P.; Nelson, A.; Marsden, S. P. Org. Biomol. Chem. 2012, 10, 3147.
(15) Kiener, C. A.; Shu, C. T.; Incarvito, C.; Hartwig, J. F. J. Am. Chem. Soc. 2003, 125, 14272.
(16) Markovic, D.; Hartwig, J. F. J. Am. Chem. Soc. 2007, 129, 11680.
(17) Madrahimov, S. T.; Markovic, D.; Hartwig, J. F. J. Am. Chem. Soc. 2009, 131, 7228.
(18) Madrahimov, S. T.; Hartwig, J. F. J. Am. Chem. Soc. 2012, 134, 8136.
(19) Takeuchi, R.; Ue, N.; Tanabe, K.; Yamashita, K.; Shiga, N. J. Am. Chem. Soc. 2001, 123,
9525.
(20) Takeuchi, R.; Kashio, M. J. Am. Chem. Soc. 1998, 120, 8647.
(21) Takeuchi, R.; Kashio, M. Angew. Chem., Int. Ed. Engl. 1997, 36, 263.
(22) Helmchen, G.; Dahnz, A.; Duebon, P.; Schelwies, M.; Weihofen, R. Chem. Commun. 2007,
675.
(23) Poli, G.; Prestat, G.; Liron, F.; Kammerer-Pentier, C. Top. Organomet. Chem. 2012, 38, 1.
(24) Acemoglu, L.; Williams, J. M. J. In Handbook of Organopalladium Chemistry for Organic
Synthesis 2002; Vol. 2, p 1689.
(25) Trost, B. M.; Toste, F. D. J. Am. Chem. Soc. 1999, 121, 4545.
(26) Ward, T. R. Organometallics 1996, 15, 2836.
(27) Åkermark, B.; Hansson, S.; Krakenberger, B.; Vitagliano, A.; Zetterberg, K.
Organometallics 1984, 3, 679.
(28) Pretot, R.; Pfaltz, A. Angew. Chem. Int. Ed. 1998, 37, 323.
(29) Sprinz, J.; Kiefer, M.; Helmchen, G.; Reggelin, M.; Huttner, G.; Walter, O.; Zsolnai, L.
28
ACS Paragon Plus Environment

Page 29 of 30
Journal of the American Chemical Society
1
2
3
Tetrahedron Lett. 1994, 35, 1523.
4
(30) Brown, J. M.; Hulmes, D. I.; Guiry, P. J. Tetrahedron 1994, 50, 4493.
(31) Togni, A.; Burckhardt, U.; Gramlich, V.; Pregosin, P. S.; Salzmann, R. J. Am. Chem. Soc.
1996, 118, 1031.
5
6
7
8
(32) Pamies, O.; Dieguez, M.; Claver, C. J. Am. Chem. Soc. 2005, 127, 3646.
(33) Helmchen, G.; Pfaltz, A. Acc. Chem. Res. 2000, 33, 336.
(34) Mata, Y.; Pàmies, O.; Diéguez, M. Adv. Synth. Catal. 2009, 351, 3217.
(35) Leahy, D. K.; Evans, P. A. In Modern Rhodium-Catalyzed Organic Reactions; Evans, P. A.,
Ed.; Wiley-VCH: Weinheim, 2005, p 191.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(36) Trost, B. M.; Lautens, M. J. Am. Chem. Soc. 1983, 105, 3343.
(37) Trost, B. M.; Lautens, M. Tetrahedron 1987, 43, 4817.
(38) Watson, I. D. G.; Yudin, A. K. J. Am. Chem. Soc. 2005, 127, 17516.
(39) Dubovyk, I.; Watson, I. D. G.; Yudin, A. K. J. Am. Chem. Soc. 2007, 129, 14172.
(40) Spiess, S.; Raskatov, J.; Gnamm, C.; Brödner, K.; Helmchen, G. Chem. Eur. J. 2009, 15,
11087.
(41) Bedford, R. B.; Castillòn, S.; Chaloner, P. A.; Claver, C.; Fernandez, E.; Hitchcock, P. B.;
Ruiz, A. Organometallics 1996, 15, 3990.
(42) Lin, C.-H.; Lin, C.-Y.; Hung, J.-Y.; Chang, Y.-Y.; Chi, Y.; Chung, M.-W.; Chang, Y.-C.;
Liu, C.; Pan, H.-A.; Lee, G.-H.; Chou, P.-T. Inorg. Chem. 2012, 51, 1785.
(43)Lin, C.-H.; Chang, Y.-Y.; Hung, J.-Y.; Lin, C.-Y.; Chi, Y.; Chung, M.-W.; Lin, C.-L.; Chou,
P.-T.; Lee, G.-H.; Chang, C.-H.; Lin, W.-C. Angew. Chem., Int. Ed. 2011, 50, 3182.
(44) The reaction with octylamine in THF did not provide detectable amounts of allylic
substitution products. Considering both oxidative addition of allyl carbonate and nucleophilic
addition to allylic complex 1 occurs at room temperature in THF, we speculate that this lack in
reactivity in THF is due to competing or irreversible binding of amine to the iridium (I) complex.
(45) Bedford, R. B.; Castillón, S.; Chaloner, P. A.; Claver, C.; Fernández, E.; Hitchock, P. B.;
Ruiz, A. Organometallics 1996, 15, 3990.
(46) The balance of the iridium in reactions with a 1:1 ratio of iridium to ligand is, presumably,
unreacted [Ir(COD)Cl]2.
(47) Hartley, F. R. Chem. Rev. 1973, 73, 163.
(48) Hayashi, T.; Konishi, M.; Kumada, M. J. Chem. Soc., Chem. Commun. 1984, 107.
(49) Hayashi, T.; Yamamoto, A.; Hagihara, T. J. Org. Chem. 1986, 51, 723.
(50) Mackenzie, P. B.; Whelan, J.; Bosnich, B. J. Am. Chem. Soc. 1985, 107, 2046.
(51) Trost, B. M.; Crawley, M. L. Chem. Rev. 2003, 103, 2921.
(52) Dubovyk, I.; Watson, I. D. G.; Yudin, A. K. J. Org. Chem. 2013, 78, 1559.
(53) Åkermark, B.; Zetterberg, K.; Hansson, S.; Krakenberger, B.; Vitagliano, A. J. Organomet.
Chem. 1987, 335, 133.
(54) The difference in free energies of activation for formation of the linear and branched
products in the absence of sodium trifluoroacetate (Figure 17) fits the value deduced from the
experimental regioselectivity more closely than the value in the absence of sodium trifluoroacteatee.
The experimental linear:branched selectivity is between 2:1 and 9:1, depending on the allylic
electrophile and the nucleophile. However, the energies of activation in Figure 18 are 17-22 kcal/mol
lower than those in Figure 17. Although the difference in selectivity in Figure 18 is slightly larger
than that corresponding to the experimental selectivity, the transition states do favor formation of the
29
ACS Paragon Plus Environment