Synergistic cascade catalysis by metal nanoparticles and Lewis acids in hydrogen autotransfer

Article abstract of DOI:10.1039/c4sc03627a

Of the many types of catalysis involving two or more catalysts, synergistic catalysis is of great interest because novel reactions or reaction pathways may be discovered when there is synergy between the catalysts. Herein, we describe a synergistic cascade catalysis, in which immobilized Au/Pd bimetallic nanoparticles and Lewis acids work in tandem to achieve the N-alkylation of primary amides to secondary amides with alcohols via hydrogen autotransfer. When Au/Pd nanoparticles were used with metal triflates, a significant rate acceleration was observed, and the desired secondary amides were obtained in excellent yields. The metal triflate is thought to not only facilitate the addition of primary amides to aldehydes generated in situ, but also enhance the returning of hydrogen from nanoparticles to hydrogen-accepting intermediates. This resulted in a more rapid turnover of the nanoparticle catalyst, and ultimately translated into an increase in the overall rate of the reaction. The two catalysts in this co-catalytic system work in a synergistic and cascade fashion, resulting in an efficient hydrogen autotransfer process.

Full text of DOI:10.1039/c4sc03627a

Chemical
Science
View Article Online
View Journal | View Issue
EDGE ARTICLE
Synergistic cascade catalysis by metal
nanoparticles and Lewis acids in hydrogen
Cite this: Chem. Sci., 2015, 6, 1719
autotransfer†
*
Gerald C. Y. Choo, Hiroyuki Miyamura and Shu Kobayashi
¯
Of the many types of catalysis involving two or more catalysts, synergistic catalysis is of great interest
because novel reactions or reaction pathways may be discovered when there is synergy between the
catalysts. Herein, we describe a synergistic cascade catalysis, in which immobilized Au/Pd bimetallic
nanoparticles and Lewis acids work in tandem to achieve the N-alkylation of primary amides to
secondary amides with alcohols via hydrogen autotransfer. When Au/Pd nanoparticles were used with
metal triflates, a significant rate acceleration was observed, and the desired secondary amides were
obtained in excellent yields. The metal triflate is thought to not only facilitate the addition of primary
amides to aldehydes generated in situ, but also enhance the returning of hydrogen from nanoparticles to
hydrogen-accepting intermediates. This resulted in a more rapid turnover of the nanoparticle catalyst,
and ultimately translated into an increase in the overall rate of the reaction. The two catalysts in this
co-catalytic system work in a synergistic and cascade fashion, resulting in an efficient hydrogen
autotransfer process.
Received 24th November 2014
Accepted 17th December 2014
DOI: 10.1039/c4sc03627a
www.rsc.org/chemicalscience
in its high atom economy because no external oxidant is
required for the activation of substrates, and no external
Introduction
Catalysis with metal nanoparticles1,2 is a hot research eld that
has gained much attention. Metal nanoparticles have great
potential as catalysts because of their facile heterogenization,
robustness, and unique reactivity and selectivity that result
from their characteristic electronic state. From the early reports
of Au nanoparticle-catalyzed aerobic oxidation reactions3–14 to
recent reports of bond-forming reactions, chiral15–19 or other-
wise, metal nanoparticles have been widely investigated as
extremely active catalysts, and applied to many reactions.20–26
The concept of employing two distinct catalysts in one reaction
system is a powerful strategy in organic synthesis to accelerate
reactions efficiently in a synergistic manner,27–29 but the use of
metal nanoparticles in such systems is less well-developed
possibly due to catalyst incompatibility; a second catalyst
sometimes deactivates metal nanoparticles. If compatibility
issues are ironed out,30 metal nanoparticles show great poten-
tial and possibility for use in synergistic catalysis.27
reductant is required for the reduction of intermediates
generated in situ. The hydrogen autotransfer methodology has
been developed using homogeneous Ir, Rh, Pd and Ru metal
complexes as catalysts.31–38 However, the recovery and reuse of
the precious metals in these reactions are usually difficult. The
use of metal with an organocatalyst39 for hydrogen autotransfers
has been reported recently. Metal nanoparticles have also been
demonstrated to be effective catalysts for the hydrogen auto-
transfer process.40,41 The alkylation of amines using alcohols via
hydrogen autotransfer has been widely reported but reports of
the N-alkylation of primary amides via hydrogen autotransfer
are quite limited compared to those of the alkylation of amines
despite the potential synthetic utility of the reaction.38,42–45 This
could be because amides are generally unreactive when
compared to amines so the nucleophilic attack of a primary
amide to an aldehyde generated in situ during the hydrogen
autotransfer process is difficult.
Our group has been investigating polymer-incarcerated (PI)
metal nanoparticles as catalysts for a variety of reactions such as
coupling reactions, aerobic oxidation of alcohols to aldehydes/
ketones, hydrogenation/reduction reactions and tandem
oxidation processes.46–49 In many cases, the reaction conditions
are mild because the immobilized metal nanoparticles are very
active and facilitate the above-mentioned reactions effectively.
More recently, we have been interested in employing immobi-
lized metal nanoparticles and other functional molecules in
reactions systems, the synergy of which has paved the way for
Hydrogen autotransfer, also known as “borrowing
hydrogen,” is a useful methodology for the formation of C–C
and C–N bonds. The attractiveness of hydrogen autotransfer lies
Department of Chemistry, School of Science, The University of Tokyo, Hongo,
Bunkyo-ku, Tokyo 113-0033, Japan. E-mail: shu_kobayashi@chem.s.u-tokyo.ac.jp
† Electronic supplementary information (ESI) available: General procedures,
materials, and instrumentation; synthesis, characterization and relevant
spectra/charts; procedures and results for optimization and additional
experiments. See DOI: 10.1039/c4sc03627a
This journal is © The Royal Society of Chemistry 2015
Chem. Sci., 2015, 6, 1719–1727 | 1719

View Article Online
Chemical Science
Edge Article
Table 1 Effect of oxygen and additive on the reaction
Entry
Ma
Additive
Yieldb (%)
1
2
3
Ir, Ru, Rh, Ni or Co
Au
Pd
Au
—
—
—
—
—
—
—
n.dc
n.dc
Tracec
0
7
24
43
89
0
4d
5d
6d
7d
8d
9d
10d
Pd
Scheme 1 Proposed dual catalysis for the challenging N-alkylation of
primary amides via hydrogen autotransfer.
Au (2 mol%) + Pd (2 mol%)
Au/Pd (Au : Pd ¼ 1 : 1)
Au/Pd (Au : Pd ¼ 1 : 1)
MgSO4 (1.66 eq.)
MgSO4 (1.66 eq.)
—
—
—
0
many interesting reactions50 and tandem oxidation
processes.30,51 We were, therefore, interested in the synergistic
catalysis between the PI metal nanoparticle catalyst and a
second catalyst for the challenging hydrogen autotransfer
reaction between primary amides and alcohols. We expected the
PI metal nanoparticle catalyst to be an effective catalyst for
hydrogen autotransfer because we are able to immobilize
various metal nanoparticles, including multi-metallic nano-
particles,17,30,50–55 and therefore, we are able to tune catalytic
activity easily by choosing appropriate metal sources.53–55 The
second catalyst is expected to enhance the efficiency of the
overall reaction by facilitating the nucleophilic addition of the
primary amide to the carbonyl compound generated in situ,
which is a key but slow step due to the poor nucleophilicity of
primary amides (Scheme 1).
a
Catalyst loading was set to 2 mol%. In the case of bimetallic catalysts,
the catalyst loading was set to 2 mol% with respect to the rst metal
stated. Determined by GC analysis with dodecane as the internal
standard. Determined by GCMS analysis of crude aer the stipulated
b
c
d
reaction time (n.d. ¼ not detected). Deoxidized benzyl alcohol was
used.
amount obtained earlier. PI/CB-Au, however, still did not afford
any product (entry 4), interestingly, a physical mixture of PI/CB-
Au and PI/CB-Pd catalysts afforded more of the desired product,
although the yield, at 24%, was still unsatisfactory (entry 6).
When the PI/CB-Au/Pd bimetallic nanoparticle catalyst was
employed, we observed a dramatic increase in yield to 43%
(entry 7). The metal nanoparticles in the catalyst was conrmed
to be alloyed by scanning transmission electron microscopy
(STEM) and energy-dispersive X-ray spectroscopy (EDS) anal-
yses; the catalyst is not a mixture of Au and Pd nanoparticles
that are independent of each other. Alloyed bimetallic nano-
particles oen demonstrate catalytic properties that are unique
from their monometallic counterparts;69–73 Au is reported to
have a promotional effect on Pd such that when the two are
combined, it results in a more active catalyst.74,75 We believe that
the promotional effect of Au is more pronounced when the
metal nanoparticles are bimetallic alloy nanoparticles, due to
the proximity of Au to Pd. In addition, the better catalytic
activity may also be attributed to the polarization of electric
charge on the surface of the alloyed bimetallic nanoparticle
arising from the difference in electronegativity between Au and
Pd.69,71
In spite of all our subsequent attempts, it was difficult to
improve the yield beyond 43%. We then decided to examine the
postulated mechanism of the reaction, and focused our atten-
tion on water that was formed as a byproduct (bottom of
Scheme 1). We hypothesized that removing water from the
reaction system would favor the formation of the acylimine
intermediate and improve the yield. Pleasingly, when MgSO4
(50 mg, 1.66 eq.) was employed as an additive, the yield
improved signicantly to 89% (Table 1, entry 8). A control
Results and discussion
Discovery of a suitable immobilized nanoparticle catalyst for
hydrogen autotransfer between benzamide and benzyl alcohol
With benzamide (1a) and benzyl alcohol (2a) as substrates for
the model reaction, initial screening of various polymer-incar-
cerated metal nanoparticle catalysts with carbon black as a
secondary support (PI/CB-M catalysts) was conducted. Initially,
we followed an earlier report and adopted benzyl alcohol as the
solvent.56 Under the reaction conditions shown in Table 1, the
reaction did not proceed when typical metals for hydrogen
autotransfer, such as Ir, Ru and Rh, were used. No product was
observed for PI/CB-Ni or PI/CB-Co either (entry 1). We then
turned our attention to Au3–14 and Pd24,57–68 nanoparticle cata-
lysts because these catalysts have been widely investigated and
demonstrated to be effective catalysts for aerobic oxidation,
dehydrogenative oxidation,68 hydrogenation and bond forming
reactions. While PI/CB-Au did not afford any product (entry 2), a
trace amount of product was detected with PI/CB-Pd (entry 3).
Aer many attempts at improving the yield, the desired N-
benzylbenzamide product (3aa) was obtained in 7% yield with
PI/CB-Pd, using benzyl alcohol that was carefully degassed
(entry 5), which was a marked improvement over the trace
1720 | Chem. Sci., 2015, 6, 1719–1727
This journal is © The Royal Society of Chemistry 2015

View Article Online
Edge Article
Chemical Science
experiment in which only MgSO4 was employed without the afford the desired product almost quantitatively with Mg(OTf)2
catalyst conrmed that MgSO4 was not the main catalyst (entry 5). Encouraged by this result, we reduced the amount of
because no product 3aa was observed (entry 9).
Mg(OTf)2 in subsequent experiments to determine if Mg was
acting as a Lewis acid for the reaction. Satisfyingly, Mg(OTf)2
worked well as an additive even at 0.5 equivalents and 5 mol%
(entries 6 and 7). That Mg(OTf)2 could be employed catalytically
as a co-catalyst while MgSO4 could not was probably due to a
difference in Lewis acidity arising from the difference in coun-
teranions.76 Other Group 2 metal triates were also screened
and they were found to be effective co-catalysts for the reaction
as well (entries 8 and 9). We then examined some metal triates
from the neighboring groups and most of them worked well
(entries 10–14); >95% of the desired product was obtained, with
the exception of KOTf.
Lewis acid effect outweighs desiccant effect
Based on these initial results, we proceeded to reduce the
amount of benzyl alcohol required from solvent amount to 4
equivalents (see ESI 4-2†). We found that toluene, in place of
benzyl alcohol (2a) as the solvent, was effective for the reaction.
Aer optimization of the reaction conditions using toluene as
the solvent, we obtained the desired amide quantitatively with 4
equivalents of benzyl alcohol (Table 2, entry 1). It should be
noted that the amide alkylation reaction proceeded under
neutral conditions. Decreasing the amount of MgSO4 from
excess to catalytic resulted in a decrease in yield (see ESI† 4-2†).
Despite this result, we were, at this juncture, unable to rule out
MgSO4 working as a Lewis acid for the addition of benzamide to
benzaldehyde. Several other additives were screened to deter-
mine their effect on the reaction and also to determine if there
was a more effective additive that could be used in catalytic
amounts (Table 2, entries 2–17).
When water was deliberately introduced into the reaction
vessel, the reaction still proceeded to give the desired product in
good yield (entry 9 with footnote d, see ESI 4-5†). Furthermore,
when molecular sieves were used as the additive (entries 16 and
˚
17), only molecular sieves 5 A gave good yield (entry 17), sug-
gesting that rather than the dehydrating properties of the
additive, it was the acidity of the additive that was crucial for the
reaction. Thus, the results from these control experiments ruled
out MgSO4 working as a desiccant.
When triic acid was examined as the co-catalyst, the yield
was 70% (entry 15). This yield, which was higher than when no
co-catalyst was employed (entry 18), demonstrated that acidity
was important for the reaction but it also suggested that Lewis
acidity is more crucial than Brønsted acidity because the yield
was still lower than when a Lewis acid such as Mg(OTf)2 was
employed.
Initially, Mg salts were examined. Neither Mg halides nor
MgO nor Mg(OH)2 worked well as additives for the reaction
(entries 2–4). However, the reaction proceeded smoothly to
Table 2 Screening of additives and equivalents of benzyl alcohol
We then further optimized the reaction by employing the
effective Lewis acid co-catalysts to the model reaction with
3 equivalents of benzyl alcohol (X ¼ 3 column in Table 2). Group
2 metal triates worked extremely well for the reaction, afford-
ing the desired product quantitatively (Table 2, entries 7–9). On
the other hand, LiOTf (entry 10) and Sc(OTf)3 (entry 13) did not
perform as well. Yb(OTf)3 also gave the desired product quan-
titatively (entry 14), albeit with the formation of several side
Entry Additive
Amount X ¼ 4a X ¼ 3a
1
MgSO4
1.66 eq. Quant. 91
2
3
MgZ2 (Z ¼ Fl, Cl, Br) 1.66 eq. 1–17
—
—
—
—
—
MgO
1.66 eq. 31
1.66 eq. 31
1.66 eq. 97
0.5 eq. Quant.
ꢀ
products. Under a lower temperature of 120 C (heating plate),
4
5
6
7
8
9
10
11
12
13
14
15
16e
17e
18
Mg(OH)2
Mg(OTf)2
Mg(OTf)2
Mg(OTf)2
Ca(OTf)2
Ba(OTf)2
LiOTf
NaOTf
KOTf
Sc(OTf)3
Yb(OTf)3
Ba(OTf)2 outperformed the other Group 2 metal triates
screened (entries 7–9). In addition, because excellent yield
(95%) was also achieved with 2.5 equivalents of benzyl alcohol
with Ba(OTf)2 (entry 9 with footnote b), we decided to adopt it as
the co-catalyst for our reaction system.
5 mol% Quant. Quant. (64)c
5 mol% 99
5 mol% 98
Quant. (63)c
Quant. (95)b (85)c (94)d
5 mol% Quant. 90
5 mol% 98
5 mol% 73
5 mol% 98
—
—
92
Substrate scope: Excellent yields achieved for difficult
aliphatic amide substrates
5 mol% Quant. Quant.
TfOH
MS 3 A or MS 4 A
MS 5 A
5 mol% 70
—
—
—
—
With the optimized reaction conditions in hand, we proceeded
to examine various substrates for the reaction (Table 3). In
general, when benzyl alcohol (2a) was used, benzamide (1a) and
its analogs worked well to afford the products in excellent
yields, especially those with electron-donating substituents on
˚
˚
20 mg
20 mg
—
<10
87
50
˚
—
a
Yield was determined by GC analysis with dodecane as the internal
b
c
standard. 2.5 eq. of benzyl alcohol were used. GC yield obtained
when the reaction was conducted at 120 ꢀC (hot plate temperature). the aromatic ring (entries 1–4). This can be attributed to the
d
4-Methylbenzyl alcohol used as the substrate;
a 5 : 1 ratio of
increased nucleophilicity of the amide. Even with a p-uoro-
substitution, the reaction proceeded well to afford the desired
product in high yield (entry 6). Heteroaromatic benzamide
e
toluene : H2O was used as the solvent. Catalyst loading: 2 mol% Au;
solvent: xylene (C ¼ 0.25 M).
This journal is © The Royal Society of Chemistry 2015
Chem. Sci., 2015, 6, 1719–1727 | 1721

View Article Online
Chemical Science
Edge Article
Table 3 Substrate scope
analog was employed, the yields were moderate (entry 18),
possibly due to the reduced electrophilicity of the carbon on the
carbonyl moiety of the corresponding aldehyde generated in situ.
Conversely, when an electron-poor benzyl alcohol analog was
employed, the yields were slightly higher (entries 16 & 19).
Attempts at using aliphatic alcohols have proven futile and
examination of the reaction mixture indicated to us that the
problem was the conversion of the alcohol (vide infra).
Entry
R
3
Yielda (%)
1
2
3
4
5
6
H
3aa
3ba
3ca
3da
3ea
3fa
95
98
Quant.
89
53
Reuse of heterogeneous immobilized gold–palladium
nanoparticle catalyst
p-Me
p-MeO
o-EtO
o-OH
p-F
We then proceeded to examine the reusability of our PI/CB-Au/
Pd catalyst in our reaction system (Table 4). With the addition of
Ba(OTf)2 for each run (Table 4, upper row), the heterogeneous
catalyst could be reused in run 2 with no pre-treatment required
but in run 3, a signicant decrease in yield was observed. We
believed that the Au/Pd bimetallic nanoparticles might have
been deactivated. Taking cues from our previous work,10,17,53–55
the recovered catalyst from run 3 was reactivated by heating it at
90
7b
8c
2-Pyridyl 3ga
3-Pyridyl 3ha
Me
n-C5H11
i-Pr
t-Bu
Bn
63
52
77
95
94
90
9b
10
11
12
13
14
3ia
3ja
3ka
3la
3ma 91
3na
ꢀ
170 C for 5 hours under open air before it was used in run 4.
Remarkably, catalytic activity recovered and excellent yields
were obtained in subsequent runs; the recovered catalyst had to
be treated by the method mentioned above only when the
conversion of benzyl alcohol showed signs of slowing down
(aer runs 6 and 9). High yields of >95% were achieved for all
runs thereaer up to run 11 and we conrmed no leaching of
both Au and Pd in each and every run, demonstrating the
robustness of the heterogeneous PI/CB-Au/Pd nanoparticle
catalyst for the hydrogen autotransfer process.
c-Hex
Quant.
15
Me
CO2Me
3ab
3ad
Quant.
61
16d
17
18
19
Me
MeO
CO2Me
3jb
3jc
3jd
86
44
68
Reusability of the heterogeneous catalyst is not restricted to
the reaction conditions where Ba(OTf)2 was used in combina-
tion with benzyl alcohol (2a). Gratifyingly, when Ca(OTf)2 was
used as the co-catalyst and 4-methylbenzyl alcohol (2b) was
used as the substrate, the heterogeneous catalyst could be
reused up to 5 times (Table 4, lower row). We also discovered
that if the reaction work-up and the recovery of the catalyst was
performed using dichloromethane, no additional Ca(OTf)2 was
a
b
c
Isolated yield. 5 eq. of 2a were used. 2 mol% Au and 10 mol%
d
Ba(OTf)2 were used. With some impurity (alcohol starting material);
refer to ESI.
analogs were also applicable to this reaction system, albeit with
modied reaction conditions to improve the yields (entries 7
and 8). We suspect that the heteroatom on the aromatic ring
could have coordinated to the Au/Pd bimetallic nanoparticles or
the Lewis acid, resulting in a slight deactivation of the desired
catalysis. We then turned our attention to aliphatic substrates
(compound 1, R1 ¼ alkyl), which are difficult substrates in
hydrogen autotransfer. Satisfyingly, all aliphatic substrates
afforded the desired products in more than 90% yield (entries
10–14), with the exception of acetamide, for which more benzyl
alcohol (2a) was required to obtain a good yield of 77% (entry 9).
This is a marked improvement over earlier reports that used
aliphatic substrates (1i–1n) with benzyl alcohol in hydrogen
autotransfer38,44 and this highlights one of the advantages of our
synergistic catalytic system.
Table 4 Reusing of the heterogeneous catalyst – PI/CB-Au/Pd
2a; M ¼ Ba
2b; M ¼ Ca
Run
1b
>99
1
2b
99
2d
95
3b
53
3d
93
4–11b,c
Yielda (%)
Run
95 to >99
4c,d
89
5d
93
Yielda (%)
93
a
Determined by GC analysis with dodecane as the internal standard.
b
Other benzyl alcohol analogs were then employed in the
reaction (entries 15–19). For p-methyl-substituted benzyl alcohol,
high yields were obtained for both benzamide and hexanamide
(entries 15 & 17). When an even more electron-rich benzyl alcohol
No leaching of Au or Pd was detected (under detection limit;
c
determined by ICP analysis). The recovered catalyst from the
previous run was reactivated before use in runs 4,
Recovered catalyst was treated with DCM and no additional Lewis
acid was added for the new run.
7 and 10.
d
1722 | Chem. Sci., 2015, 6, 1719–1727
This journal is © The Royal Society of Chemistry 2015

View Article Online
Edge Article
Chemical Science
required for each run, indicating that the Lewis acid was also
recovered in the process (footnote d in Table 4 and ESI 3-3†).
Capturing reaction intermediates and demonstrating N-
alkylation of amides involving aliphatic substrates
In our quest to gain some insights into the reaction mechanism,
we conducted an experiment starting from tolualdehyde (4) and
benzamide (1a) under hydrogen atmosphere. We obtained
neither alcohol 2b nor desired product 3ab, which suggested
that even if hydrogen gas was generated during the reaction
(from the hydrogen accepted by the nanoparticle catalyst),
hydrogen gas cannot serve as a reductant, and that any reduc-
tion that occurs in our reaction system was due to transfer
hydrogenation (see ESI 4-6 and 4-14†). Instead, we isolated a
solid, in large amounts, that was highly insoluble in many
solvents, and we identied it to be N,N0-(p-methylphenyl-
methylene)dibenzamide (N,N0-diamide 5a), formed from one
molecule of tolualdehyde (4) and two molecules of benzamide
(1a) (Scheme 2a).77 Interestingly, this unexpected compound
was also formed under argon atmosphere, regardless of
whether Lewis acid was added or not (see ESI 4-7, 4-8 and 4-9†).
When 5a was subjected to the optimized conditions with 4-
methylbenzyl alcohol (2b), the desired product (3ab) was
obtained (Scheme 2b). This strongly implied that N,N0-diamide
5a could be a key intermediate in the reaction.
In addition, when we synthesized N,N0-diamide (5b) using 3-
phenylpropionaldehyde (aliphatic aldehyde) and benzamide
(1a), and subjected the newly formed N,N0-diamide (5b) to the
optimized conditions with 4-methylbenzyl alcohol (2b), we
obtained two secondary amides – N-(4-methylbenzyl)benzamide
(3ab) and N-(3-phenylpropyl)benzamide (3af) (Scheme 3a).
Interestingly, when we started out with benzamide (1a) and an
aliphatic aldehyde, and used a secondary alcohol as the
reductant, benzamide (1a) was N-alkylated smoothly and the
desired secondary amide (3af) was isolated in 81% yield under
our dual catalysis conditions (Scheme 3b). This result demon-
strated that our catalytic system is also effective for the
N-alkylation of amides via transfer hydrogenation, when both
benzylic and aliphatic aldehydes are used. Furthermore, this
reinforces the notion that aliphatic alcohols do not work for our
reaction system not because the addition of an amide to an
aldehyde is problematic, but because there is difficulty in the
Scheme 3 Amide alkylation using aliphatic substrates.
initial conversion of the aliphatic alcohol to the corresponding
aldehyde.
Evidence of synergy between metal nanoparticle and Lewis
acid
To clarify the synergistic effect between metal nanoparticles and
a Lewis acid, we monitored the hydrogen autotransfer reactions
under various reaction conditions. The reaction prole when 4
equivalents of 4-methylbenzyl alcohol (2b) were employed was
examined. We monitored the formation/consumption of 2b,
tolualdehyde (4), the N,N0-diamide (5a) and the desired product
(3ab). The resulting reaction proles obtained without Ba(OTf)2
(Fig. 1) and with Ba(OTf)2 (Fig. 2) were then compared. During
our monitoring, we also observed the formation of two side
products – xylene (6) and di(4-methylbenzyl) ether (7). From
control experiments (see ESI 4-12, 4-13 and 4-15†), it was clear
that tolualdehyde (4) and these two side products were formed
from 4-methylbenzyl alcohol (2b), only when PI/CB-Au/Pd was
Fig. 1 Reaction profile with 4-methylbenzyl alcohol (2b) as substrate
and no Lewis acid as co-catalyst.
Scheme 2 N,N0-Diamide as a key intermediate of the reaction.
This journal is © The Royal Society of Chemistry 2015
Chem. Sci., 2015, 6, 1719–1727 | 1723

View Article Online
Chemical Science
Edge Article
reaction was almost complete aer 6 h and the formation of the
desired product began much earlier. In addition, we observed a
lower concentration of the aldehyde at the steady-state for
Ba(OTf)2, which implied that the initiation of the reaction was
faster and that the induction period was shorter. The results
demonstrate that the choice of Lewis acid is important.
Proposed reaction mechanism
Based on the observations made from the reaction proles and
various control experiments, we propose the reaction mecha-
nism shown in Scheme 4. The reaction begins with the accu-
mulation of the aldehyde (“initiation” process within green box,
Fig. 2 Reaction profile with 4-methylbenzyl alcohol (2b) as substrate
and with Ba(OTf)2 as the co-catalyst.
present, due to a possible disproportionation-like reaction
(steps I and II0 in Scheme 4, ESI 4-4†).78 No reaction was
observed when an attempt was made to reduce the ether (7)
using either 4-methylbenzyl alcohol (2b) as the hydrogen
source, or using hydrogen gas with the PI/CB-Au/Pd catalyst (see
ESI 4-16 and 4-17†).
When Ba(OTf)2 was absent, a large amount of tolualdehyde
(4) (approx. 120%, based on the amount of benzamide) was
formed aer 3 h while the formation of the desired product
(3ab) was slow, reaching only a mere 20% aer 6 h. In addition,
xylene (6) was steadily formed, reaching 40% aer 6 h, and
almost no N,N0-diamide (5a) was observed (Fig. 1).
Fig. 3 Reaction profile with benzyl alcohol (2a) as the substrate and
In contrast, when Ba(OTf)2 was present (Fig. 2), the rate of
formation of the desired product (3ab), xylene (6) and the ether
(7) was accelerated. In particular, for the same duration of 6 h,
the amount of the desired product (3ab) and ether (7) formed
was almost 4–5 times as much. As a result of this Lewis acid
acceleration, the consumption of 4-methylbenzyl alcohol (2b)
was also much faster. Aer 30 min into the reaction, we
observed the formation of the N,N0-diamide (5a), the concen-
tration of which remained steady until 3 h into the reaction, and
then returned to almost zero thereaer. Concurrently, there was
a swi increase in the yield of the desired product (3ab) during
the same period. The amount of tolualdehyde (4) increased
during this period and reached a steady-state concentration of
50%. This implied that an induction period existed, and that a
certain amount of the aldehyde had to be rst accumulated
before the desired product (3ab) started to form. A similar
phenomenon was observed even when a Lewis acid was not
present (Fig. 1).
Ba(OTf)2 as the co-catalyst.
We also made a comparison between reaction proles
obtained with 5 mol% of Ba(OTf)2 (Fig. 3) and 1.66 equivalents
of MgSO4 (Fig. 4), with the focus on the alcohol, the aldehyde
and the desired product. We observed a dramatic rate acceler-
ation with Ba(OTf)2 than with MgSO4 (see ESI 4-23†) because the
Fig. 4 Reaction profile with benzyl alcohol (2a) as the substrate and
MgSO4 as the co-catalyst.
1724 | Chem. Sci., 2015, 6, 1719–1727
This journal is © The Royal Society of Chemistry 2015

View Article Online
Edge Article
Chemical Science
Scheme 4 Schematic representation of the processes within the reaction system, which includes an initiation process and Lewis acid
acceleration.
step I) through a disproportionation-like reaction, which results a Lewis acid when the aldehyde and benzamide was heated
in the formation of toluene or xylene, and the sacricial under reux with toluene (see ESI 4-9†). Therefore, that we did
consumption of an alcohol (step II0). The aldehyde then reacts not observe any N,N0-diamide when no Lewis acid was present
with the primary amide or the alcohol to form hydrogen- in our reaction system must imply that there exists a very fast
acceptors – N,N0-diamide (5a) or acetal (8), which equilibrate step aer the formation of the N,N0-diamide that leads to the
with the respective acylimine (9)/hemiaminal (10)/N,O-acetal desired product. When a Lewis acid was present, however, the
(100) or oxocarbenium ion (11). Concurrently, hydrogen is being N,N0-diamide is at steady-state, indicating that the formation of
abstracted from the alcohol by the nanoparticle catalyst (step I), the N,N0-diamide is no longer the rate-determining step (Fig. 2).
and that hydrogen is then “returned” to the hydrogen-acceptors The overall rate of the sequential reaction is then governed by
generated in the system to afford the desired product or the the turnover rate of the nanoparticle catalyst, in particular, by
ether (step II). The N,N0-diamide (5a) is the most thermody- the rate of hydrogen “returning”.80
namically stable compound among the potential hydrogen
The catalytic cycle of the Au/Pd nanoparticle involves two
accepting intermediates (5a, 9 and 10 in Scheme 4) because only steps – hydrogen abstraction (step I) and returning (steps II0 &
the N,N0-diamide was observed in the control experiments II), which are interdependent processes. While there is
starting from aldehydes and primary amides (see ESI 4-6† and competition between the desired reaction pathway and the side
Scheme 2a).
reaction pathways with regard to accepting hydrogen from the
The Au/Pd nanoparticle catalyst plays the crucial role of Au/Pd-H2 catalyst, the presence of a Lewis acid would lead to the
transferring hydrogen from the alcohol to the various hydrogen- production of various highly reactive hydrogen acceptors (5a, 9,
acceptors. The Lewis acid, on the other hand, must be involved and 10),81 which would result in a faster turnover (step II) for the
in both the formation of the N,N0-diamide79 and the hydrogen catalyst from Au/Pd-H2 (resting state) to Au/Pd nanoparticle. In
“returning” process aer the said formation. We postulate this turn, that would lead to the production of more aldehyde and
based on our observations of the different steady state thus more hydrogen-acceptor intermediates. As a result of the
concentrations of the N,N0-diamide and the different rates at acceleration of various steps within the reaction system, and
which the desired product was formed, for experiments with also the faster catalytic turnover of the Au/Pd nanoparticle
and without the Lewis acid (vide infra).
catalyst, the consumption of 4-methylbenzyl alcohol is signi-
Without a Lewis acid in the system, we expect the formation cantly quickened. However, in the absence of an amide, even
of the N,N0-diamide to be the rate-determining step because the with a Lewis acid, consumption of the alcohol was not full even
N,N0-diamide is very quickly consumed aer it is produced, aer 18 h (see ESI 4-13 vs. 4-24†). This implies that hydrogen
resulting in the close to zero concentration observed (Fig. 1). We abstraction itself is not accelerated by the Lewis acid. Further-
inferred this from the fact that the N,N0-diamide was formed in more, the concentration of the aldehyde at steady-state is lower
high concentration under thermodynamic control even without with
a
more efficient Lewis acid because that lower
This journal is © The Royal Society of Chemistry 2015
Chem. Sci., 2015, 6, 1719–1727 | 1725

View Article Online
Chemical Science
Edge Article
concentration is presumably sufficient for the hydrogen 11 B. K. Min and C. M. Friend, Chem. Rev., 2007, 107, 2709–
acceptors to form (Fig. 1 vs. Fig. 2 and Fig. 3 vs. Fig. 4).
2724.
12 C. Della Pina, E. Falletta, L. Prati and M. Rossi, Chem. Soc.
Rev., 2008, 37, 2077–2095.
13 C. H. Christensen and J. K. Norskov, Science, 2010, 327, 278–
279.
Conclusion
We have discovered a synergistic cascade catalytic system that
employs immobilized Au/Pd nanoparticles and Ca(OTf)2/
Ba(OTf)2 Lewis acid for the N-alkylation of primary amides with
benzyl alcohol and its analogs via hydrogen autotransfer. The
choice of metal(s) for the nanoparticle catalyst and choice of
Lewis acid is the key to achieve an efficient system. In particular,
the performance of the Lewis acid is crucial for overall efficient
catalytic turnover. This is a very unique catalytic system in
which metal nanoparticles and Lewis acid work synergistically
within a complex and elaborated catalytic cycle. This is also the
rst example of a metal nanoparticle-catalyzed hydrogen auto-
transfer process that employs primary amides as substrate. The
substrate scope was broad and in particular, excellent yields
were observed for many difficult aliphatic primary amide
substrates. Both metal nanoparticle and Lewis acid were reus-
able and no leaching of Au and Pd to the product was observed.
We strongly believe that such a synergistic system paves the way
for us to achieve reactions that with only heterogeneous cata-
lysts, including metal nanoparticles, are currently either
impossible or inefficient.
14 T. Tsukuda, H. Tsunoyama and H. Sakurai, Chem.–Asian J.,
2011, 6, 736–748.
¨
15 K. V. S. Ranganath, J. Kloesges, A. H. Schafer and F. Glorius,
Angew. Chem., Int. Ed., 2010, 49, 7786–7789.
16 K. V. S. Ranganath and F. Glorius, Catal. Sci. Technol., 2011,
1, 13–22.
17 T. Yasukawa, H. Miyamura and S. Kobayashi, J. Am. Chem.
Soc., 2012, 134, 16963–16966.
18 E. Gross, J. H. Liu, S. Alayoglu, M. A. Marcus, S. C. Fakra,
F. D. Toste and G. A. Somorjai, J. Am. Chem. Soc., 2013,
135, 3881–3886.
19 T. Yasukawa, H. Miyamura and S. Kobayashi, Chem. Soc.
Rev., 2014, 43, 1450–1461.
´
20 A. Grirrane, A. Corma and H. Garcıa, Science, 2008, 322,
1661–1664.
21 H. Cong and J. A. Porco Jr, ACS Catal., 2011, 2, 65–70.
´
22 A. Corma, A. Leyva-Perez and M. J. Sabater, Chem. Rev., 2011,
111, 1657–1712.
23 Y. Zhang, X. Cui, F. Shi and Y. Deng, Chem. Rev., 2011, 112,
2467–2505.
24 Y. M. A. Yamada, S. M. Sarkar and Y. Uozumi, J. Am. Chem.
Soc., 2012, 134, 3190–3198.
Acknowledgements
´
25 M. Stratakis and H. Garcıa, Chem. Rev., 2012, 112, 4469–
This work was partially supported by a Grant-in-Aid for Scien-
tic Research from JSPS, the Global COE Program, the Univer-
sity of Tokyo, and MEXT (Japan). We also thank Mr. Noriaki
Kuramitsu (The University of Tokyo) for STEM and EDS
analysis.
4506.
´
26 J.-F. Soule, H. Miyamura and S. Kobayashi, J. Am. Chem. Soc.,
2013, 135, 10602–10605.
27 A. E. Allen and D. W. C. MacMillan, Chem. Sci., 2012, 3, 633–
658.
28 N. T. Patil, V. S. Shinde and B. Gajula, Org. Biomol. Chem.,
2012, 10, 211–224.
29 Z. T. Du and Z. H. Shao, Chem. Soc. Rev., 2013, 42, 1337–1378.
Notes and references
1 Nanoparticles and Catalysis, ed. D. Astruc, Wiley-VCH Verlag 30 H. Miyamura, G. C. Y. Choo, T. Yasukawa, W.-J. Yoo and
GmbH & Co. KGaA, 2008. S. Kobayashi, Chem. Commun., 2013, 49, 9917–9919.
2 A. Corma and H. Garcıa, Chem. Soc. Rev., 2008, 37, 2096– 31 M. H. S. A. Hamid, P. A. Slatford and J. M. J. Williams, Adv.
2126. Synth. Catal., 2007, 349, 1555–1575.
3 M. Haruta, T. Kobayashi, H. Sano and N. Yamada, Chem. 32 T. D. Nixon, M. K. Whittlesey and J. M. J. Williams, Dalton
´
Lett., 1987, 405–408.
Trans., 2009, 753–762.
4 M. Haruta, Chem. Rec., 2003, 3, 75–87.
5 M. Haruta, Nature, 2005, 437, 1098–1099.
33 A. J. A. Watson and J. M. J. Williams, Science, 2010, 329, 635–
636.
´
´
´
6 A. Abad, P. Concepcion, A. Corma and H. Garcıa, Angew. 34 G. Guillena, D. J. Ramon and M. Yus, Chem. Rev., 2010, 110,
Chem., Int. Ed., 2005, 44, 4066–4069. 1611–1641.
7 H. Tsunoyama, H. Sakurai, Y. Negishi and T. Tsukuda, J. Am. 35 G. E. Dobereiner and R. H. Crabtree, Chem. Rev., 2010, 110,
Chem. Soc., 2005, 127, 9374–9375. 681–703.
8 C. H. Christensen, B. Jørgensen, J. Rass-Hansen, K. Egeblad, 36 F. Alonso, P. Riente and M. Yus, Acc. Chem. Res., 2011, 44,
R. Madsen, S. K. Klitgaard, S. M. Hansen, M. R. Hansen,
379–391.
H. C. Andersen and A. Riisager, Angew. Chem., Int. Ed., 37 P. J. Black, G. Cami-Kobeci, M. G. Edwards, P. A. Slatford,
2006, 45, 4648–4651.
9 A. S. K. Hashmi and G. J. Hutchings, Angew. Chem., Int. Ed.,
2006, 45, 7896–7936.
M. K. Whittlesey and J. M. J. Williams, Org. Biomol. Chem.,
2006, 4, 116–125.
´
´
38 A. Martınez-Asencio, M. Yus and D. J. Ramon, Synthesis,
10 H. Miyamura, R. Matsubara, Y. Miyazaki and S. Kobayashi,
Angew. Chem., Int. Ed., 2007, 46, 4151–4154.
2011, 2011, 3730–3740.
1726 | Chem. Sci., 2015, 6, 1719–1727
This journal is © The Royal Society of Chemistry 2015

View Article Online
Edge Article
Chemical Science
39 Y. Zhang, C.-S. Lim, D. S. B. Sim, H.-J. Pan and Y. Zhao, 61 N. T. S. Phan, M. van der Sluys and C. W. Jones, Adv. Synth.
Angew. Chem., Int. Ed., 2014, 53, 1399–1403. Catal., 2006, 348, 609–679.
40 K.-i. Shimizu, N. Imaiida, K. Kon, S. M. A. Hakim Siddiki and 62 M. Seki, Synthesis, 2006, 2006, 2975–2992.
A. Satsuma, ACS Catal., 2013, 998–1005.
63 R. Nishio, M. Sugiura and S. Kobayashi, Org. Biomol. Chem.,
41 X. Liu, R. S. Ding, L. He, Y. M. Liu, Y. Cao, H. Y. He and
K. N. Fan, ChemSusChem, 2013, 6, 604–608.
2006, 4, 992–995.
64 D. Astruc, Inorg. Chem., 2007, 46, 1884–1894.
42 X. Cui, Y. Zhang, F. Shi and Y. Deng, Chem.–Eur. J., 2011, 17, 65 L. Yin and J. Liebscher, Chem. Rev., 2007, 107, 133–173.
1021–1028.
66 M. Lamblin, L. Nassar-Hardy, J.-C. Hierso, E. Fouquet and
F.-X. Felpin, Adv. Synth. Catal., 2010, 352, 33–79.
67 A. L. Isfahani, I. Mohammadpoor-Baltork, V. Mirkhani,
A. R. Khosropour, M. Moghadam, S. Tangestaninejad and
R. Kia, Adv. Synth. Catal., 2013, 355, 957–972.
43 F. Li, P. Qu, J. Ma, X. Zou and C. Sun, ChemCatChem, 2013, 5,
2178–2182.
44 K.-i. Fujita, A. Komatsubara and R. Yamaguchi, Tetrahedron,
2009, 65, 3624–3628.
45 A. J. A. Watson, A. C. Maxwell and J. M. J. Williams, J. Org. 68 M. Conte, H. Miyamura, S. Kobayashi and V. Chechik, J. Am.
Chem., 2011, 76, 2328–2331. Chem. Soc., 2009, 131, 7189–7196.
46 R. Akiyama and S. Kobayashi, Chem. Rev., 2009, 109, 594– 69 N. Toshima and T. Yonezawa, New J. Chem., 1998, 22, 1179–
642. 1201.
47 S. Kobayashi and H. Miyamura, Chem. Rec., 2010, 10, 271– 70 X. Peng, Q. Pan and G. L. Rempel, Chem. Soc. Rev., 2008, 37,
290. 1619–1628.
48 S. Kobayashi and H. Miyamura, Aldrichimica Acta, 2013, 46, 71 K. Kaizuka, H. Miyamura and S. Kobayashi, J. Am. Chem.
3–19. Soc., 2010, 132, 15096–15098.
49 H. Miyamura and S. Kobayashi, Acc. Chem. Res., 2014, 47, 72 B. M. Munoz-Flores, B. I. Kharisov, V. M. Jimenez-Perez,
˜
´
´
´
´
1054–1066.
P. E. Martınez and S. T. Lopez, Ind. Eng. Chem. Res., 2011,
50 H. Yuan, W. J. Yoo, H. Miyamura and S. Kobayashi, J. Am.
Chem. Soc., 2012, 134, 13970–13973.
50, 7705–7721.
73 A. K. Singh and Q. Xu, ChemCatChem, 2013, 5, 652–676.
51 W.-J. Yoo, H. Miyamura and S. Kobayashi, J. Am. Chem. Soc., 74 M. Chen, D. Kumar, C.-W. Yi and D. W. Goodman, Science,
2011, 133, 3095–3103. 2005, 310, 291–293.
52 H. Miyamura, R. Matsubara and S. Kobayashi, Chem. 75 L. Kesavan, R. Tiruvalam, M. H. Ab Rahim, M. I. bin Saiman,
Commun., 2008, 2031–2033.
D. I. Enache, R. L. Jenkins, N. Dimitratos, J. A. Lopez-
Sanchez, S. H. Taylor, D. W. Knight, C. J. Kiely and
G. J. Hutchings, Science, 2011, 331, 195–199.
´
53 J.-F. Soule, H. Miyamura and S. Kobayashi, J. Am. Chem. Soc.,
2011, 133, 18550–18553.
´
˜
54 J.-F. Soule, H. Miyamura and S. Kobayashi, Chem. Commun., 76 S. Antoniotti, V. Dalla and E. Dunach, Angew. Chem., Int. Ed.,
2013, 49, 355–357. 2010, 49, 7860–7888.
55 J.-F. Soule, H. Miyamura and S. Kobayashi, Chem.–Asian J., 77 A. Milenkovic, F. Fache, R. Faure and M. Lemaire, Synth.
2013, 8, 2614–2626. Commun., 1999, 29, 1535–1546.
56 Y. Watanabe, T. Ohta and Y. Tsuji, Bull. Chem. Soc. Jpn., 78 W. Ipatiew, Ber. Dtsch. Chem. Ges., 1908, 41, 993–995.
´
1983, 56, 2647–2651.
79 With no Lewis acid, no 5 was accumulated even though the
concentration of 4 was high, whereas with a Lewis acid,
some 5 was accumulated. This demonstrated that the
Lewis acid promoted the formation of the 5.
57 H.-U. Blaser, A. Indolese, A. Schnyder, H. Steiner and
M. Studer, J. Mol. Catal. A: Chem., 2001, 173, 3–18.
58 S. W. Kim, M. Kim, W. Y. Lee and T. Hyeon, J. Am. Chem.
Soc., 2002, 124, 7642–7643.
80 This argument is limited to the steady-state period of 5.
59 R. Akiyama and S. Kobayashi, J. Am. Chem. Soc., 2003, 125, 81 The formation of 6 and 7 was also denitely accelerated by
3412–3413.
the Lewis acid, although we do not have direct evidence on
what the actual hydrogen-accepting intermediates are.
60 K. Mori, T. Hara, T. Mizugaki, K. Ebitani and K. Kaneda, J.
Am. Chem. Soc., 2004, 126, 10657–10666.
This journal is © The Royal Society of Chemistry 2015
Chem. Sci., 2015, 6, 1719–1727 | 1727