Ag nanoparticles on mixed Al2O3-Ga2O 3 supports as catalysts for the N-alkylation of amines with alcohols

Article abstract of DOI:10.1016/j.apcata.2013.09.044

The combination of AgNO₃ with NaH results in Ag nanoparticles that can selectively perform alcohol aminations under mild reaction conditions (110 C). NaH not only serves as a reducing agent for the Ag salt, but also activates the alcohol for dehydrogenation to the corresponding ketone/aldehyde. The stability of the particles can be improved by immobilizing them onto mixed Al₂O₃-Ga₂O₃ supports; the combination of Ga and Al provides materials with stronger Lewis acidic sites compared to pure alumina or gallium oxide supports. This leads to catalysts with enhanced activities, without the necessity of adding external Lewis acids. Detailed TEM characterization also reveals a close interaction between the Ag NPs and the gallium oxide phase. The obtained catalysts are recyclable and show activity for the alcohol amination using a variety of aliphatic and aromatic amines under mild conditions.

Full text of DOI:10.1016/j.apcata.2013.09.044

Applied Catalysis A: General 469 (2014) 373–379
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Ag nanoparticles on mixed Al2O3–Ga2O3 supports as catalysts for the
N-alkylation of amines with alcohols
Inge Geukensa, Frederik Vermoortelea, Maria Meledinab, Stuart Turnerb,
Gustaaf Van Tendeloob, Dirk E. De Vosa,∗
a Department of Microbial and Molecular Systems, KU Leuven, Centre for Surface Chemistry and Catalysis, Kasteelpark Arenberg 23, 3001 Leuven, Belgium
b Electron Microscopy for Materials Science (EMAT), University of Antwerp, Groenenborgerlaan 171, 2020 Antwerp, Belgium
a r t i c l e i n f o
a b s t r a c t
Article history:
The combination of AgNO3 with NaH results in Ag nanoparticles that can selectively perform alcohol
aminations under mild reaction conditions (110 ◦C). NaH not only serves as a reducing agent for the
Ag salt, but also activates the alcohol for dehydrogenation to the corresponding ketone/aldehyde. The
stability of the particles can be improved by immobilizing them onto mixed Al2O3–Ga2O3 supports; the
combination of Ga and Al provides materials with stronger Lewis acidic sites compared to pure alumina or
gallium oxide supports. This leads to catalysts with enhanced activities, without the necessity of adding
external Lewis acids. Detailed TEM characterization also reveals a close interaction between the Ag NPs
and the gallium oxide phase. The obtained catalysts are recyclable and show activity for the alcohol
amination using a variety of aliphatic and aromatic amines under mild conditions.
Received 6 August 2013
Received in revised form
25 September 2013
Accepted 28 September 2013
Available online 9 October 2013
Keywords:
Alcohol amination
Silver
Nanoparticles
Heterogeneous catalysis
Support influence
© 2013 Elsevier B.V. All rights reserved.
1. Introduction
less expensive elements such as Cu [11], Fe [12], Ni [13], or com-
binations thereof [14,15]. Raney Ni has often been used, though
Functionalized secondary or tertiary amines are important
intermediates for pharmaceuticals, polymers, dyes and electronic
materials [1–5]. Catalytic C N coupling of amines with alkyl or aryl
halides allows selective synthesis of these higher amines (Fig. 1a)
[2,6,7]. We recently showed that Ni nanoparticles in phospho-
nium ionic liquids are excellent catalysts for these reactions, even
using aryl chlorides [8]. Nevertheless, halide salts are formed as
side products (Fig. 1a). Working with widely available alcohols
solves this problem, as H2O is the only side product (Fig. 1b) [9].
A downside of using alcohols is that hydroxyl groups are poor
leaving groups, especially compared with halides. Therefore, the
commonly applied reaction sequence is the “hydrogen borrowing”
cycle (Fig. 2). First, the alcohol is dehydrogenated over the metal
catalyst to form a carbonyl compound, which reacts with the amine
to form an imine and water. The imine is reduced to the alkyl-
ated amine product, thereby reoxidizing the catalyst. The metal
thus “borrows” the hydrogen atoms from the alcohol to reduce the
imine [9]. A downside to this cycle is that tertiary alcohols are not
able to react. Catalysts for the hydrogen borrowing reaction include
homogeneous Ru or Ir complexes [9,10]. There is a trend to use
quite drastic reaction conditions are necessary, such as an over-
stoichiometric amount of the catalyst (200 mol%) or of the alcohol
(500 mol%), or high temperatures [10].
One element which has gained attention lately is Ag. This rela-
tively inexpensive metal shows remarkable activity and selectivity
for the coupling of alcohols and amines. The lower stability of the
Ag H bond compared to that of other noble elements (Au, Pd,
Pt) is thought to facilitate the imine reduction [16]. In using Ag
nanoparticles on Al2O3, Shimizu found a synergetic effect between
the acid–base sites of Al2O3 and the Ag sites [16,17]; even then,
high temperatures (140 ◦C) were required and the addition of Lewis
acids like FeCl3·6H2O was necessary to facilitate the imine reduc-
tion. Moreover, the catalysts were only applied to anilines [16].
Later catalysts, e.g. Cu0.95Ag0.05 on Al2O3, were applicable to a
broader substrate range and did not require Lewis acids, but even
higher temperatures were used (155 ◦C) [18]. Deng’s Ag6Mo10O33
catalyst for the alkylation of amines, carboxamides and sulfona-
mides with alcohols requires high temperatures (160 ◦C) and the
addition of 20 mol% of KOtBu [19]. Finally, by increasing the surface
area of the Al2O3 and decreasing the Ag particle size, the reac-
tion can be conducted with smaller amounts of Ag/Al2O3 catalyst
and in milder reaction conditions (120 ◦C), but a base like Cs2CO3
(20 mol%) is still essential for high yields [20]. These examples show
that despite the favourable interactions between the Ag metal and
∗
Corresponding author. Tel.: +32 16 321639.
E-mail address: dirk.devos@biw.kuleuven.be (D.E. De Vos).
0926-860X/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.apcata.2013.09.044

374
I. Geukens et al. / Applied Catalysis A: General 469 (2014) 373–379
Catalyst
M-X
H-B
a)
b)
base (M-B)
Catalyst
'
'
X = I, Br, Cl
M = alkali metal
'
'
Fig. 1. N-alkylation reactions of amines with (a) alkyl or aryl halides and (b) alcohols.
supports like Al2O3, efficient catalysis still often requires extra addi-
tives, like bases or Lewis acids.
analysis of the materials was performed on an Ultima apparatus
equipped with a Burgener atomizer and a radial optic detector,
using Ar as carrier gas and as plasma. N2 physisorption experiments
were performed after pretreatment in N2 at 423 K on a Micromeri-
tics Tristar apparatus at 77 K. FT-IR measurements were performed
with a Nicolet 6700 spectrometer, using a self-supporting sample
wafer. The sample was pretreated at 798 K in air. After cooling to
room temperature, 20 mbar pyridine was introduced and allowed
to adsorb on the sample. The FT-IR spectrum was measured after
temperature programmed desorption (5 K/min) at 423 K. Amines,
alcohols and solvents were used as received from Sigma–Aldrich.
In this paper, we first examine the use of unsupported Ag
nanoparticles in the reaction of 4-ethylaniline and benzyl alcohol.
While significant yield improvements can be realized by changing
the nanoparticle preparation and reaction procedures, the stabil-
ity of the nanoparticles under reaction conditions calls for use of a
support. By changing the acid–base properties of the support, we
succeed not only in increasing the reaction rate in mild reaction
conditions (110 ◦C); moreover, the use of additives becomes super-
fluous. Finally, the substrate scope of the new Ag/Al2O3–Ga2O3
catalysts was explored by testing a variety of aromatic and aliphatic
amines and alcohols.
2.2. Reaction procedure with unsupported Ag NPs
2. Experimental
Representative example (Table 2, entry 3): 0.054 mmol AgNO3
(9.2 mg), 0.5 mmol NaH (60% dispersion in mineral oil; 12 mg) and
0.5 mL of toluene were mixed in a vial; the Ag NPs were formed
by heating this mixture to 110 ◦C under reflux for 1 h. 1.8 mmol of
4-ethylaniline (220 ␮l) and 1.8 mmol benzyl alcohol (190 ␮l) were
added and the reaction mixture was heated to 110 ◦C for 20 h. After
cooling and centrifugation, the supernatant was isolated, washed
with water to remove inorganic impurities and analyzed by GC and
GC–MS.
2.1. General
Reactions were carried out under inert atmosphere in dried,
magnetically stirred glass reaction vials (8 mL) with reflux setup.
Reaction mixtures were analyzed with a Shimadzu 2014 GC
equipped with a CP Sil-8 column and a FID detector using n-
tetradecane as internal standard. Mass spectra were obtained on
an Agilent 6890N GC with a HP-5MS column coupled to a 5973
MSD mass spectrometer. Samples for TEM investigation were pre-
pared by crushing the powders with ethanol in a mortar and placing
several drops of the suspension onto a holey carbon grid. TEM and
HRTEM investigations were performed on a JEOL 3000F transmis-
sion electron microscope, operated at 300 kV. HAADF-STEM and
EDX analysis were performed using a FEI Titan “cubed” microscope,
equipped with an aberration corrector for the probe-forming lens
and a Super-X, large solid-angle EDX detector, operated at 120 kV.
Powder X-ray diffraction (PXRD) patterns were measured with a
STOE COMBI P diffractometer in high throughput mode using a Cu
2.3. Synthesis of the materials
2.3.1. Ag supported on commercial ꢁ-Al2O3 via wet impregnation
Following a literature recipe [16], 37.1 mg AgNO3 was dissolved
in 15 ml of water and this solution was used to suspend 977 mg
of a commercial ␥-Al2O3 support (crushed Saint-Gobain NorPro®
Unispheres®; BET 227 m2/g). After stirring for 4 h at room tem-
perature, the water was evaporated at 80–100 ◦C. The sample was
calcined at 600 ◦C for 1 h (10 ◦C/min). This material is referred to
as Ag(2.4 wt%)/␥-Al2O3 commercial. Part of the Ag-loaded, calcined ␥-
Al2O3 was reduced in a H2 flow for 30 min at 300 ◦C and the other
part was reduced with NaH.
◦
˚
K␣1 source (1.5406 A) with 2ꢀ ranging from −15 to 62 . ICP-AES
2.3.2. Ag supported on ꢁ-Al2O3 via in situ sol–gel synthesis
Following a literature recipe [20], 4.17 ml aluminium tri-sec-
butoxide was added to 31.5 mg AgNO3 together with 0.99 ml
2-butanol, followed by stirring until the AgNO3 was completely dis-
solved and a gel was formed. The gel was stirred for 3 h at 100 ◦C,
after which 2.25 ml H2O was added dropwise, followed by stirring
for 1 h at 100 ◦C. The resulting brown–black solid was centrifuged,
washed with acetone (3×), dried at 100 ◦C overnight and finally
calcined at 600 ◦C for 1 h (10 ◦C/min). This material is referred to as
Ag(2.4 wt%)/␥-Al2O3 sol–gel. The material was reduced with NaH, or
with H2 (30 min, 300 ◦C).
M
MH2
2.3.3. Preparation of the Ag/Al2O3–Ga2O3 series
Representative example Ag/Al2O3–Ga2O3 (0.6) (Table 4, entry 4):
following a literature procedure [21], 1.394 g of Ga(NO3)3·xH2O and
Fig. 2. The hydrogen borrowing cycle.

I. Geukens et al. / Applied Catalysis A: General 469 (2014) 373–379
375
Table 1
1.125 g of Al(NO3)3·9H2O were dissolved in 15 ml of water. This
solution was added very quickly to a vigorously stirred (NH4)2CO3
aqueous solution (5.1 g (NH4)2CO3 in 30 ml H2O) and stirred for
1 h. The precipitate was centrifuged, washed with water (2×) and
ethanol (2×), and dried at 80 ◦C after which the sample was calcined
at 700 ◦C for 3 h (5 ◦C/min). The material was then loaded with Ag
via the wet impregnation technique: 5 ml of a 0.028 M AgNO3 aque-
ous solution was added to 300 mg of the as-synthesized support,
stirred for 4 h at room temperature, followed by water evaporation
at 80–100 ◦C and calcination at 600 ◦C for 1 h (10 ◦C/min). The other
mixed or pure oxides were prepared in similar ways, by adjusting
the quantities of the nitrate salts (see supporting information for
more details). The materials are denoted as Ag/Al2O3–Ga2O3 (x),
with x the Ga/(Ga + Al) content.
Application of Ni, Fe, Cu and Ag NPs in the N-alkylation of 4-ethylaniline with benzyl
alcohol.a
Entry
Metal precursor
X2 (%)b
S3 (%)c
S4 (%)d
1
2
3
4
5
6
7
8
NiBr2
Ni(acac)2
FeCl3
Fe(OAc)2
CuCl2
Cu(OAc)2
AgOAc
AgNO3
31
6
6
10
8
4
90
90e
93e
90e
87e
86e
7
6
7
12
14
34
37
12
12
93
93
7
a
Reaction conditions: benzyl alcohol (1, 1.8 mmol), 4-ethylaniline (2, 1.8 mmol),
metal precursor (0.054 mmol, 3 mol%), lithium tert-butoxide (LiOtBu; 0.054 mmol),
NaH (0.5 mmol), toluene (0.5 ml), 110 ◦C, 20 h, Ar.
b
Conversion of 2.
Selectivity for 3.
Selectivity for 4.
c
d
2.4. Reaction procedure with the supported Ag NPs
e
Other identified products (GC–MS): benzyl benzoate and 4,4ꢀ-
diethylazobenzene.
Representative example (Table 5, entry 1): Prior to use, the catalyst
(162 mg; 3 mol%) was dried at 150 ◦C. Then, NaH (60% dispersion
in mineral oil; 12 mg) and 1 ml of toluene were added, followed
by 1 h heating at 110 ◦C. 1.8 mmol of 4-ethylaniline (220 ␮l) and
3.6 mmol benzyl alcohol (380 ␮l) were added. After 26 h reaction at
110 ◦C, analysis was performed as in Section 2.2. To recycle the cat-
alyst, the material was washed with water after reaction to remove
inorganic impurities, followed by washing with 2-propanol (2×)
and acetone (2×). The material was dried overnight at 150 ◦C after
which again NaH (0.5 mmol; 12 mg) was added and a new run was
started.
better results than in the absence of an extra base. It was also
shown that AgNO3 or NaH alone cannot catalyze the reaction; their
combination is essential for catalytic activity (Table 2, entries 4–5).
Besides reducing the Ag+, excess NaH also easily deprotonates the
benzyl alcohol, which is the compound with the lowest pKa in
the mixture. This results in the formation of an alkoxide, as was
also visually observed by H2 gas formation when adding the alco-
hol. To further examine this, the reaction was also performed by
first mixing amine, alcohol and NaH prior to addition of AgNO3
(entry 6). In such case, no hydride is left, since the benzyl alcohol
is present in excess with respect to NaH. Nevertheless, an imme-
diate reduction of the Ag+ was observed, with formation of a black
suspension. This suggests that in this case, the alkoxide is supply-
ing the electrons for the Ag reduction while it is simultaneously
dehydrogenated to a carbonyl compound. With this procedure, the
same conversion was obtained after 20 h, though selectivity was
significantly lower, indicating that the pre-reduction of the Ag salt
with NaH is still preferred (entry 6). The amount of NaH used is also
important: adding less NaH did not significantly affect conversion,
but lowered selectivity (entry 7). With a large excess of NaH, both
3. Results and discussion
3.1. Reactions with metal nanoparticles
In the search for an economically favourable catalyst that is
able to selectively perform a wide variety of alcohol aminations
under mild conditions, nickel nanoparticles (NPs) were used as
starting point. These were synthesized by the reduction of a nickel
salt in toluene with NaH in the presence of lithium tert-butoxide
(LiOtBu). Such nanoparticles were previously used successfully for
the racemization of primary amines [22], a reaction with mech-
anistic similarities to the hydrogen borrowing cycle. The nickel
particles were applied for the coupling of benzyl alcohol (1) and
4-ethylaniline (2) (Fig. 3). A frequent reoccurring side product is
the imine (4), which is formed when the final step of the hydro-
gen borrowing cycle does not take place. If this is the case, the
hydrogen atoms adsorbed on the Ag surface need to be removed
in order to close the mass and redox balance, e.g. as hydrogen gas
(Fig. 2).
Preliminary results using NiBr2 as metal precursor showed
indeed some conversion (31%), but with only 10% selectivity for the
secondary amine 3, the remainder being the imine (4; Table 1, entry
1). Using another nickel precursor even lowered the conversion
(Table 1, entry 2). Because of the poor selectivity with nickel, we
tested other elements, employing the same synthesis procedure for
the particle formation. As can be seen from Table 1, the selectivity
did not improve when using Fe or Cu nanoparticles, the imine being
the dominant product. Ag, on the other hand, showed markedly
better selectivities of up to 93% for the alkylated amine 3 (Table 1,
entries 7–8). Of the different tested precursors, AgNO3 seemed to
provide the best results and was used for further optimization.
In a next step, the influence of the different additives of the
Ag nanoparticle system, i.e. LiOtBu and NaH, on the reaction was
examined (Table 2). This showed that the presence of LiOtBu is not
necessary for the activity of the NPs (Table 2, entry 2–3). Other
bases were also tested (see supporting information), but none gave
Table 2
Influence of the precursors and additives on the amine alkylation catalyzed by Ag
NPs.a
AgNO3
NaH (mmol)
1:2b
X2 (%)c
S3 (%)d
S4 (%)e
1
–
–
1:1
1:1
1:1
1:1
1:1
1:1
1:1
1:1
0.2
37
38
1
0
93
95
8
32
7
5
2f
3
AgNO3
AgNO3
AgNO3
–
AgNO3
AgNO3
AgNO3
NaH (0.5)
NaH (0.5)
–
NaH (0.5)
NaH (0.5)
NaH (0.25)
NaH (0.75)
4
3
5
1
8
83
48
14
36
6g
7
37
34
19
51
86
64
8
9
10
11
12
AgNO3
AgNO3
AgNO3
AgNO3
NaH (0.5)
NaH (0.5)
NaH (0.5)
NaH (0.5)
1:2
2:1
3:1
4:1
48h
56
58
64
95
95
91
74
5
5
9
26
13i
AgNO3
NaH (0.5)
1:1
9
36
64
a
Reaction conditions as in Table 1, but without LiOtBu.
Molar ratio of benzyl alcohol (1) to 4-ethylaniline (2).
Conversion of 2.
Selectivity for 3.
Selectivity for 4.
b
c
d
e
f
0.054 mmol of LiOtBu was added.
0.5 ml toluene, 1 (1.8 mmol), 2 (1.8 mmol) and 0.5 mmol NaH were mixed and
g
then added to 0.054 mmol AgNO3; 20 h, 110 ◦C, Ar.
h
Conversion of 1.
After recycling of the catalyst, second run.
i

376
I. Geukens et al. / Applied Catalysis A: General 469 (2014) 373–379
3
NP catalyst
toluene, 110 °C, 20h
4
1
2
Fig. 3. N-alkylation of 4-ethylaniline (2) with benzyl alcohol (1) to form N-benzyl-4-ethylaniline (3), with the imine (4) as the side product.
Table 3
3.2. Reactions with Ag on mixed oxides
Effect of the reduction procedure for Ag/␥-Al2O3 catalysts.a
As a starting point, the Ag/␥-Al2O3 catalysts used in literature
were reproduced and activated by reduction under flowing H2
[16,17]. To compare the reduction methods, the materials were
also reduced with NaH. As shown in Table 3, NaH clearly gives bet-
ter catalytic results, irrespective of the method used to prepare the
catalyst. For the heterogenized catalyst, an additional role of NaH
could be to deprotonate the surface hydroxyl groups of ␥-Al2O3.
This would increase the local basicity and enhance the catalyst’s
activity.
To further enhance the activity, gallium was introduced into
the oxide support. Similar Al2O3–Ga2O3 mixed oxides have been
investigated for reactions such as the reduction of NOx by
methane [23,24]. Comparing the acid-base properties of a mixed
Al2O3–Ga2O3 oxide with those of pure Al2O3, it has been found that
a larger number of strong Lewis acidic sites is created by the pres-
ence of Ga [23]. The Al2O3–Ga2O3 materials have a local superficial
spinel structure which differs significantly from bulk Ga2O3, also
leading to enhanced Lewis acidity compared to pure Ga2O3 [24].
The mixed supports were prepared by coprecipitation, with fast
addition of Al(NO3)3 and Ga(NO3)3 to an aqueous (NH4)2CO3 solu-
tion, followed by calcination of the precipitate [25]. The Ga/(Ga + Al)
ratios were varied in the Al2O3–Ga2O3 series between 0 (pure
Al2O3) and 1 (pure Ga2O3). XRD confirmed that the crystal struc-
tures of the Al2O3–Ga2O3 mixed oxides were similar to those
reported in literature (supporting information) [25]. Strikingly, the
aluminas and the solid solutions have much higher specific sur-
face areas than pure Ga2O3 (Table 4). Finally, Fourier transform
infrared (FT-IR) spectroscopy with pyridine chemisorption, e.g. on
Al2O3–Ga2O3 (0.8), confirmed the presence of Lewis acid sites (sup-
porting information).
Catalyst
Reducing
agent
X2 (%)b
S3 (%)c
S4 (%)d
1e
2e
3f
4f
Ag(2.4 wt%)/␥-Al2O3 commercial
Ag(2.4 wt%)/␥-Al2O3 commercial
Ag(2.4 wt%)/␥-Al2O3 sol–gel
Ag(2.4 wt%)/␥-Al2O3 sol–gel
NaH
H2
NaH
H2
44
5
41
15
94
27
88
53
6
73
11
47
a
Reaction conditions: 2.4 mol% Ag catalyst, reduced with 0.5 mmol NaH or using
a hydrogen stream (30ꢀ at 300 ◦C); 1 mL toluene, 1 (3.6 mmol), 2 (1.8 mmol), 20 h,
110 ◦C, Ar.
b
Conversion of 2.
Selectivity for 4.
Selectivity for 3.
c
d
e
Synthesis by wet impregnation of a commercial ␥-Al2O3 support.
The Ag NPs are entrapped in an in situ sol–gel synthesis of the alumina support
f
(see Section 2).
conversion and selectivity were significantly lowered (entry 8). The
optimum amount of NaH for which the highest yield was obtained
was 0.5 mmol for a reaction at the 1.8 mmol scale (entry 3). The
double role of NaH, as reducing agent and as base, is in line with
other Ag systems, which need extra base to facilitate the alcohol
dehydrogenation [20].
Despite the good selectivity of the Ag NPs, the conversion was
still too low (Table 2, entry 3). Therefore, the alcohol: amine ratio
was adjusted. An alcohol: amine ratio of 2:1 provided the best con-
version and selectivity (entry 10). With a larger excess of alcohol,
selectivity was negatively affected (Table 2, entries 11–12). Pro-
longing the reaction for more than 20 h did not increase the yield
any further, showing that the particles lose most of their activity
during reaction. This is due to aggregation and precipitation of the
particles, as was observed visually, and was also shown by the dras-
tic drop in activity after recycling of the Ag catalyst (Table 2, entry
13). Attempts to stabilize the particles by polymer addition were
not successful (see supporting information). For this reason, the
focus was placed on using oxide supports to enhance the stability
of the particles and, simultaneously, boost their activity.
Next, the materials were loaded with Ag via wet impreg-
nation, followed by calcination at 600 ◦C. Some compositional
and physicochemical characteristics of the obtained Ag catalysts
are summarized in Table 4 (entries 1–6; see also supporting
information). For all materials, the surface areas decreased after
Table 4
Characteristics of the Ag/Al2O3–Ga2O3 materials and their turnover frequency (TOF) for the N-alkylation reaction.a
Catalyst
Ag content (wt%)
Ga/(Ga + Al)
BET surface area (m2/g) with Ag
BET surface area (m2/g) without Ag
TOF (1/h)b
1
Ag/Al2O3 (0)
3.1
5.4
0
0
94
n.d.
152
227
0.64
0.68
2c
Ag/␥-Al2O3
commercial
3
4
5
6
Ag/Al2O3–Ga2O3 (0.3)
Ag/Al2O3–Ga2O3 (0.6)
Ag/Al2O3–Ga2O3 (0.8)
Ag/Ga2O3 (1)
4.4
4.0
3.6
3.2
0.3
0.6
0.8
1
149
108
68
197
138
110
24
1.07
1.32
1.63
1.44
15
7d
Ag NPs
–
–
–
–
0.74
a
b
c
Reaction conditions: 1.7 mol % Ag catalyst (0.03 mmol Ag), 0.5 mmol NaH, 1 mL toluene, 1 (3.6 mmol), 2 (1.8 mmol), 20 h, 110 ◦C, Ar.
Mole of product (3) formed per mole of Ag per h after 20 h.
Commercial ␥-Al2O3 support loaded with Ag by wet impregnation.
d
0.03 mmol AgNO3 (5.2 mg) was used. n.d.: not determined.

I. Geukens et al. / Applied Catalysis A: General 469 (2014) 373–379
377
Table 5
N-alkylations of amines with alcohols catalyzed by Ag/Al2O3–Ga2O3 (0.8).a
.
Entry
1g
Alcohol
Amine
X6 (%)b
S7 (%)c
S8 (%)d
S9 (%)e
84
96
3
–
2h
58
85
93
96
6
3
–
–
3g
4f
75
56
86
72
36
14
17
14
58
–
6
79
–
5f,g
6g
7f
CH3(CH2)5NH2
CH3(CH2)7NH2
83
83
3
8
9
CH3(CH2)7OH
20
54
42
4
CH3(CH2)5OH
CH3(CH2)7OH
15
25
53
–
44
96
3
4
10f
CH3(CH2)7NH2
a
Reaction conditions: 3 mol% Ag catalyst, 0.5 mmol NaH, 1 mL toluene, alcohol (3.6 mmol), amine (1.8 mmol), 26 h, 110 ◦C, Ar.
b
c
d
e
f
Conversion of 6.
Selectivity for 7.
Selectivity for 8.
Selectivity for 9.
48 h.
g
h
Other identified products (GC–MS): benzyl benzoate.
Other identified products (GC–MS): benzoic acid, 4-methyl-, (4-methylphenyl)methyl ester.
Ag loading. High resolution transmission electron microscopy
alcohol:amine ratio. Averaged turnover frequencies after 20 h are
displayed in Table 4, together with data for the reaction with the
unsupported Ag NPs in the same conditions. Immobilizing the Ag
particles clearly improves their activity, as all supported Ag cata-
lysts displayed higher TOFs than unsupported Ag NPs, except when
the support was Al2O3, either self-prepared or from a commercial
source (entries 1 and 2). This can be attributed to a suitable stabi-
lization of the Ag NPs on the high surface area mixed oxides, but also
to a synergetic action between the Ag sites and the Lewis acid sites
of the Ga-containing supports [16,17]. The Lewis acidity of the sup-
port could improve the activity of the Ag NPs in various ways. First,
Lewis acidity could result in the withdrawal of some electron den-
sity from the metal NPs towards the metal–support interface, which
would facilitate the alcohol dehydrogenation on the Ag [26]. The
close proximity of Lewis acid sites and Ag sites could also provide
(HRTEM) and high-angle annular dark-field scanning transmis-
sion electron microscopy (HAADF STEM), e.g. performed on the
Ag/Al2O3–Ga2O3 (0.8) sample, showed that the Ga and Al are indeed
present as oxides in the material, whereas the spherical Ag par-
ticles are present in their metallic form (Fig. 4). This was further
confirmed by energy-dispersive X-ray spectroscopy (EDX) map-
ping, in Fig. 5, showing the dispersion of the Ag nanoparticles in
the Al2O3–Ga2O3 matrix. In some cases, the Ag nanoparticles are
covered by a partially crystalline shell (see for example the layered
structure indicated by arrows in Fig. 4b), which was shown to be
GaxOy by spatially resolved electron energy-loss spectroscopy (see
supplementary information).
Finally, the supported Ag catalysts were tested for their activ-
ity in the reference N-alkylation (Fig. 3), using the optimum 2:1

378
I. Geukens et al. / Applied Catalysis A: General 469 (2014) 373–379
a
Fig. 6. Right: Turnover frequency (TOF = mole of product formed per mole of Ag
per h based on GC analysis of the reaction mixture after 26 h of reaction) for the
Ag/Al2O3–Ga2O3 (0.8) catalyst in three runs: R1, R2 and R3. Left: hot filtration exper-
iment: after 6 h, the reaction liquid without the catalyst was allowed to react further
(♦). For comparison, the yield with the Ag catalyst is also displayed (ꢀ). Reaction con-
ditions: 3 mol% Ag catalyst, 0.5 mmol NaH, 1 mL toluene, 1 (3.6 mmol), 2 (1.8 mmol),
110 ◦C, Ar.
adsorption sites for the amine, bringing this reactant in close con-
tact with the aldehyde formed on the Ag sites. Finally, Lewis acids
have previously been shown to accelerate the imine reduction step
of the hydrogen borrowing cycle [16]. Besides enhanced Lewis acid-
ity, Ga2O3 could potentially also contribute by its redox properties,
allowing it to adsorb hydrides, which could also facilitate the imine
reduction step [27,28].
Close inspection of the results in the series of mixed oxides
(Table 4) shows that an increasing Ga content leads to higher
TOFs (entries 3–5). TEM showed that Ag/Al2O3–Ga2O3 (0.3) and
Ag/Al2O3–Ga2O3 (0.8) display a similar mean Ag particle size of
Fig. 4. (a) TEM image of the Ag/Al2O3–Ga2O3 (0.8) sample. The sample consists
of strongly agglomerated, non-uniform nanoparticles. Spherical Ag nanoparticles
(dark contrast features) are spread throughout the agglomerates; (b) HRTEM image
of a Ag nanoparticle imaged along the [0 1 1] zone axis, as evidenced by the inset
FT. The white square indicates the area of the FT; (c) HAADF STEM image of the
Ag/Al2O3–Ga2O3 (0.8) sample. Ag nanoparticles are seen as bright contrast features
and (d) HAADF STEM image of a single Ag nanoparticle. The nanoparticles have a
near-spherical shape and are generally covered by a 1–2 nm partially crystalline
GaxOy shell.
8
1 nm (see supporting information). Hence particle size does not
explain the activity differences. The highest TOF was obtained for
the Ag/Al2O3–Ga2O3 (0.8) catalyst (entry 5). At this high Ga con-
tent, potential interactions between Ag sites and Lewis acid sites
are maximized, while the remaining Al ensures a high surface area.
The pure, Ag-free supports were not catalytically active; they hardly
showed any conversion (<1%) for the alcohol amination reaction.
The best catalyst, Ag/Al2O3–Ga2O3 (0.8), was tested on other
substrates. As in literature often only aniline or aniline deriva-
tives are used, sufficient attention was given to the variation of
the amine (Table 5). Generally, good amine conversions were
obtained with the catalyst, though with the aliphatic alcohols, the
catalysts reacted more slowly (entries 8–10). Using benzylic or
aliphatic amines (entries 4, 6 and 7), the secondary amine:imine
ratio became lower compared to those obtained for aniline sub-
strates (entries 1–3 and 8–9) due to the higher stability of the
imine. When coupling two aliphatic reagents, only the imine was
formed (entry 10). For none of the reactions, tertiary amines were
detected, not even for the aliphatic substrates, probably due to
the application of mild reaction conditions. Noticeable are also the
small amounts of amides encountered in some reaction mixtures
(entries 4, 7–10), indicating that the intermediate hemiaminal,
formed through reaction between the aldehyde and the amine [16],
sometimes undergoes a dehydrogenation, rather than a dehydra-
tion. In the case of morpholine as substrate, the amide was even the
dominant product (entry 5), which has also been observed by oth-
ers when using Ag/Al2O3 catalysts [20,29]. Besides the formation
of amides, traces of the aldehyde formed by the dehydrogenation
of the alcohol were also encountered as side product at high amine
conversions. When comparing the performance of this catalytic
system with other Ag catalysts used in literature [16,20], gener-
ally, similar yields are obtained for similar reactions, but with the
advantage that mild reaction conditions can be applied.
Finally, the recyclability of the catalysts was tested. As can
be seen from Fig. 6, the catalyst can be used at least three times
without loss in activity. Further, to check the heterogeneity of the
reaction, a hot filtration test was performed (Fig. 6). After removal
Fig. 5. HAADF STEM image and EDX maps of the Ag/Al2O3–Ga2O3 (0.8) sample. The
area of EDX map acquisition is marked by a white rectangle. Ga and Al are present
as oxides, the Ag nanoparticles appear in their metallic form.

I. Geukens et al. / Applied Catalysis A: General 469 (2014) 373–379
379
of the catalyst, the yield remained identical even after 20 h of
References
reaction, indicating that Ag does not leach to the reaction liquor
during reaction to homogeneously perform the reaction.
[1] Y.J. Chang, T.J. Chow, Tetrahedron 65 (2009) 9626–9632.
[2] B. Schlummer, U. Scholz, Adv. Synth. Catal. 346 (2004) 1599–1626.
[3] S. Lengvinaite, J.V. Grazulevicius, S. Grigalevicius, B. Zhang, J. Yang, Z. Xie, L.X.
Wang, Synth. Met. 158 (2008) 213–218.
[4] M.R. Dobler, I. Bruce, F. Cederbaum, N.G. Cooke, L.J. Diorazio, R.G. Hall, E. Irving,
Tetrahedron Lett. 42 (2001) 8281–8284.
4. Conclusions
Ag nanoparticles, formed by the reduction of AgNO3 in the
presence of NaH show high selectivity for the desired secondary
amine in a reference alcohol amination reaction using mild reac-
tion conditions (110 ◦C, 20 h). NaH is not only necessary to reduce
the silver salt, but also plays a role in the reaction mechanism by
activating the alcohol for subsequent oxidation by the Ag nanopar-
ticles. In an attempt to enhance their stability, a series of mixed
Al2O3–Ga2O3 supports were synthesized to immobilize the parti-
cles. The supported Ag catalysts show higher turnover frequencies
when immobilized on the mixed Al2O3–Ga2O3 supports compared
to the pure alumina supports. This is related to the higher sur-
face density of strong Lewis acidic sites compared to pure alumina,
caused by the presence of Ga. The mixed oxides with a higher Ga
content display higher turnover frequencies and the best results
were obtained with the support with a Ga/(Ga + Al) ratio of 0.8,
for which TEM investigations revealed a close interaction between
the Ag species and the GaxOy phase. The catalyst proved active for
the alcohol amination of a variety of aliphatic and aromatic amines
using mild reaction conditions and remained active in at least three
runs without loss in activity.
[5] F.B. Koyuncu, S. Koyuncu, E. Ozdemir, Electrochim. Acta 55 (2010)
4935–4941.
[6] A. Guram, R. Rennels, S. Buchwald, Angew. Chem. Int. Ed. 34 (1995)
1348–1350.
[7] J. Louie, J. Hartwig, Tetrahedron Lett. 36 (1995) 3609–3612.
[8] I. Geukens, J. Fransaer, D.E. De Vos, ChemCatChem 3 (2011) 1431–1434.
[9] S. Bähn, S. Imm, L. Neubert, M. Zhang, H. Neumann, M. Beller, ChemCatChem 3
(2011) 1853–1864.
[10] G. Guillena, D.J. Ramón, M. Yus, Chem. Rev. 110 (2009) 1611–1641.
[11] A. Martínez-Asencio, D.J. Ramón, M. Yus, Tetrahedron 67 (2011)
3140–3149.
[12] R. Martinez, D.J. Ramón, M. Yus, Org. Biomol. Chem. 7 (2009) 2176–2181.
[13] K.-i. Shimizu, K. Kon, W. Onodera, H. Yamazaki, J.N. Kondo, ACS Catal. 3 (2013)
112–117.
[14] J. Sun, X. Jin, F. Zhang, W. Hu, J. Liu, R. Li, Catal. Commun. 24 (2012)
30–33.
[15] X. Cui, X. Dai, Y. Deng, F. Shi, Chem. -Eur. J. 19 (2013) 3665–3675.
[16] K. Shimizu, M. Nishimura, A. Satsuma, ChemCatChem 1 (2009) 497–503.
[17] K. Shimizu, K. Sugino, K. Sawabe, A. Satsuma, Chem. -Eur. J. 15 (2009)
2341–2351.
[18] K.-i. Shimizu, K. Shimura, M. Nishimura, A. Satsuma, RSC Adv.
1310–1317.
1 (2011)
[19] X. Cui, Y. Zhang, F. Shi, Y. Deng, Chem. -Eur. J. 17 (2011) 1021–1028.
[20] H. Liu, G.-K. Chuah, S. Jaenicke, J. Catal. 292 (2012) 130–137.
[21] T. Watanabe, Y. Miki, Y. Miyahara, T. Masuda, H. Deguchi, H. Kanai, S. Hosokawa,
K. Wada, M. Inoue, Catal. Lett. 141 (2011) 1338–1344.
[22] I. Geukens, E. Plessers, J.W. Seo, D.E. De Vos, Eur. J. Inorg. Chem. (2013)
2623–2628.
[23] M. Haneda, E. Joubert, J.-C. Menezo, D. Duprez, J. Barbier, N. Bion, M. Daturi,
Acknowledgements
J. Saussey, J.-C. Lavalley, H. Hamada, Phys. Chem. Chem. Phys.
1366–1370.
[24] A.L. Petre, A. Auroux, P. Gélin, M. Caldararu, N.I. Ionescu, Thermochim. Acta 379
(2001) 177–185.
[25] T. Masuda, T. Watanabe, Y. Miyahara, H. Kanai, M. Inoue, Top. Catal. 52 (2009)
699–706.
[26] S. Handjani, E. Marceau, J. Blanchard, J.-M. Krafft, M. Che, P. Mäki-Arvela, N.
Kumar, J. Wärnå, D.Y. Murzin, J. Catal. 282 (2011) 228–236.
[27] W. Jochum, S. Penner, K. Föttinger, R. Kramer, G. Rupprechter, B. Klötzer, J. Catal.
256 (2008) 268–277.
3 (2001)
IG and ST thank the Research Foundation Flanders (FWO) for
financial support as research assistant and postdoctoral researcher
respectively, and under project number G004613N. We are also
grateful to IWT for support in the SBO project MAPIL, to the federal
government for support in IAP-PAI P7/05 project and to the KU
Leuven for the Methusalem grant CASAS.
[28] S.E. Collins, M.A. Baltanás, A.L. Bonivardi, Langmuir 21 (2004) 962–970.
[29] K.-i. Shimizu, K. Ohshima, A. Satsuma, Chem. -Eur. J. 15 (2009) 9977–9980.
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.apcata.
2013.09.044.