Electrostatic Control of Regioselectivity in Au(I)-Catalyzed Hydroarylation

Article abstract of DOI:10.1021/jacs.6b11971

Competing pathways in catalytic reactions often involve transition states with very different charge distributions, but this difference is rarely exploited to control selectivity. The proximity of a counterion to a charged catalyst in an ion paired comple

Full text of DOI:10.1021/jacs.6b11971

Subscriber access provided by UB + Fachbibliothek Chemie | (FU-Bibliothekssystem)
Article
Electrostatic control of regioselectivity in Au(I)-catalyzed hydroarylation
Vivian M Lau, William C. Pfalzgraff, Thomas E. Markland, and Matthew W. Kanan
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.6b11971 • Publication Date (Web): 22 Feb 2017
Downloaded from http://pubs.acs.org on February 22, 2017
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 14
Journal of the American Chemical Society
1
2
3
4
5
Electrostatic control of regioselectivity in Au(I)ꢀcatalyzed hydroarylation
Vivian M. Lau, William C. Pfalzgraff, Thomas E. Markland, and Matthew W. Kanan*
Department of Chemistry, Stanford University, 337 Campus Drive, Stanford, California 94305, United States
6
7
8
9
ABSTRACT
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
Competing pathways in catalytic reactions often involve transition states with very different charge distributions, but this
difference is rarely exploited to control selectivity. The proximity of a counterion to a charged catalyst in an ion paired
complex gives rise to strong electrostatic interactions that could be used to energetically differentiate transition states. Here
we investigate the effects of ion pairing on the regioselectivity of the hydroarylation of 3ꢀsubstituted phenyl propargyl ethers
catalyzed by cationic Au(I) complexes, which forms a mixture of 5ꢀ and 7ꢀsubstituted 2Hꢀchromenes. We show that changing
–
the solvent dielectric to enforce ion pairing to a SbF6 counterion changes the regioselectivity by up to a factor of 12
depending on the substrate structure. Density functional theory (DFT) is used to calculate the energy difference between the
putative productꢀdetermining isomeric transition states (ꢁꢁE‡) in both the presence and absence of the counterion. The
change in ꢁꢁE‡ upon switching from the unpaired transition states in high solvent dielectric to ion paired transition states in
low solvent dielectric (ꢁ(ꢁꢁE‡)) was found to be in good agreement with the experimentally observed selectivity changes
across several substrates. Our calculations indicate that the origin of ꢁ(ꢁꢁE‡) lies in the preferential electrostatic stabilization
of the transition state with greater charge separation by the counterion in the ion paired case. By performing calculations at
multiple different values of the solvent dielectric, we show that the role of the solvent in changing selectivity is not solely to
enforce ion pairing, but rather that interactions between the ion paired complex and the solvent also contribute to ꢁ(ꢁꢁE‡).
Our results provide a foundation for exploiting electrostatic control of selectivity in other ion paired systems.
INTRODUCTION
Homogeneous catalysis often proceeds through catalytic cycles with one or more charged intermediates. Numerous
strategies to control selectivity in catalysis rely on ion pairing to charged intermediates. Most commonly, counterions are
used to induce phase transfer or to provide a steric environment that biases the reactivity of an intermediate involved in a
productꢀdetermining step.1ꢀ4 For example, in asymmetric ion pairing catalysis, the proximity of a chiral counterion to a
charged reactive intermediate provides a chiral environment that affords asymmetric induction. In some cases, counterions
have also been shown to bind to reactive centers or directly participate in a chemical step.5ꢀ7 Aside from enforcing proximity,
ion pairing gives rise to electrostatic interactions that affect the energies of charged species and could thereby impact
selectivity. Specifically, the strength of the electrostatic interaction between a counterion and a charged transition state
depends on the distribution of charge in the transition state. If competing transition states have different charge distributions,
electrostatic interactions induced by ion pairing could change the relative barrier heights.
We previously investigated the ability of ion pairing to effect electrostatic control of selectivity for a sulfoxide
rearrangement reaction catalyzed by cationic Au(I) complexes.8 We found that ion pairing changes the regioselectivity of the
rearrangement by a magnitude that depends on the dielectric constant (ε) and the substrate structure. Density functional
theory (DFT) was used to calculate dipole moments (µ) of competing transition states for the putative selectivityꢀdetermining
step. A correlation was observed between the magnitude of the selectivity change and the difference in the magnitude of the
1
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 14
1
2
3
4
5
transition state dipole moments (ꢁ|ꢂ|) for a set of substrates. The results supported an electrostatic model in which ion pairing
changes selectivity by preferentially stabilizing transitions states with greater charge separation.
6
7
8
9
Using ion pairing–induced electrostatic effects in other reactions would benefit from a better understanding of the
electrostatic interactions that impact selectivity. In this study, we provide a second example of regiocontrol afforded by ion
pairing and a more comprehensive computational study of the effects of ion pairing on the energy barriers that determine
selectivity. The reaction studied here is a Au(I)ꢀcatalyzed hydroarylation of 3ꢀsubstituted phenyl propargyl ethers, which
forms a mixture of 5ꢀ and 7ꢀsubstituted 2Hꢀchromenes and is mechanistically distinct from the previously studied Au(I)ꢀ
catalyzed sulfoxide rearrangement reaction (Table 1).9ꢀ13 Mechanistic studies suggest that the reaction proceeds through
initial πꢀactivation of the alkyne by Au(I) followed by electrophilic hydroarylation.14,15 We show that ion pairing to a small
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
–
nonꢀcoordinating counterion such as SbF6 changes the regioselectivity to favor the 7ꢀsubstituted isomer to an extent that
depends on the nature of the substituent. We then use DFT to calculate the relative energies (ꢁꢁE‡) of the putative productꢀ
–
determining transition states in the presence and absence of an SbF6 counterion. The calculated change in the relative
energies induced by ion pairing, ꢁ(ꢁꢁE‡), is found to be consistent with the experimentally observed selectivity changes
across multiple substrates. The calculation of ꢁ(ꢁꢁE‡) provides much more direct support of the electrostatic effect on
selectivity than what was previously possible when comparing selectivity changes to ꢁ|ꢂ|. Furthermore, our calculations
enable us to assess the dependence of ꢁ(ꢁꢁE‡) on the position of the counterion and the contribution of ion pair–solvent
interactions to the electrostatic effect.
RESULTS AND DISCUSSION
The Au(I)ꢀcatalyzed hydroarylation reaction was initially examined using Clꢀsubstituted 2a in CH2Cl2 (ε = 8.9) and
toluene (ε = 2.4). We and others have previously shown by diffusion NMR studies that cationic Au(I) complexes form tight
ion pairs with their counterions in solvents with ε≤5, but exhibit little or no ion pairing in solvents with ε≥9.8,16 To assess the
effect of the ligand on reaction selectivity, we surveyed a series of cationic Au(I) catalysts bearing ligands with different
–
electronic and steric properties. The counterion used was SbF6 , which previously showed the strongest electrostatic effects
for Au(I)ꢀcatalyzed sulfoxide rearrangements.8 Reactions were performed for 3 h at ambient temperature using a catalyst
loading of 2 mol% and the yields and product distributions were measured by NMR with an internal standard (Table 1). In
all cases, the 2Hꢀchromenes 3a and 4a were the major products, though the yields varied substantially depending on the
catalyst (11%–67%). Regioisomeric benzofurans 5a and 6a were formed as minor products (comprising less than 25% of
products (Table S1)), and the mass balance was composed of unreacted starting material. In CH2Cl2, the ratio of 4a:3a
showed relatively little dependence on the electronic or steric properties of the Au(I) catalyst used, ranging from 0.8:1 to
1.4:1. When the reaction solvent was changed to toluene, however, regioisomer 4a became the dominant product for all
catalysts. The magnitude of the selectivity change obtained using CH2Cl2 versus toluene as solvent was the largest when
using [BisPhePhosXDAu]SbF6 ([1]SbF6).17 With [1]SbF6, the product ratio 4a:3a increased from 0.8:1 in CH2Cl2 to 5.0:1 in
toluene with similar yields, corresponding to a 6.3ꢀfold change in selectivity. Control experiments showed no product
interconversion upon 24 h incubation of either 2Hꢀchromenes 3a and 4a or minor products 5a and 6a under reaction
conditions with 2 mol% of [1]SbF6.
2
ACS Paragon Plus Environment

Page 3 of 14
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Table 1. Ligand influence on regioselectivity of Au(I)ꢀcatalyzed hydroarylation of 2a, in CH2Cl2 (ε = 8.9) and toluene (ε =
2.4)
+
+
Cy
O
[Au]
[Au]
P
Au+
R
Au(I) catalyst
2 mol%
Cl
O
O
+
+
O
O
solvent (5 mM)
R
2
Cl
3 h, r.t.
R = OiPr
[1]+
Cl
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
Cl
Cl
ꢀ = 1.5 D
ꢀ = 3.0 D
2a
[3a‡]+
[4a‡]+
3a
4a
P
toluene
Solvent = CH2Cl2
Au(I) catalyst
4a : 3aa
1.3 : 1
1.3 : 1
1.0 : 1
1.4 : 1
0.8 : 1
1.2 : 1
1.4 : 1
Yield [%]b
P
c
CH2Cl2
[Ph3PAu(NCMe)]SbF6
63
63
56
50
66
29
32
1.5
3.1
2.8
3.0
6.3
1.8
2.6
[(oꢀtol)3PAu(NCMe)]SbF6
[SPhosAu(NCMe)]SbF6
[tBuXPhosAu(NCMe)]SbF6
[BisPhePhosXDAu]SbF6 ([1]SbF6)d
[IMesAu(NCMe)]SbF6
[IPrAu(NCMe)]SbF6
a Determined by NMR of crude reaction mixture. b Combined yield of 3a and 4a, determined by NMR using an internal standard. c ꢀ
ꢁꢂꢃꢄꢅꢆꢅ
and ꢀꢇꢈ ꢇꢃ are the product ratios (4a:3a) in toluene and CH2Cl2. d [1]SbF6 was generated by halide metathesis of [1]Cl with AgSbF6. No
ꢉ
ꢉ
reaction was observed when using only BisPhePhosXDAuCl or AgSbF6.
To further explore the role of the solvent, the hydroarylation of 2a catalyzed by [1]SbF6 was performed in a series of
nonꢀcoordinating solvents and solvent mixtures with ε ranging from 2.4 to 10.4 (Figure 1, Table S3). While the product ratio
of 4a:3a remained unchanged at 0.8:1 when ε ≥ 9, selectivity for 4a increased monotonically as ε was decreased below 9. The
influence of solvent on regioselectivity in this reaction reflects only the ε, and not the molecular structure of the solvent, and
is thus inconsistent with a solventꢀdependent change in reaction mechanism. The εꢀdependent selectivity of the reaction of 2a
suggests that electrostatic interactions, facilitated by tight ion pairing in lowꢀε solvents, alter the relative activation barriers of
regioisomeric cationic Au(I)ꢀbound transition states.
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 14
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 1. Influence of solvent dielectric (ε) on regioselectivity in reaction of 2a is dependent on catalyst counterion (X).
Solvent mixture ratios are v/v.
The role of electrostatic interactions in the hydroarylation of 2a was further investigated by comparing SbF6– to B(ArF)4 ,
a much larger counterion. Active catalyst [1]B(ArF)4 was generated by in situ halide metathesis of [1]Cl with [Na]B(ArF)4. In
contrast to [1]SbF6, when the hydroarylation of 2a was carried out with [1]B(ArF)4, the product ratio of 4a:3a remained at
0.7:1 in all solvents tested, with ε ranging from 2.4 to 11.4. Similar results were observed when [IPrAu(NCPh)]B(ArF)4 was
–
used as the catalyst (Table S1). We have previously demonstrated that B(ArF)4 forms tight ion pairs with cationic Au(I)
–
complexes in CDCl3 at catalytically relevant concentrations.8 However, because of its large radius and chargeꢀdiffuse nature,
–
B(ArF)4 does not effectively discriminate between cationic transition states that have different charge distributions. Taken
–
together, the change in selectivity with decreasing ε when using a small counterion such as SbF6 , and the absence of a
–
response to ε when using the large counterion B(ArF)4 , indicate that the regioselectivity depends on the strength of the
Coulombic interaction between the reactive cationic Au(I) species and its counterion.
Effect of substrate on ε–dependent selectivity. We next examined the effects of changing the steric and electronic
properties of the hydroarylation substrate. A series of 3ꢀsubstituted phenyl propargyl ether substrates was reacted with
[1]SbF6 in toluene (ε=2.4), CHCl3 (ε=4.8), and CH2Cl2, (ε=8.9) (Figure 2, Table S4). The regioselectivity of reactions with
substrates bearing a halogen, triflate, or CF3 substituent showed substantial dependence on ε. As seen with 2a, the reactions
of Br, I, and triflateꢀsubstituted substrates (2b–d) all showed >6ꢀfold changes in regioselectivity when ε was changed from
8.9 to 2.4. Though the magnitude of the εꢀdependent change in selectivity was similar across the substrates 2aꢀd, the absolute
selectivity for product 4 obtained in toluene increased in the order I < Br < Cl < OTf, such that the hydroarylation of 2d in
4
ACS Paragon Plus Environment

Page 5 of 14
Journal of the American Chemical Society
1
2
3
4
5
this solvent gave a product ratio for 4d:3d of 13:1. In CH2Cl2 where ion pairing is weak, the absolute selectivity for product 3
increases in the order OTf < Cl < Br ≈ I. This enhanced selectivity for product 3 corresponds to the increasing C–R bond
length in substrates 2a–d, and is explained by the increased steric accessibility of 3.
6
7
8
9
In order to investigate the selectivity of substrates with strongly electron withdrawing F and CF3 substituents, methyl
groups were installed at the propargylic position to facilitate the reaction. Fluorineꢀbearing substrate 2e showed a 3ꢀfold
change in regioselectivity from ε=8.9 to ε=2.4. Trifluoromethylꢀsubstituted substrate 2f showed the largest εꢀdependent
change in regioselectivity, with the product ratio 4f:3f increasing from 0.5:1 in CH2Cl2 to 6.0:1 in toluene, corresponding to
an 12ꢀfold change. The εꢀdependence of the selectivity is not perturbed by the presence of propargylic methyl groups. The
product ratio of dimethylpropargylꢀfunctionalized Clꢀbearing substrate 2g increased 6.3ꢀfold, from 0.6:1 in CH2Cl2 to 3.8:1 in
toluene, matching the selectivity change of parent substrate 2a.
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
In contrast to substrates 2aꢀg, the regioselectivity of the reaction of methylꢀsubstituted 2h catalyzed by [1]SbF6 showed
very little dependence on ε. The product ratio 4h:3h increased only slightly from 1:1 to 1:1.4 when ε was decreased from 8.9
to 2.4. Similarly, the selectivity with 5ꢀpropargyloxy dihydroindene 2i showed very weak dependence on ε, although this
substrate intrinsically favors the more sterically accessible product 4i. Weak εꢀdependence was also observed with substrates
that have a mesylate substituent (2j), OMe (2k), and OAc (2l). The absence of significant εꢀdependent selectivity for
substrates 2h–l demonstrates that the ion pairing effect seen with 2a–g does not correlate with the steric demand of the aryl
substituent. Instead, insensitivity to ε for 2h–l suggests that the competing regioisomeric transition states cannot be
discriminated through electrostatic effects for these substrates.
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 14
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 2. Substrateꢀdependent regioselectivity of hydroarylation in representative solvents toluene, CHCl3, and CH2Cl2.
Electrostatic regiodivergent synthesis of 2Hꢀchromenes. As illustrated in Table 1, the regioselectivity of Au(I)ꢀ
catalyzed hydroarylation is difficult to control through manipulation of catalyst structure. Previously reported examples of
regioselective 2Hꢀchromene synthesis via Au(I)ꢀcatalyzed hydroarylation require coordinating directing groups and are
limited to bulky internal alkynes, giving rise to 4ꢀsubstituted 2Hꢀchromenes.13 Electrostatic interactions can be exploited in
conjunction with steric effects to achieve the regioselective hydroarylation of 3ꢀhalophenyl propargyl ether substrates in the
absence of directing groups or alkyne substitution. Steric selectivity can be modulated by varying the identity of the haloꢀ
substituent while electrostatic selectivity is controlled by solvent ε. Selection of reaction substrate and solvent provides
regioselective access to 2Hꢀchromenes bearing (pseudo)halide functionality at either the 5ꢀ or 7ꢀ position (Scheme 1).
6
ACS Paragon Plus Environment

Page 7 of 14
Journal of the American Chemical Society
1
2
3
Scheme 1. Application to regiodivergent 2Hꢀchromene synthesis.
4
5
O
6
7
8
R = I (2m)
CH2Cl2
(50 mM)
O
O
N
7
O
O
O
9
B
O
O
98%, 7 + 8
(5:2 isomeric ratio)
48% 7 isolated
I
O
90%
(1.5 equiv)
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]SbF6
(2.5 mol%)
(3m:4m = 5:2
)
Pd(OAc)2 (5 mol%)
SPhos (10 mol%)
1,4-dioxane/H2O
16 h, 60 °C
TfO
O
R
O
R = OTf (2n)
Toluene
O
85%
(25 mM)
(4n:3n = 14:1)
8
100%, 8 + 7
(14:1 isomeric ratio)
To access 5ꢀsubstituted 2Hꢀchromenes, hydroarylation of 3ꢀiodophenyl propargyl ether 2m was catalyzed by [1]SbF6 in
CH2Cl2 with high efficiency to give a combined isolated yield of regioisomers 3m and 4m of 90% in a 5:2 ratio. The product
distribution is consistent with that observed for similar substrate 2c using the same solvent and catalyst. The inseparable
mixture of 3m and 4m was functionalized with 2ꢀbenzofuranboronic acid MIDA ester using Pdꢀcatalyzed SuzukiꢀMiyaura
coupling under literature conditions18 to give the regioisomeric (benzofuranꢀ2ꢀyl)ꢀ2,2ꢀdimethylꢀ2Hꢀchromenes 7 and 8 in a
5:2 ratio, in 98% total yield. Recrystallization of the product mixture yielded pure isomer 7 in 48% isolated yield, and its
structure was confirmed by Xꢀray crystallography (Figure S1). The regioisomer 8 was accessed by cyclization of 3ꢀ
trifluoromethanesulfonylphenyl propargyl ether 2n using catalyst [1]SbF6 in toluene to enforce strong ion pairing. The 7ꢀ
substituted product 4n was obtained as the major isomer in a 14:1 ratio, and in 85% combined yield. Crossꢀcoupling of the
cyclization products proceeded quantitatively to access 8 in ~93% isomeric purity.
MECHANISTIC INSIGHTS FROM SIMULATION
The ε–dependent regioselectivity seen above for ionic Au(I)ꢀcatalyzed hydroarylation suggests that the relative energies
of competing transition states can be perturbed by the reaction medium and proximity of a counterion. These results present a
potentially new avenue for enhancing catalytic selectivity, provided that the underlying physical effects that give rise to it are
understood. To elucidate the origin of the selectivity changes, we performed electronic structure calculations, which can be
used to calculate the product ratios measured in experiments and to unravel the physical effect that gives rise to the observed
enhancement in selectivity.
Theoretical approach. If the selectivityꢀdetermining step of a particular reaction is irreversible, then the product ratio
depends on the relative rates of the two competing reactions in that selectivityꢀdetermining step. Previous experimental and
computational studies suggest that the hydroarylation proceeds through an electrophilic aromatic substitution pathway that
involves πꢀcoordination of the alkyne to the Au(I) complex and subsequent 6ꢀendoꢀdig cyclization.14,15 This step is likely to
be irreversible because the carbocation intermediate that it forms undergoes rapid, thermodynamically favorable proton
transfer.19 If the 2Hꢀchromene products 3 and 4 originate from a common intermediate, then the product distribution, P, of
the reaction in a particular solvent can be approximated using transition state theory as:
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 14
1
2
3
4
5
‡
‡
ꢐ
‡
∆ꢎꢏꢊ ꢐꢑ∆ꢒꢏꢋ
ꢖꢖꢒ
ꢀ = [[ꢊꢋ]] = ꢌꢍ
≡ ꢌꢍ
(1)
ꢓ
ꢕ
ꢓ ꢕ
ꢔ
ꢔ
where kB is Boltzmann’s constant, T is the temperature, and ꢀG(3‡) and ꢀG(4‡) are the free energies required to reach the
transition states. Here 3‡ and 4‡ are the cyclization transition states corresponding to 2Hꢀchromene products 3 and 4 (see
Table 1). If the entropic contributions to the barrier heights are similar for the two competing transition states (which is likely
if both transition states originate from a common intermediate), then these contributions will largely cancel and the product
ratio will be dominated by the difference in activation energies, ꢁꢁE‡:
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
‡
ꢖꢖꢗ
ꢀ ≈ ꢌꢍ
(2)
ꢓ
ꢕ
ꢔ
For the Au(I)ꢀcatalyzed hydroarylation studied here, since 3‡ and 4‡ originate from a common starting point, the
difference in their activation energies ꢁꢁE‡ for this reaction is independent of the energy of their common intermediate 2 and
is given by
‡
ꢙ
‡ꢚ ꢙ ꢚ
ꢙ
‡ꢚ ꢙ ꢚ
ꢙ
‡ꢚ
ꢙ
‡ꢚ
ΔΔE = ꢏꢘ ꢊ – ꢘ ꢛ ꢐ – ꢏꢘ ꢋ – ꢘ ꢛ ꢐ = ꢘ ꢊ – ꢘ ꢋ
(3)
where E(2) is the energy of the common starting point and E(3‡) and E(4‡) are the energies of the transition states. Therefore,
if the relative energies of the two competing regioisomeric transition states are known, then the product ratios measured
above can be approximated using eqs. (2) and (3). Another quantity of particular relevance to experiment is the change in
product ratio when comparing toluene to CH2Cl2 as reaction media, given by:
‡
‡
ꢖꢖꢥ
ꢑ ꢖꢖꢥ
‡
ꢖꢙꢖꢖꢥꢚ
ꢝꢞꢟꢠꢡꢢꢡ
ꢣꢤ ꢣꢟ
ꢉ
ꢉ
ꢍ
ꢝꢞꢟꢠꢡꢢꢡ = ꢌ
≡ ꢌꢍ
(4)
ꢜ
ꢓ
ꢕ
ꢓ ꢕ
ꢔ
ꢔ
ꢜ
ꢣꢤ ꢣꢟ
ꢉ
ꢉ
In eq. (4), ꢁ(ꢁꢁE‡) is the change in the difference between the barrier heights of the two competing transition states in
toluene versus CH2Cl2. Therefore, in eq. (4), the simulation needs only to capture how the difference in the barrier heights
responds to changes in the solvent, which is likely to be given more accurately than absolute activation energies, as
systematic errors in individual barrier heights will cancel when taking the difference.
Table 2. Relative activation energies of unpaired and ion paired regioisomeric transition statesa
ꢀ
ꢀ
ꢁꢂꢃꢄꢅꢆꢅ
ꢁꢂꢃꢄꢅꢆꢅ
ꢁꢁE‡
[kcal/mol]
4 : 3
(theory)
4 : 3
(expt.)
ꢁ(ꢁꢁE‡)
[kcal/mol]
R =
Transition state
ꢀꢇꢈ ꢇꢃ
(theory)
ꢀꢇꢈ ꢇꢃ
ꢉ
ꢉ
ꢉ ꢉ
(expt.)
6.3
Unpaired
Ion Pair
Unpaired
Ion Pair
Unpaired
Ion Pair
Unpaired
Ion Pair
Unpaired
Ion Pair
–0.07
–2.30
–0.30
–0.13
0.00
1.1 : 1
50 : 1
1.7 : 1
1.2 : 1
1.0 : 1
1.0 : 1
4.5 : 1
1.4 : 1
1.5 : 1
22 : 1
0.8 : 1
5.0 : 1
1.0 : 1
1.3 : 1
2.0 : 1
2.6 : 1
0.9 : 1
1.4 : 1
1.9 : 1
13 : 1
Cl (2a)
–2.2
+0.2
0.0
45
Me (2h)
OMe (2k)
OMs (2j)
OTf (2d)
0.8
1.0
0.3
15
1.3
1.3
1.6
6.8
0.00
–0.88
–0.20
–0.23
–1.81
+0.7
–1.6
a Based on energies of ion unpaired transition states in CH2Cl2, and ion paired transition states in toluene
8
ACS Paragon Plus Environment

Page 9 of 14
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
We first performed DFT calculations to obtain the energies and geometries of 3‡ and 4‡ for various substrates in implicit
solvents matching the dielectric constants of toluene and CH2Cl2 (see the SI). On the basis of diffusion NMR studies,16,20 the
spectator anion is expected to be tightly paired to the cationic Au(I)ꢀbound reaction intermediates and transition states in the
lowꢀε solvent toluene, whereas ion pairing is minimal in the higherꢀε solvent CH2Cl2.8 Hence, for toluene we calculated the
energies assuming an ion paired transition state while for CH2Cl2 we calculated the energies assuming an unpaired transition
state. When these energy differences are converted to product ratios and compared to the experimentally observed solventꢀ
dependent regioselectivities for several substrates (2a,d,h,jꢀk), very good qualitative (and in many cases semiꢀquantitative)
agreement is observed (Table 2). In particular, the differences between the calculated ꢁ(ꢁꢁE‡) and the ꢁ(ꢁꢁE‡) that
corresponds to the experimentally observed Ptoluene/PCH2Cl2 range from 0.1 to 1 kcal/mol, which are within the error expected
for DFT and on the order of the thermal energy kBT at 300 K. For example, as shown in Table S8, changing the density
functional used can change ꢁ(ꢁꢁE‡) by between 0.1 and 0.8 kcal/mol, but qualitative agreement is seen in all cases.
To understand the origins of the selectivity changes, it is particularly instructive to consider the two substituents of
contrasting polarity 2a (R = Cl) and 2h (R = Me). For both of these substrates, when hydroarylation is catalyzed by [1]SbF6
in CH2Cl2 and more polar solvents, poor regioselectivity is observed, with 4:3 ranging from 0.8:1 to 1.0:1. The weak
selectivity is reflected in the cationic transition state energies calculated in CH2Cl2, which show that in the absence of a
counterion, the competing regioisomeric transition states [3‡]+ and [4‡]+ exhibit comparable activation barriers (ꢁꢁE‡ ~ 0).
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
–
However, when the transition states are ion paired with SbF6 , the relative activation barriers are perturbed in a substituentꢀ
dependent manner, with the Clꢀsubstituted transition state [4a‡]SbF6 lower in energy than [3a‡]SbF6 by 2.3 kcal/mol. This
corresponds to a calculated 45ꢀfold increase in selectivity for 4a when the transition state is ion paired. This is in qualitative
agreement with the regioselectivity observed for the reaction of 2a in toluene, where strong ion pairing is expected, which
favors 4a over 3a by a factor of 5. The calculated ion paired transition states are optimized structures—i.e. the counterion
occupies the position that affords maximum stabilization of the cationic transition state and consequently maximum
electrostatic differentiation of the competing pathways. The fact that the observed selectivity change is smaller than what is
calculated may arise from incomplete ion pairing or nonꢀideal counterion positioning. In contrast to the Clꢀsubstituted case,
our DFT calculations do not indicate a significant energy difference between ion paired Meꢀsubstituted transition states
[3h‡]SbF6 and [4h‡]SbF6, (ꢁ(ꢁꢁE‡) ~ 0) in agreement with the experimentally observed lack of regioselectivity for the
substrate in solvents ranging in ε from 2.4 to 10.4 (Figure 2, Table S4). It is important to note that the lack of selectivity
change seen from this substrate arises from a cancellation of energies, where the energy of both unpaired complexes is
lowered upon ion pairing, but by the same amount, a result that is not obvious and arises from the calculations. Because the
energy is lowered by the same amount for both, no change in ꢁꢁE‡ is observed, in agreement with experiment.
The presence of a polar substituent does not guarantee εꢀdependent selectivity, as evidenced by the fact that substrates
with OMe, OAc, and OMs substituents (2j, 2k, 2l) behave similarly to 2a. Interestingly, replacing OMs with OTf (2d)
restores a strong εꢀdependence. Simulations for substrates with conformationally flexible substituents are complicated by the
presence of multiple transition state conformers with similar energies. Nonetheless, the calculations qualitatively recapitulate
the strong εꢀdependence for 2d and the weak dependence for 2j and 2k (Table 2).
Origins of the εꢀdependent selectivity. To understand why different substrates respond differently to changing ε, we
first mapped the electrostatic potential on a surface of constant electron density for the two regioisomeric transition states for
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 14
1
2
3
4
5
6
7
8
9
substrates 2a and 2h (Figure 3). From inspection of the electrostatic potential, the two Clꢀsubstituted transition states [3a‡]+
and [4a‡]+ have noticeably different charge distributions, with the more polar [4a‡]+ having greater localization of negative
charge around the Cl substituent. In contrast, the Meꢀsubstituted transition states [3h‡]+ and [4h‡]+ have very similar charge
distributions. The different selectivities of 2a and 2h can therefore be rationalized because the more polar charge distribution
of [4a‡]+ is better stabilized by the presence of the counterion SbF6– compared to [3a‡]+, while [3h‡]+ and [4h‡]+ respond
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
–
similarly to the presence of SbF6 . This explanation is consistent with the previous observation that ion pairing induces a
selectivity change for a sulfoxide rearrangement if the two selectivityꢀdetermining transition states have very different charge
distributions, which was roughly estimated by calculating the dipole moments.8
The electrostatic potential surfaces are similar for the transition states corresponding to 2j, as expected given that the
calculated ꢁ(ꢁꢁE‡) ~ 0 for this substrate (Figure S2). In contrast, large differences are evident in the electrostatic potential
surfaces for the transition states corresponding to 2d, for which ꢁ(ꢁꢁE‡) is substantial. Interestingly, the surfaces for
transition states corresponding to 2k also show substantial differences, despite the fact that ꢁ(ꢁꢁE‡) is small for this
substrate. These results underscore the limitations of analyzing charge densities alone when calculating the selectivity
change.
Figure 3. Electrostatic potential surfaces of unpaired transition states leading to regioisomeric products 3 and 4 for substrates
2a and 2h. The electrostatic potential is shown on a surface of constant electron density of 0.0004 e/Å3.
While ε clearly determines the propensity for ion pairing, it could also affect the strength of electrostatic stabilization
within the ion pair. In other words, the interaction between the ion paired complex and the polarizable dielectric medium of
the solvent may be a contributor to the εꢀdependence of the selectivity. In order to assess the role of these interactions, we
calculated ꢁꢁE‡ for cationic and ion paired transition states in toluene (low ε), CHCl3 (intermediate ε), and CH2Cl2 (large ε)
for 2a and 2h (Table 3). For the ion paired transition states of 2a, ꢁꢁE‡ exhibits strong εꢀdependence, indicating that
interactions with the solvent play a key role in driving εꢀdependent selectivity for this substrate. In particular, while the ion
paired transition state [4a‡]SbF6 is more stable than [3a‡]SbF6 by 2.30 kcal/mol in lowꢀε solvent (toluene), both ion paired
10
ACS Paragon Plus Environment

Page 11 of 14
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
transition states have similar energies in highꢀε solvent (CH2Cl2). This effect arises because the highꢀε solvent attenuates the
electrostatic interactions between the transition states and the counterions that give rise to ꢁꢁE‡ion pair. This result also
indicates that even if ion pairing could be enforced in highꢀε solvents it would be expected to have little or no effect on
selectivity.
9
Unlike the ion paired transition states, the unpaired transition states [3a‡]+ and [4a‡]+ have essentially the same energy in
all solvents. The insensitivity of ion unpaired transition states to ε is consistent with experimental results for the reaction of
2a catalyzed by [1]B(ArF)4, for which the regioselectivity shows essentially no εꢀdependence. Although diffusion NMR
suggests that [1]B(ArF)4 is ion paired in lowꢀε solvents,8 the large radius of B(ArF)4– attenuates the electrostatic interactions in
the ion pair, which results in weak discrimination between the different charge distributions of the cationic transition states.
In other words, [3a‡]B(ArF)4 and [4a‡]B(ArF)4 behave like unpaired transition states. For 2h, ꢁꢁE‡ for both the paired and
unpaired transition states is effectively independent of ε, indicating that the charge distributions of these two transition states
are similar enough that the polarizable dielectric stabilizes both by the same amount. Overall, the results in Table 3 suggest
that the role of the solvent in changing selectivity is not solely to enforce ion pairing, but also that reducing ε strengthens the
electrostatic interactions in the ion pair that energetically differentiate transition states with large differences in charge
distribution.
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. Solvent influence on ꢁꢁE‡unpaired and ꢁꢁE‡
ion pair
R = Cl (2a)
unpaired
R = Me (2h)
unpaired
Solventa ꢁꢁE‡
ꢁꢁE‡
ꢁꢁE‡
ꢁꢁE‡
ion pair
ion pair
[kcal/mol] [kcal/mol] [kcal/mol] [kcal/mol]
Toluene
CHCl3
CH2Cl2
0.19
0.03
–0.07
–2.30
–1.10
–0.45
–0.42
–0.34
–0.30
–0.13
–0.14
–0.19
a Modeled using IEFꢀPCM implicit solvent
Effect of ion position on catalytic selectivity. DFT optimizations of the regioisomeric transition states [3a‡]SbF6 and
[4a‡]SbF6 find the lowest energy ion pair geometry when the anion is positioned adjacent to the surface of greatest positive
charge density on the cation. However, even if there is only one dominant minimum energy ion position, there may be other
similar structures that are close in energy and hence relevant to catalytic selectivity. In order to assess the importance of such
structures, we performed a relaxed scan along the Au–Sb interatomic distance, in which we held the positions of the
transition state atoms fixed while allowing the anion position to relax as it is moved away from the transition state. We then
calculated the effect of these displacements on the relative barrier heights for the two competing reactions (Figure 4). These
calculations show that the lowest energy ion pairs [3a‡]SbF6 and [4a‡]SbF6 in toluene exhibit Au–Sb distances of 5.1 Å and
4.4 Å, respectively, with the favored product having a closer Au–Sb distance at its minimum energy structure. Thus, the
greater polarization of [4a‡]+ results in a shorter ion–ion separation in the ion pair, which is responsible for much of the
ꢁꢁE‡. Furthermore, because the two curves in Figure 4 have similar shapes, these calculations also show that ion pairing
induced selectivity is relatively insensitive to small perturbations in the ion–ion distances. For example, when both structures
are at their minima, ꢁꢁE‡ = –2.2 kcal/mol. When both structures are displaced from their minima by 1 Å (at an energetic cost
of ~2 kcal/mol for both), ꢁꢁE‡ remains essentially constant at –2.5 kcal/mol.
11
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 14
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 4. Effect of ion pairing distance on difference in activation barrier heights for [3a‡]SbF6 and [4a‡]SbF6. Differences in
relative activation barriers (ꢁꢁE‡) are given for the lowest energy geometry and at +1 Å displacement. Barrier heights were
found by a scan in which the positions of the transition state atoms were fixed and the ion orientation was allowed to relax at
a fixed Au–Sb distance.
CONCLUSIONS
In summary, we have demonstrated that ion pairing of a cationic Au(I) catalyst to a small spectator counterion can change the
regioselectivity of intramolecular hydroarylation by a factor of up to 12 depending on the substrate structure. By modulating
the strength of electrostatic interactions, 5ꢀ and 7ꢀsubstituted 2Hꢀchromenes can be accessed regioselectively without the
need for significant substrate bulk or directing groups. Calculations of transition state energies in paired and unpaired forms
indicate that this effect arises from electrostatic interactions between the transition state charge distributions and the
counterion. Transition states that have a greater degree of charge separation are better stabilized by the proximal counterion.
The energy difference of ion paired transition states diminishes rapidly as ε is increased even if ion pairing is preserved
because of interactions between the ion pair and the dielectric medium. In contrast, the energies of unpaired transition states
show very little εꢀdependence, which explains why reaction selectivity is not dependent on solvent ε when large, diffuse
counterions are used. Using ion pairing to electrostatically differentiate competing transition states is in principle applicable
to any reaction where ionic, productꢀdetermining transition states have substantially different charge distributions. While our
results show that ~1–1.5 kcal/mol perturbation is readily achievable with an ion pair in a lowꢀε medium, accessing larger
12
ACS Paragon Plus Environment

Page 13 of 14
Journal of the American Chemical Society
1
2
3
4
5
perturbations may require coꢀlocalizing multiple ions, or designing catalysts/substrates that react via transition states with
more divergent charge distributions.
6
7
8
ASSOCIATED CONTENT
9
Supporting Information
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
The Supporting Information is available free of charge on the ACS Publications website.
Experimental details, tables, figures, optimized Cartesian coordinates and structural characterization (PDF).
AUTHOR INFORMATION
Corresponding Author
*mkanan@stanford.edu
ACKNOWLEDGEMENTS
We thank the Air Force Office for Scientific Research (FA9550ꢀ11ꢀ1ꢀ0293) for funding. V.M.L acknowledges the Natural
Sciences and Engineering Research Council of Canada for support through the PGSꢀD program. T.E.M was supported by the
U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences (DEꢀSC0014437). T.E.M also acknowledges
support from a Cottrell Scholarship from the Research Corporation for Science Advancement. W.C.P acknowledges support
from the Melvin and Joan Lane Stanford Graduate Fellowship. We also thank Stanford University and the Stanford Research
Computing Center for providing computational resources and support.
REFERENCES
(1) Starks, C. M.; Liotta, C. L.; Halpern, M. PhaseꢁTransfer Catalysis; Chapman & Hall: New York, NY, 1994.
(2) Denmark, S. E.; Gould, N. D.; Wolf, L. M. J. Org. Chem. 2011, 76, 4337.
(3) Brak, K.; Jacobsen, E. N. Angew. Chem. Int. Ed. 2012, 52, 534.
(4) Mahlau, M.; List, B. Angew. Chem. Int. Ed. 2013, 52, 518.
(5) Delferro, M.; Marks, T. J. Chem. Rev. 2011, 111, 2450.
(6) Jia, M.; Bandini, M. ACS Catal. 2015, 5,1638.
(7) Trinchillo, M.; Belanzoni, P.; Belpassi, L.; Biasiolo, L.; Busico, V.; D’Amora, A.; D’Amore, L.; Del Zotto, A.; Tarantelli,
F.; Tuzi, A. Organometallics 2016, 35, 641.
(8) Lau, V. M.; Gorin, C. F.; Kanan, M. W. Chem. Sci. 2014, 5, 4975.
(9) Menon, R. S.; Findlay, A. D.; Bissember, A. C.; Banwell, M. G. J. Org. Chem. 2009, 74, 8901.
(10) Lykakis, I. N.; Efe, C.; Gryparis, C.; Stratakis, M. Eur. J. Org. Chem. 2011, 2011, 2334.
(11) Arcadi, A.; Blesi, F.; Cacchi, S.; Fabrizi, G.; Goggiamani, A.; Marinelli, F. Org. Biomol. Chem. 2012, 10, 9700.
(12) Majumdar, N.; Paul, N. D.; Mandal, S.; de Bruin, B.; Wulff, W. D. ACS Catal. 2015, 5, 2329.
(13) Ding, D.; Mou, T.; Feng, M.; Jiang, X. J. Am. Chem. Soc. 2016, 138, 5218.
(14) Nevado, C.; Echavarren, A. M. Chem. Eur. J. 2005, 11, 3155.
(15) Tang, D.; Chen, Z.; Zhang, J.; Tang, Y.; Xu, Z. Organometallics 2014, 33, 6633.
(16) Pregosin, P. S. Prog. Nucl. Magn. Reson. Spec. 2006, 49, 261.
(17) Malhotra, D.; Mashuta, M. S.; Hammond, G. B.; Xu, B. Angew. Chem. Int. Ed. 2014, 53, 4456.
(18) Knapp, D. M.; Gillis, E. P.; Burke, M. D. J. Am. Chem. Soc. 2009, 131, 6961.
(19) Tunge, J. A.; Foresee, L. N. Organometallics 2005, 24, 6440.
(20) Kumar, P. G. A.; Pregosin, P. S.; Goicoechea, J. M.; Whittlesey, M. K. Organometallics 2003, 22, 2956.
13
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 14
1
2
3
4
Graphic Abstract
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
14
ACS Paragon Plus Environment