Palladium-catalyzed redox cascade for direct β-arylation of ketones

Article abstract of DOI:10.1016/j.tet.2018.03.017

Herein we report a full article about the detailed design and development of two palladium-catalyzed redox cascade methods that enable direct β-arylation of ketones. Palladium-catalyzed ketone dehydrogenation, aryl-X bond activation and conjugate addition were merged into a redox-neutral catalytic cycle. Non-metal-based aryl electrophiles were used as both the oxidant and the aryl source. The β-arylation with aryl iodides was achieved site-selectively with Pd(TFA)₂/P(i-Pr)₃ as the precatalyst and AgTFA as the iodide scavenger. Both cyclic and linear ketones can react to give β-aryl ketones with excellent functional group tolerance. The β-arylation with diaryliodonium salts was realized without stoichiometric heavy metal additives, and proved to be redox-neutral. A wider substrate scope regarding aryl groups and ketones was obtained for the arylation with diaryliodonium salts, and the possible involvement of palladium nanoparticles as the active catalyst was examined and discussed.

Full text of DOI:10.1016/j.tet.2018.03.017

Accepted Manuscript
Palladium-catalyzed redox cascade for direct β-arylation of ketones
Zhongxing Huang, Guangbin Dong
PII:
S0040-4020(18)30260-6
10.1016/j.tet.2018.03.017
TET 29357
DOI:
Reference:
To appear in: Tetrahedron
Received Date: 12 January 2018
Revised Date: 6 March 2018
Accepted Date: 7 March 2018
Please cite this article as: Huang Z, Dong G, Palladium-catalyzed redox cascade for direct β-arylation of
ketones, Tetrahedron (2018), doi: 10.1016/j.tet.2018.03.017.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT
Palladium-Catalyzed Redox Cascade for Direct β-Arylation of Ketones
Zhongxing Huang and Guangbin Dong
Department of Chemistry, University of Chicago, Chicago, Illinois 60637, United States
1 Introduction
Reactions of ketones have been cornerstones in organic chemistry.1 Classic transformations of ketones hinge on
the inherent acidity of the α-C−H bonds and the electrophilicity of the ipso carbon. Conventionally, the β-C−H
bonds of ketones are considered inert due to their remote positions to the carbonyl group, and thus less likely to
be directly functionalized (Scheme 1A).2
Scheme 1. Functionalization of Ketones
Nevertheless, β-substituted ketones are highly sought after as they are versatile synthetic intermediates and
prevalent in bioactive compounds (Scheme 1B).3 For more than a century, conjugate addition of metal-based
nucleophiles (e.g. R-MgX, R-Li) to α,β-unsaturated ketones has been widely used to synthesize ketones with β-
substituents (Scheme 2).4 Continuous efforts from the synthetic community have led to a number of efficient
catalytic and non-catalytic systems that enable the conjugate addition of a wide range of nucleophiles and enone
acceptors. We envision that direct β-functionalization methods that employ saturated ketones as the substrate
would offer an attractive alternative to the conjugate addition reactions considering the generally higher
availability and lower cost of saturated ketones compared with the corresponding enones.5 It is also desirable to
avoid the use of basic and nucleophilic metal-based reagents for the β-functionalization reaction, as they often
require stoichiometric reductant to synthesize and compromise the tolerance of certain functional groups, such as
carbonyl groups and acidic protons. Herein, we present a detailed account of the design, development, and
mechanistic understanding of a palladium-catalyzed redox cascade that enables the direct β-arylation of saturated
ketones using readily available and non-metal based aryl sources.
1

ACCEPTED MANUSCRIPT
Scheme 2. Direct β-Functionalization of Saturated Ketones with Non-Metal Aryl Sources
2 Design of a palladium-catalyzed redox cascade
We conceived the idea of using saturated ketones and aryl halides as the substrates for the direct β-arylation
reaction by merging palladium-catalyzed ketone dehydrogenation, C−X (X: halides) bond activation, and
conjugate addition into a single catalytic cycle (Scheme 3). Since the seminal discovery by Theissen,6 palladium-
mediated dehydrogenation has been widely employed to prepare α,β-enones from saturated ketones.7, 8 The
mechanism of this transformation has been studied and is proposed to proceed through a sequence of palladium
enolate formation (Step A), β-hydride elimination (Step B), and reductive elimination to give the enone product
and a Pd(0) species (Step C). A variety of catalytic dehydrogenation has been developed based on this
mechanism by using different stoichiometric oxidants to regenerate the active Pd(II) catalyst from the Pd(0)
intermediate. Here we hypothesize that aryl halides can serve as a stoichiometric oxidant for the catalyst
regeneration as the oxidative addition of Pd(0) to aryl halides is well established (Step D).9 It could also be
envisaged the resulting Pd(II)-enone complex would undergo further migratory insertion to give Pd-enolate (Step
E), which upon protonation, would yield the β-arylation product and release the Pd(II) catalyst (Step F). The
migratory insertion and protonation steps have been established as key components of the mechanism of a large
number of palladium-catalyzed conjugate addition reactions.4b A key feature of the designed catalytic cycle is the
use of aryl halides as both the stoichiometric oxidant for the dehydrogenation step, and the aryl source for the
conjugate addition step. Overall, saturated ketones and aryl halides serve as a redox pair for the β-arylation
reaction, which in principle should be redox-neutral and requires no external oxidant or reductant.
Scheme 3. Proposed Palladium-Catalyzed Redox Cascade
While the constituent ketone dehydrogenation and conjugate addition are supported by a large number of
precedents, we were aware of several challenges associated with the merged catalytic cycle. First,
dehydrogenation of ketones often requires an electrophilic palladium catalyst to facilitate the ketone coordination
2

ACCEPTED MANUSCRIPT
and subsequent palladium enolate formation. 7g However, oxidative addition of C
−X (X: halides) bonds prefers
an electron-rich palladium complex.9 Thus the choice of palladium catalysts for the merged redox cascade is non-
trivial. Second, examples from prior literatures indicate several side reactions can take place, and thus need to be
inhibited during the optimization of the β-arylation reaction. It is known that aryl halides can dimerize to biaryls
in the presence of palladium catalysts and stoichiometric reductant (Scheme 4A).10 Over-oxidation of the ketone
substrates and/or the β-aryl ketone products should also be minimized, as palladium-catalyzed sequential
dehydrogenation of ketones11 (Scheme 4B) and dehydrogenative β-arylation of ketones12 (Scheme 4C) have been
reported. Third, carboxylate-type ligands have been proved important to maintain the reactivity of Pd(II)
catalysts for the ketone dehydrogenation (Scheme 4D).13 Nevertheless, during the oxidative addition step in our
proposed catalytic cycle, a halide ligand is transferred to the palladium catalyst, which would presumably
deactivate the catalyst for the next dehydrogenation. Therefore, efficient halide scavengers must be found to
restore the palladium carboxylate catalyst by replacing the halide ligand.
Scheme 4. Precedents of Possible Side-reactions and Requirement for Halide Scavenger
3 β-Arylation with aryl iodides
3.1 Optimization of conditions
Guided by the proposed redox cascade, we were delighted to find the β-arylation product from the coupling
between cyclohexanone and iodobenzene with Pd(TFA)2/DMSO (palladium trifluoroacetate/dimethylsulfoxide)
as the precatalyst (Scheme 5, Eq. 4). Our early screening of the reaction parameters, including precatalysts,
additives and solvents, has revealed several preferences of the β-arylation reaction. First, silver salt can greatly
promote the reaction. As designed in the catalytic cycle, the role of silver salt is to scavenge the iodide ligand and
restore the active Pd(II) catalyst. Second, acetate or trifluoroacetate counter anion is indispensable for the
3

ACCEPTED MANUSCRIPT
reaction. In the catalytic cycle, when the silver scavenges iodide from the palladium center, it also delivers its
counter anion to the palladium. As the acetate-type ligand on palladium is crucial to maintain the palladium
catalyst sufficiently electrophilic, silver trifluoroacetate is favored for the reaction. Second, acidic and
coordinating solvents are both important for the reaction, as a mixed solvent consisting of equal volumes of
dioxane and trifluoroacetic acid was found beneficial for the arylation reaction. It is anticipated that dioxane can
stabilize the palladium intermediates as a weak ligand, and acidic reaction medium is proposedly critical for the
last protonation step (vide supra, Scheme 3, Step F). Fourth, although different sorts of ligands (i.e. phosphine-
and nitrogen-based ligands) are screened, they seemed to have undifferentiated effects on the reaction outcome.
We hypothesize that due to the use of strongly acidic trifluoroacetic acid, a vast majority of phosphine, phosphite
or nitrogen-based ligand can be protonated, and thus sluggish to coordinate to the palladium center. In addition,
further optimization of other reaction parameters, including additives and solvents, did not turn out to be fruitful.
Scheme 5. Preliminary Conditions for β-Arylation with Aryl Iodides
The protonation issue was largely alleviated when HFIP (hexafluoro-2-propanol) was used as the co-solvent
instead of trifluoroacetic acid (Scheme 6, Eq. 5). Given that the pKa of HFIP is only 9.3, it is acidic enough to
pronate enolates, but not acidic enough to protonate the majority of phosphine- and nitrogen-based ligands.
Scheme 6. Use of HFIP as Co-solvent
While HFIP initially didn’t lead to a higher yield than trifluoroacetic acid, with HFIP as a co-solvent for the
ligand screening, the differences among various types of ligands were revealed to be significant (Scheme 7).
While simple pyridine was not an effective ligand (entry 1), bis-nitrogen-based ligands generally promoted the
reaction (entries 2-5). Among these bidentate ligands, electron-deficient 4,5-diazafluoren-9-one14 (entry 4) and
2,2’-dipyridyl ketone (entry 5) gave the highest yields, probably due to their π-acidity that facilitates the
dehydrogenation. It is noteworthy that simple triphenylphosphine (entry 7) outperformed DMSO (entry 6) to
deliver the β-arylation product in 42% yield. Further screening of phosphine ligands with a wide range of aryl or
alkyl substitutions demonstrated electron-rich trialkylphosphine, especially PCy3 (entry 14) and P(i-Pr)3 (entry
15), are the best candidates for the reaction, while most bidentate phosphine ligands gave very low conversion
(entries 18-21).
4

ACCEPTED MANUSCRIPT
Scheme 7. Ligand Effect of β-Arylation with Aryl Iodides
The ratio between two reactants also turned out to be crucial for the efficiency of the β-arylation reaction
(Scheme 8). While reactions with excess iodobenzene led to yields around 50%, a reverse ratio of substrates
resulted in a much improved yield (76%). We attributed the requirement for excess ketone substrates to the
challenging dehydrogenation step, as the Lewis basicity of ketones is relatively low and thus they are sluggish to
complex with palladium for dehydrogenation. These observations were also supported by Stahl’s mechanistic
5

ACCEPTED MANUSCRIPT
studies7g of the palladium-catalyzed aerobic dehydrogenation of ketones, where higher concentration of ketone
substrates leads to higher initial rates.
Scheme 8. Effect of the Ratio of Substrates
Upon reaching the optimal reaction conditions (Scheme 9, Eq. 6), detailed analysis of the reaction showed the
major side-product of this transformation was the biphenyl resulted from the reductive homo-coupling of
iodobenzene (vide supra, Scheme 4A). α-Arylation (2) and dehydrogenative β-arylation (3) products were not
observed. It is worthy to mention that similar combinations of palladium catalysts and electron-rich phosphine
ligands are frequently used in the Buchwald-Hartwig-Miura α-arylation of ketones (Eq. 7).15 We proposed that
the different pathway and site-selectivity observed here roots from different reaction medium employed: strong
bases are used in the α-arylation reaction, whereas this β-arylation employs acid conditions (HFIP and
trifluoroacetic acid16).
Scheme 9. Optimized Conditions for β-Arylation with Aryl Iodides
3.2 Substrate scope
Several features regarding the substrate scope of the direct β-arylation reaction are noteworthy. The palladium-
catalyzed redox cascade shows a broad scope of aryl iodides. Aryl iodides with different substitution patterns
(ortho, meta, or para substitution) as well as different electronic properties (electron-neutral, -rich, and -poor) all
reacted to give the desired β-aryl ketones under the reaction conditions (Scheme 10A). Besides, we were
delighted to discover that many sensitive functional groups, such as ketones, aldehydes, and acidic protons, are
compatible with the reaction conditions (Scheme 10B). These functional groups, on the other hand, are often not
tolerated in conjugate addition reactions when using strongly basic and nucleophilic metal-based reagents (e.g.
Grignard and organolithium reagents). The tolerance of these base- and nucleophile-sensitive functional groups
is probably attributed to the acidic reaction medium in the absence of stoichiometric aryl-metal species. However,
it should be pointed out that the acidic reaction medium and in situ generated trifluoroacetic acid during the
course of the reaction has compromised the compatibility of some acid-labile functional groups, such as silyl-
protected alcohol (4.1) and acetate (4.2) (Scheme 10C). In addition, with the electrophilic nature of the palladium
catalyst, aryl halides with Lewis-basic (4.3-4.5) or electron-rich (4.6) heterocycles, as well as unsaturated C−C
bonds (4.7-4.8) failed to give the β-arylation products, probably due to the catalyst poisoning and side-reactions
of alkenes/alkynes.
6

ACCEPTED MANUSCRIPT
Scheme 10. Advantage and Limitation of the Scope of Aryl Iodidesa
Aryl bromides exhibit much lower reactivity (Scheme 11). While methyl 4-bromobenzoate and simple
bromobenzene gave low conversions, 3,5-dimethoxybromobenzene afforded the desired β-arylation product in a
moderate yield. It is proposed that the major obstacle to achieving efficient β-arylation with aryl bromides is their
slower oxidative addition compared with iodides. Without a rapid oxidative addition, the Pd(0) intermediate is
prone to decomposition to palladium black through aggregation in the absence of excess ligands. It was found
that addition of more phosphine ligands inhibited the ketone dehydrogenation step likely by blocking the
coordination site for β-H elimination. The use of bidentate phosphine ligands may increase the rate of oxidative
addition of Pd(0) and prevent decomposition. However, they are not effective for the ketone dehydrogenation
step. Thus, the key to address the challenge of using ArBr as the coupling partner in the future is to find a
suitable ligand that can stabilize the Pd(0) intermediate but meanwhile not interfere with the dehydrogenation
step.
7

ACCEPTED MANUSCRIPT
Scheme 11. β-Arylation with Aryl Bromides Using Pd(TFA)2/P(i-Pr)3
The scope of ketones was examined next (Scheme 12). The β-arylation reaction appears to be highly
diastereoselective when cyclohexanones with substituents at the C4 position were submitted to the reaction,
giving the trans products (5.2, 5.3). Nevertheless, we were disappointed to find that substituents at the C2 or C3
position (6.1, 6.2), as well as cyclohexanones with a heteroatom in the skeleton (6.3), are unsuitable substrates
for the β-arylation reaction. Reactions with these ketones often resulted in a low conversion of ketones,
indicating a difficult dehydrogenation step. On the other hand, simple cyclic ketones with ring-sized other than
cyclohexanones (5.4, 5.5) succeeded to give the β-arylation product. In addition, acyclic ketones are also able to
participate in the β-arylation reaction (5.5-5.7). Intriguingly, propiophenone gave a large amount of diarylation
product, which was not observed for the cyclic ketones. The diarylation product was proposed to be accessed
from intermediate 7.1 after the first conjugate addition, instead of a separate β-arylation of the monoarylated
product, as the β-arylation of β-phenyl propiophenone only gave a trace amount of diarylated product. We
hypothesized that, due to the more flexible conformation and rotation of β C−C bond, the second β-hydrogen
elimination of 7.2 is facile to take place, followed by conjugate addition and protonation to give the diarylation
product.
8

ACCEPTED MANUSCRIPT
Scheme 12. Scope and Limitation of Ketone Substratesa
3.3 Silver issue
The use of stoichiometric AgTFA in our conditions was the key to avoid poisoning of halide and restore the
active Pd(II) dicarboxylate catalyst. However, the disadvantages of using silver salts are, first, they are expensive
and thus impractical on a process scale; second, they have cast some doubts on the reaction mechanism. In our
proposed catalytic cycle, the saturated ketone and aryl halide serve as a redox-pair for the β-arylation reaction,
rendering the catalysis redox-neutral (Scheme 13A). Nevertheless, it is known that Ag(I) salts are able to oxidize
Pd(0) species. Thus, an alternative mechanism of the β-arylation reaction may involve two separate
transformations, the palladium-catalyzed dehydrogenation and reductive Heck reaction (Scheme 13B). In this
alternative mechanism, the Ag(I) additive serves as an stoichiometric oxidant for the palladium-catalyzed
dehydrogenation (Cycle I), and the resulting enone enters the second catalytic cycle to undergo the palladium-
catalyzed reductive Heck reaction (Cycle II), where excess ketones or solvents could serve as the reductant.
Therefore, instead of being redox-neutral, the β-arylation with aryl halides requires both external oxidants and
reductants if undergoing the ‘two-cycle’ mechanism. When stoichiometric silver salt is employed, the “two-cycle”
mechanism cannot be excluded.
9

ACCEPTED MANUSCRIPT
Scheme 13. Comparison between Redox-neutral cycle and ‘Two-cycle’ Mechanism
4 Arylation with diaryliodonium salts
One indirect approach to solve the silver issue is to use aryl electrophiles that do not transfer a halide anion to
palladium during the oxidative addition, thus avoiding the use of a halide scavenger at all. Towards this end, we
found the diaryliodonium salt as a promising aryl source for the β-arylation without halide scavenger (Scheme
14).17 Unlike aryl iodides, the oxidative addition of diaryliodonium salts to Pd(0) transfers an aryl and a non-
halide anion to the metal, and release an molecule of aryl iodide at the same time.18 As a result, silver salt is not
necessary to extract the iodide ligand and regenerate the active palladium catalyst any more. Besides the silver
issue, the use of diaryliodonium salts would also address the limitation of using air-sensitive P(i-Pr)3 ligand
under our previous β-arylation conditions, which necessitates air-free operations and highly purified reagents. As
a more reactive aryl electrophile, oxidative addition of diaryliodonium salts generally does not need the
assistance of electron-rich phosphine ligand. Thus, the new reaction conditions using diaryliodonium salts as the
aryl source is expected to tolerate air.
Scheme 14. Oxidative Addition of Diaryliodonium Salts to a Pd(0)-Enone Complex
4.1 Optimization of conditions
Indeed, the use of diphenyliodonium triflate as the aryl source led to a new set of β-arylation conditions that
require no stoichiometric heavy metals as a halide scavenger, where Pd(OAc)2 (palladium acetate)/DMSO serves
as the precatalyst (Scheme 15A, Eq. 8). However, one concern about the use of diphenyliodonium salt in our
preliminary conditions is that it can act as two equivalents of aryl source, as the oxidative addition of the salt to
Pd(0) will produce one molecule of iodobenzene as a byproduct (Scheme 15B). Besides the complication of
stoichiometry, the iodobenzene byproduct could potentially poison the palladium catalyst by undergoing
oxidative addition with Pd(0) intermediate in the reaction. In order to avoid such an issue, we switched to
mesitylphenyliodonium triflate as the aryl source. The rationale for the use of mesityl-based iodonium salts is
two-fold. First, mesitylaryliodonium salts are known to transfer the less sterically hindered aryl group
chemoselectively due to the bulk of mesityl group. Second, also due to the sterics, the byproduct from the aryl
transfer, iodomesitylene, is hard to undergo oxidative addition to the Pd(0) intermediate, thus avoiding poisoning
the palladium catalyst. As expected, replacement of diphenyliodonium salt with its mesityl counterpart resulted
in improved efficiency of the β-arylation reaction under the same conditions (Eq. 9).
10

ACCEPTED MANUSCRIPT
Scheme 15. Preliminary Conditions for the β-Arylation with Iodonium Salts
In terms of the choice of ligands, the new reaction conditions with iodonium salts are distinct from our previous
Pd(TFA)2/P(i-Pr)3 system (Scheme 16). While phosphine and phosphite ligands still afforded the β-arylation
product (entry 1-2), the yields were generally low, despite their outstanding performances for the previous
conditions (vide supra, Scheme 7). Nitrogen-based ligands (either mono- or bidentate) were not suitable for the
reaction with diaryliodonium salts (entry 4-9); the only exception being 4,5-diazafluoren-9-one that gave the
product in 56% yield (entry 8). On the other hand, sulfur-based ligands are more favored for the new conditions,
with simple DMSO outcompeting phosphine- and nitrogen-based ligands (entry 10).
11

ACCEPTED MANUSCRIPT
Scheme 16. Effects of Different Types of Ligands
Subsequently, a large variety of sulfide- and sulfoxide-based ligands with different backbones was synthesized
and examined (Scheme 17). While monodentate sulfide and sulfoxide ligands all gave the β-arylation product in
good yields (entry 1-7), they are generally outperformed by their bidentate derivatives. Especially, 1,2-
bis(phenylthio)ethane (8.1), as well as its sulfoxide counterpart (8.2),19 originally employed by White and
coworkers for the allylic C−H activation, proved to be a superior ligand for the reaction. Increasing the bite angle
of the bis-sulfide ligand (entry 10) or using a conformationally fixed derivative (entry 11) decreases the
efficiency of the reaction; eliminating one carbon on the backbone of the ligand (entry 12) totally shut down the
reaction. Replacing the phenyl groups on the ligand with alkyl groups resulted in lower yields (entry 13), and the
use of an electron-donating tert-butyl-substituted bis-sulfide ligand (entry 14) gave no β-arylation product,
probably by inhibiting the dehydrogenation step.
12

ACCEPTED MANUSCRIPT
Scheme 17. Effects of Different Sulfide and Sulfoxide Ligands
Based on the backbone of 8.1 and 8.2, we modified the aryl substituents to examine their effects on the reaction
efficiency (Scheme 18). However, alternating the electronic properties of the arenes seemed to have little effects
on the reaction; ligands with either electron-donating or -withdrawing substituents gave comparable or lower
yields than 8.2. Increasing the steric hindrance of the sulfur ligands (entry 6 and 10), however, only afforded the
product in a lower or similar yield.
13

ACCEPTED MANUSCRIPT
Scheme 18. Ligand Optimization Based on 8.1 and 8.2
Finally, several sulfilimine ligands were synthesized, examined, and compared with sulfide and sulfoxide ligands
for the β-arylation reaction (Scheme 19). To our delight, bis-N-tosylsulfilimine ligand 9 turned out to be superior,
giving a higher yield than its sulfide and sulfoxide counterparts (entry 2 and 3). This ligand was easily prepared
in one step from 1,2-bis(phenylthio)ethane and Chloramine-T, and the meso and racemic ligand demonstrated
nearly identical reactivity. Although sulfides and sulfoxides are frequently employed as ligands, to the best of our
knowledge, the class of bis-sulfilimines has not been previously used as ligands for transition metals.
Monodentate sulfilimine ligands (entry 4-6) and the bis-sulfilimine ligand with an elongated backbone (entry 7)
were all found inferior.
14

ACCEPTED MANUSCRIPT
Scheme 19. Effects of Sulfilimine Ligands
Besides the ligand effect, several other features of this new generation of β-arylation are noteworthy. First, a
combination of KTFA (potassium trifluoroacetate) and TFA (trifluoroacetic acid) was employed (Scheme 20A).
While acidic medium is generally required for the final protonation of the palladium enolate to release the
product, the role of KTFA here is proposed to prevent the acidity of the reaction from going too high by
neutralizing strong acids generated (i.e. triflic acid and its equivalent). This hypothesis was also supported by the
detrimental effect of added strong acids (i.e. triflic acid). Third, partly due to the absence of electron-rich
phosphine ligands, the new reaction conditions tolerate air/moisture and can be set up without a glovebox or any
Schlenk techniques (Scheme 20B). Furthermore, the reactivity of the new conditions can be sustained under
lower temperatures with an elongated reaction time (Scheme 20C).
15

ACCEPTED MANUSCRIPT
Scheme 20. Optimized Conditions for the β-Arylation with Mesitylphenyliodonium Salts
4.2 Redox count
More importantly, as no redox-active additives were used under the new conditions, a clear analysis of the redox
property of the β-arylation can be executed (Scheme 21). A detailed identification and quantification of products
and byproducts of the new β-arylation reaction was carried out. For the cyclohexanone part, while excess ketone
was necessary for a fast reaction initiation, a high recovery of the unreacted ketone was obtained. Only a trace
amount of cyclohexen-1-one was formed, and no phenol from the over-oxidation was observed. For the
mesitylphenyliodonium salt part, there are two major reaction pathways. About three quarters of the iodonium
salts participated in the β-arylation pathways to give the desired β-arylated ketone with iodomesitylene as the
byproduct. Most of the remaining iodonium salt proceeded through a decomposition pathway to mesitylene and
iodobenzene. Altogether, the substrates being oxidized and reduced under the palladium catalysis are nearly of
the equal amount, thus supporting that the reaction between cyclohexanone and mesitylphenyliodonium salt is
indeed redox-neutral.
Scheme 21. Redox Property of the β-Arylation with Mesitylphenyliodonium Salts
4.2 Substrate scope
When a wide range of mesitylaryliodonium salts were subjected to the optimized conditions, a similar scope of
aryl groups was found compared with the Pd(TFA)2/P(i-Pr)3 system. Besides aryl groups with different electronic
properties and substitution patterns (Scheme 22A), the tolerance of functional groups that are hard to survive
under conjugate addition conditions, e.g. aldehyde and a second ketone, was also observed (11.1 and 11.2).
Moreover, it is interesting to note that aryl bromide (11.3), a functional group that was not compatible with β-
arylation using P(i-Pr)3, can be tolerated under the new conditions, probably due to the less electron-rich
palladium catalyst used. The iodonium salt with an electron-rich thiophenes moiety also reacted to give the
corresponding arylation product (11.4), while the previous method did not show tolerance of heterocycle
substrates. The high chemoselectivity was also demonstrated in the reaction with an estrone-derived iodonium
salt, where the cyclohexanone was selectively arylated in the presence of the cyclopentanone moiety in estrone
(11.5).
16

ACCEPTED MANUSCRIPT
Scheme 22. Scope of Aryl Groups
Both linear (12.10 and 12.11) and cyclic ketones with different ring sizes (12.1 and 12.2) participated in the β-
arylation with iodonium salts (Scheme 23). A significant improvement in terms of the ketone scope was that
cyclohexanones with C2-, C3-, or C4-substituents, including sterically demanding 2,2-dimethylcyclohexanone,
all reacted to give the corresponding arylated ketones (12.3-12.8), whereas the previous method only
accommodates C4-substituted cyclohexanones. Additionally, the previously incompatible 4-piperidinone
derivatives are suitable substrates using the new method (12.9). We hypothesized that the extended ketone scope
is attributed to a more cationic palladium catalyst generated during the reaction. While trifluoroacetate stayed as
the counter anion for the palladium catalyst throughout the reaction in the Pd(TFA)2/P(i-Pr)3 system, oxidative
addition of the iodonium salts in the new reaction would transfer less-coordinating anions (i.e. OTf) to the
palladium center, which in turn, would facilitate the ketone dehydrogenation step.
17

ACCEPTED MANUSCRIPT
Scheme 23. Improved Scope of Ketonesa
4.3 Mechanistic studies
Monitoring of the reaction progress by gas chromatography revealed an induction period for the β-arylation
reaction, during which an opaque and dark red solution was formed. Careful analysis of the reaction mixture
during the induction period also indicated the full conversion of bis-sulfilimine ligand 9 to give a pair of
elimination products 13.1 and 13.2 (Scheme 24, Eq. 10). Literature precedents20 and our control reactions further
demonstrated the elimination of sulfilimine could take place under heated conditions without any other additives
(e.g. palladium and potassium salts, trifluoroacetic acid, or water). It is also interesting to note that compared
with the sulfilimine ligand, bis-sulfoxide ligand 8.2 only afforded a trace amount of elimination products under
the reaction conditions. Further control experiments show that both of the elimination products (13.1 and 13.2)
are suitable ligands for the β-arylation reaction, giving comparable yields as using bis-sulfilimine ligand 9.
Scheme 24. Elimination of Sulfilimine Ligand
18

ACCEPTED MANUSCRIPT
Based on these above features, which are distinct from typically homogeneous palladium catalysis, we
hypothesized this new β-arylation reaction might be promoted by palladium nanoparticles or heterogeneous
palladium catalysts formed in situ. A large number of prior reports have demonstrated the facilitating effects of
sulfur-based ligands, acid, and salts for the formation and stabilization of palladium nanoparticle species.21
Especially, Stahl and coworkers have executed a detailed study and elucidated the role of palladium
nanoparticles in the aerobic dehydrogenation reaction of ketones.22 a In addition, a recent report has also
demonstrated that the reductive Heck reaction could also be catalyzed by palladium nanoparticles.22b Thus, a
series of experiments were devised to probe the involvement of palladium nanoparticles for the new β-arylation
conditions.
As the precipitation of palladium black was noted during and after the reaction, hot filtration test was first
employed to distinguish between a soluble nanoparticle and a heterogeneous catalyst.23 Parallel reactions were
set up under the reaction conditions and the conversion was monitored by gas chromatography (Scheme 25).
When the reactions initiated and reached around 20% yield, the reaction mixtures were passed through either a
short plug of Celite or a 200 nm PTFE filter to remove over-sized particles. Both reactions afforded dark red
filtrates. Subsequently, heat was restored to theses filtrates. Regarding the Celite layer, the heterogeneous filtrand
was added to a reaction vessel with newly mixed substrates, additives and solvents, and the reaction was then run
o
at 80 C for 12 h. We observed that the filtrates from both the Celite and PTFE filtration showed comparable
catalytic activity as the standard conditions. However, the filtrand from the Celite layer failed to catalyze the β-
arylation reactions with the new mixture of reactants. This hot filtration test suggested that the active palladium
catalyst generated during the induction period sustained the solubility, and those heterogeneous species were not
responsible for the transformation.
Scheme 25. Hot Filtration Tests
Dynamic lighting scattering (DLS) experiments24 were also employed to confirm the presence of small particles.
When the β-arylation product started to form after the induction period, a sample of the reaction solution was
submitted for the scan and indicated to have particles with an average size of 0.9 and 204 nm. In addition,
mercury poisoning test has also been carried out. A complete loss of reactivity was observed when excess
mercury was added to the reaction during the growth of yield (Scheme 26). Such an observation is consistent
with the involvement of palladium nanoparticles, as molecular mercury has been known to amalgamate metal
nanoparticles and thus inhibit the catalysis, 25 though the possibility of homogeneous Pd(0)-involved catalysis
cannot be completely excluded. 26 Overall, these above experiments, as well as precedents of palladium
nanoparticle-catalyzed dehydrogenation and reductive Heck reaction, are consistent with the involvement of
nanoparticle catalysts in the proposed pathway of the redox cascade.
Pd(OAc)2 (10 mol%)
Me
OTf
TsN
Ph
NTs
Ph
O
O
(10 mol%)
S
S
I
KTFA (200 mol%)
Me
Me
Ph
1,4-dioxane:TFA:H2O (20:2:1)
air, 80 C, 12 h
(100 mol%)
(250 mol%)
then Hg added at 180 min
19

ACCEPTED MANUSCRIPT
Scheme 26. Mercury Test
5 Conclusion
In summary, a palladium-catalyzed redox cascade strategy for direct β-arylation of ketones has been devised
through merging ketone dehydrogenation, aryl-halide bond activation, and conjugate addition. Two catalytic
systems have been developed. The Pd(TFA)2/P(i-Pr)3 system allows use of simple aryl halides as the aryl source.
The Pd(OAc)2/bis-sulfinimine system employs diaryliodonium salts as the aryl source, which avoids use of
stoichiometric silver salts and tolerates air and moisture. Both conditions show satisfactory substrate scopes and
functional group compatibility. These β-arylation methods represent an attractive alternative to the conjugate
addition for the synthesis of β-aryl ketones, as they employ readily available saturated ketones as the substrate
and avoid the use of strongly basic and/or nucleophilic metal reagents. Future work will be focused on
developing more efficient catalytic systems with a broader substrate scope and enabling enantioselective
transformations.
6 Experimental
6.1 General procedure for the β-arylation with aryl iodides
An 8 mL vial was charged with Pd(TFA)2 (13.3 mg, 0.1 equiv.), AgTFA (176 mg, 2.0 equiv.), Hexafluoro-2-
propanol (1 mL), ketone (1.0 mmol, 2.5 equiv.) and aryl iodide (0.4 mmol). The vial was sealed with a PTFE
lined cap and transferred to a glove box. The vial was opened and 1,4-dioxane (1 mL) and P(i-Pr)3 (16 ꢀL, 0.2
equiv.) were added under N2 purging. The vial was then sealed again, taken out of the glove box, and heated in a
pie-block at 80oC for 12 hours under stirring. Then, the vial was allowed to cool to room temperature and the
mixture was filtered through a small plug of silica gel, eluted with diethyl ether. The solvent of the filtrate was
then removed in vacuo and flash column chromatography (hexane/ethyl acetate or DCM/methanol) of the residue
gave the arylation product.
6.2 General procedure for the β-arylation with diaryliodonium salts
An 8 mL vial was charged with Pd(OAc)2 (9.0 mg, 0.1 equiv.), KTFA (122 mg, 2.0 equiv.), Mesitylaryliodonium
salt (0.4 mmol), bis-sulfilimine ligand 9 (24 mg, 0.1 equiv.), TFA (200 ꢀL), ketone (1.0 mmol, 2.5 equiv.), H2O
(100 ꢀL) and 1,4-dioxane (2 mL). The vial was sealed with a PTFE lined cap (no inert atmosphere is required)
and heated in a pie-block at 80 °C for 12 hours under stirring. Then, the vial was allowed to cool to room
temperature and the mixture was filtered through a small plug of silica gel, eluted with diethyl ether. The solvent
was then removed in vacuo and flash column chromatography (hexane/ethyl acetate or DCM/methanol) of the
residue gave the arylation product.
Acknowledgement
20

ACCEPTED MANUSCRIPT
We thank CPRIT and the Welch Foundation (F-1781) for funding. Dr. Michael Young and Dr. Vincent Lynch
are acknowledged for X-ray crystallography. Johnson Matthey is thanked for a donation of Pd salts.
References
1. Otera, J., Modern Carbonyl Chemistry. John Wiley & Sons: 2008.
2. For selected reviews of C−H functionalization of ketones and other carbonyl compounds, see: (a) Rouquet, G.;
Chatani, N. Angew. Chem., Int. Ed., 2013, 52, 11726. (b) Huang, Z.; Dong, G. Tetrahedron Lett. 2014, 55, 5869.
(c) Qiu, G.; Wu, J. Org. Chem. Front. 2015, 2, 169. (d) Zhang, B.; Guan, H.-X.; Liu, B.; Shi, B.-F. Chin. J. Org.
Chem. 2014, 34, 1487. (e) Huang, Z.; Lim, H. N.; Mo, F.; Young, M. C.; Dong, G. Chem. Soc. Rev. 2015, 44,
7764.
3. For representative examples, see: (a) Roth, G. N.; Chandra, A.; M. Nair, G. J. Nat. Prod. 1998, 61, 542. (b)
Melvin, L. S.; Johnson, M. R.; Harbert, C. A.; Milne, G. M.; Weissman, A. J. Med. Chem. 1984, 27, 67.
4. For selected reviews on transition-metal-catalyzed conjugate addition, see: (a) Rossiter, B. E.; Swingle, N. M.
Chem. Rev. 1992, 92, 771. (b) Gutnov, A. Eur. J. Org. Chem. 2008, 4547. (c) Hayashi, T.; Yamasaki, K. Chem.
Rev. 2003, 103, 2829.
5. For recent work of β-functionalization using saturated ketones, see: (a) Pirnot, M. T.; Rankic, D. A.; Martin, D.
B. C.; MacMillan, D. W. C. Science 2013, 339, 1593. (b) Petronijević, F. R.; Nappi, M.; MacMillan, D. W. C. J.
Am. Chem. Soc. 2013, 135, 18323. (c) Okada, M.; Fukuyama, T.; Yamada, K.; Ryu, I.; Ravelli, D.; Fagnoni, M.
Chem. Sci. 2014, 5, 2893. (d) Zhang, F.-L.; Hong, K.; Li, T.-J.; Park, H.; Yu, J.-Q. Science, 2016, 351, 252. (e)
Hong, K.; Park, H.; Yu, J.-Q. ACS Catal. 2017, 7, 6938.
6. Theissen, R. J. J. Org. Chem. 1971, 36, 752.
7. (a) Muzart, J.; Pete, J. P. J. Mol. Catal. 1982, 15, 373. (b) Wenzel, T. T. J. Chem. Soc., Chem. Commun. 1989,
932. (c) Park, Y. W.; Oh, H. H. Bull. Korean Chem. Soc. 1997, 18, 1123. (d) Shvo, Y.; Arisha, A. H. I. J. Org.
Chem. 1998, 63, 5640. (e) Diao, T.; Stahl, S. S. J. Am. Chem. Soc. 2011, 133, 14566. (f) Gao, W.; He, Z.; Qian,
Y.; Zhao, J.; Huang, Y. Chem. Sci. 2012, 3, 883. (g) Diao, T.; Pun, D.; Stahl, S. S. J. Am. Chem. Soc. 2013, 135,
8205.
8. For palladium-catalyzed dehydrogenation of carbonyl compounds other than ketones, see: (a) Chen, Y.;
Romaire, J. P.; Newhouse, T. R., J. Am. Chem. Soc. 2015, 137, 5875. (b) Chen, Y.; Turlik, A.; Newhouse, T. R.,
J. Am. Chem. Soc. 2016, 138, 1166. (c) Zhao, Y.; Chen, Y.; Newhouse, T. R. Angew. Chem. Int. Ed. 2017, 56,
13122.
9. Hartwig, J. F. In Organotransition Metal Chemistry: From Bonding to Catalysis; University Science Books:
Sausalito, 2010.
10. (a) Hennings, D. D.; Iwama, T.; Rawal, V. H. Org. Lett. 1999, 1, 1205. (b) Nelson, T. D.; Crouch, R. D. Org.
React. 2004, 265.
11. Izawa, Y.; Pun, D.; Stahl, S. S. Science 2011, 333, 209.
12. Gandeepan, P.; Rajamalli, P.; Cheng, C.-H. ACS Catalysis 2014, 4, 4485.
13. Ito, Y.; Hirao, T.; Saegusa, T. J. Org. Chem. 1978, 43, 1011.
14. For palladium-catalyzed dehydrogenation using 4,5-diazafluoren-9-one as the ligand, see: (a) Diao, T.;
Wadzinski, T. J.; Stahl, S. S. Chem. Sci. 2012, 3, 887. (b) Gao, W.; He, Z.; Qian, Y.; Zhao, J.; Huang, Y. Chem.
Sci. 2012, 3, 883.
15. For selected reviews, see: (a) Johansson, C. C. C.; Colacot, T. J. Angew. Chem., Int. Ed. 2010, 49, 676. (b)
Bellina, F.; Rossi, R. Chem. Rev. 2010, 110, 1082. (c) Culkin, D. A.; Hartwig, J. F. Acc. Chem. Res. 2003, 36,
234.
16. Trifluoroacetic acid is proposed to be generated during the course of the reaction.
17. For recent reviews of arylation reactions using diaryliodonium salts, see: (a) Deprez, N. R.; Sanford, M. S.
Inorg. Chem. 2007, 46, 1924. (b) Zhdankin, V. V.; Stang, P. J. Chem. Rev. 2008, 108, 5299. (c) Merritt, E. A.;
Olofsson, B. Angew. Chem., Int. Ed. 2009, 48, 9052.
18. For recent examples of the use of diaryliodonium salts in the C-H arylation reactions, see: (a) Phipps, R. J.;
McMurray, L.; Ritter, S.; Duong, H. A.; Gaunt, M. J. J. Am. Chem. Soc. 2012, 134, 10773. (b) Toh, Q. Y.;
McNally, A.; Vera, S.; Erdmann, N.; Gaunt, M. J. J. Am. Chem. Soc. 2013, 135, 3772. (c) Kieffer, M. E.; Chuang,
K. V.; Reisman, S. E. J. Am. Chem. Soc. 2013, 135, 5557. (d) Wagner, A. M.; Hickman, A. J.; Sanford, M. S. J.
Am. Chem. Soc. 2013, 135, 15710. (e) Zhang, F.; Das, S.; Walkinshaw, A. J.; Casitas, A.; Taylor, M.; Suero, M.
G.; Gaunt, M. J. J. Am. Chem. Soc. 2014, 136, 8851.
19. Chen, M. S.; Prabagaran, N.; Labenz, N. A.; White, M. C. J. Am. Chem. Soc. 2005, 127, 6970.
21

ACCEPTED MANUSCRIPT
20. For a selected review on elimination of alkyl bis-N-tosylsulfilimines, see: Oae, S.; Furukawa, N. Tetrahedron,
1977, 33, 2359.
21. For recent reviews on palladium nanoparticles, see: (a) Astruc, D.; Inorg. Chem. 2007, 46, 1884. (b) Crabtree,
R. H. Chem. Rev. 2011, 112, 1536.
22. (a) Pun, D.; Diao, T.; Stahl, S. S. J. Am. Chem. Soc. 2013, 135, 8213. (b) Parveen, N.; Saha, R.; Sekar, G.
Adv. Synth. Catal. 2017, 359, 3741.
23. For recent examples of filtration tests, see: (a) Biffis, A.; Zecca, M.; Basato, M. Eur. J. Inorg. Chem. 2001,
1131. (b) Conlon, D. A.; Pipik, B.; Ferdinand, S.; LeBlond, C. R.; Sowa, J. R., Jr.; Izzo, B.; Collins, P.; Ho, G.-J.;
Williams, M.; Shi, Y.-J.; Sun, Y. Adv. Synth. Catal. 2003, 345, 931. (c) Yu, K.; Sommer, W.; Weck, M.; Jones,
C. W. J. Catal. 2004, 226, 101. (d) Gaikwad, A. V.; Holuigue, A.; Thathagar, M. B.; ten Elshof, J. E.;
Rothenberg, G. Chem. Eur. J. 2007, 13, 6908.
24. For recent reviews and examples of DLS measurement of nanoparticle catalysis, see: (a) Bohren, C. F.;
Huffman, D. R. Absorption and Scattering of Light by Small Particles, 1st ed.; John Wiley & Sons: New York,
1998. (b) Pecora, R. J. Nanoparticle Res. 2000, 2, 123. (c) Berne, B. J.; Pecora, R. Dynamic Light Scattering
(With Applications to Chemistry, Biology, and Physics), 2nd ed.; Dover: Mineola, NY, 2000. (d) Hintermair, U.;
Hashmi, S. M.; Elimelech, M.; Crabtree, R. H. J. Am. Chem. Soc. 2012, 134, 9785.
25. For examples of mercury poisoning test, see: (a) Bayram, E.; Linehan, J. C.; Fulton, J. L.; Roberts, J. A. S.;
Szymczak, N. K.; Smurthwaite, T. D.; Ozkar, S.; Balasubramanian, M.; Finke, R. G. J. Am. Chem. Soc. 2011,
133, 18889. (b) Yu, K.; Sommer, W.; Richardson, J. M.; Weck, M.; Jones, C. W. Adv. Synth. Catal. 2005, 347,
161. (c) Finney, E. E.; Finke, R. G. Inorg. Chim. Acta 2006, 359, 2879. (d) Weck, M; Jones, C. W. Inorg. Chem.
2007, 46, 1865. (e) Jones, C. W. Top. Catal. 2010, 53, 942.
26. Mercury is also known to inhibit catalysis involving Pd(0) species, see: Phan, N. T. S.; Van Der Sluys, M.;
Jones, C. W. Adv. Synth. Catal. 2006, 348, 609.
22