Journal of Catalysis 292 (2012) 130–137

Contents lists available at SciVerse ScienceDirect

Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat

N-alkylation of amines with alcohols over alumina-entrapped Ag catalysts

using the ‘‘borrowing hydrogen’’ methodology
Huihui Liu, Gaik-Khuan Chuah, Stephan Jaenicke ⇑
Department of Chemistry, National University of Singapore, 3 Science Drive 3, Kent Ridge, Singapore 117543, Singapore

a r t i c l e i n f o a b s t r a c t

Article history: Ag/Al2O3 catalysts were found to be highly active and selective in the N-alkylation of aniline with a vari-
Received 28 February 2012 ety of aromatic and linear alcohols. In the presence of a base, for example, Cs2CO3 or K3PO4 as promoter,
Revised 4 May 2012 high yields of secondary amines were obtained. The highest activity was found with 2.4 wt.% Ag/Al2O3.
Accepted 5 May 2012
The reaction involves the transient catalytic dehydrogenation of the alkylating alcohol to the correspond-
Available online 8 June 2012
ing carbonyl compound which then reacts spontaneously with the amine. In the final step, the hydrogen
from the alcohol dehydrogenation step is transferred back to the imine, giving the secondary amine as the
Keywords:
final product. With secondary amines, piperidine and pyrrolidine, the hemiaminal intermediate can
Alcohol
Amine
undergo dehydrogenation as well as dehydration to give amides and amines, respectively. The heteroge-
Amide neous nature of Ag/Al2O3 allows for easy recovery and regeneration, which makes it an attractive system
Heterogeneous catalysis in the context of green and sustainable synthesis.
N-alkylation Ó 2012 Elsevier Inc. All rights reserved.
Borrowing hydrogen
Silver

1. Introduction In this reaction, the alcohol is formally dehydrogenated on the

catalyst and forms a carbonyl compound (an aldehyde from a pri-
Nitrogen-containing compounds such as amines and amides mary alcohol), whereas the hydrogen is bound to the catalyst as a
constitute an important group of compounds that are of key impor- hydride species (Scheme 1). The carbonyl compound reacts with
tance in industrial chemical synthesis [1]. Among these nitrogen- the amine under elimination of one molecule of water to form
containing compounds, secondary amines are of particular value. the imine. Hydrogenation of the imine with the abstracted hydro-
There are various methods to prepare them, including electrophilic gen from the first step yields the desired amine. In summary, the
alkylation [2], reductive alkylation [2] and amination of aryl ha- catalyst ‘‘borrows’’ a molecule of hydrogen and ‘‘auto transfers’’ it
lides [3,4]. The reaction of primary amines with alkyl halides in to the intermediate imine to form the final product. There are
the presence of stoichiometric amounts of inorganic bases [5] is in- several advantages to this N-alkylation method, such as the ready
deed the preferred laboratory method. Even though these proto- availability of the required alcohols and the high atom efficiency of
cols have often reached excellent yield levels, there are the transformation where both products from the alcohol dehydro-
drawbacks associated with them, that is, low atom efficiency, high genation (hydrogen and carbonyl compound) are used in the reac-
waste/by-products production and low selectivity toward the de- tion and water is formed as the only by-product [13]. In addition,
sired secondary amines because of the competing formation of this process is highly selective to the desired amine as competing
undesired tertiary amines and alkylammonium halides. One self-condensation of the carbonyl substrate is minimized due to
approach could be the hydroamination of olefins and alkynes [6]. its low concentration. Upon dehydrogenation of the alcohol, the
However, this method is severely limited by the availability of carbonyl substrate reacts to form an imine. This N-alkylation reac-
the corresponding olefin and the electrophilic character of the tion proceeds smoothly in the absence of an additional hydrogen
nitrogen reagent. Another promising protocol is the alkylation of source. In contrast, reductive amination requires high concentra-
amines by alcohols using a reaction sequence that has been vari- tion of the carbonyl substrate and high pressure of hydrogen for
ously termed ‘‘hydrogen autotransfer,’’ ‘‘borrowing hydrogen good yields.
mechanism’’ or ‘‘self-supply system for active hydrogen’’ [7–12]. Many homogeneous transition metal catalysts [14–17], espe-
cially based on ruthenium [18–22] and iridium [23–28], have been
⇑ Corresponding author. Fax: +65 6779 1691. reported for this transformation. Watanabe et al. [18] used
E-mail address: [email protected] (S. Jaenicke). RuCl2(PPh3)3 as the catalyst in the alkylation of aniline derivatives

0021-9517/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jcat.2012.05.007
 H. Liu et al. / Journal of Catalysis 292 (2012) 130–137 131

is important for the dehydrogenation step in the N-alkylation with

alcohols. To facilitate this interaction, we prepared the catalysts
following the procedure of Park’s group [47,48] where nanoparti-
cles of the active metal are entrapped in an in situ – generated
sol–gel alumina support. These authors have shown that such cat-
alysts have excellent activities for a variety of reactions [47–50].
Due to the intimate mixing, the heterogeneous system is expected
to offer a wide substrate tolerance and high catalytic activity.
Scheme 1. Hydrogen autotransfer process from alcohol to imine.

2. Experimental
by primary alcohols and obtained the corresponding N,N-dialkylat-
ed amines as the main product. Other examples include [Ru3(CO)12] 2.1. Synthesis of Al2O3-entrapped Ag catalysts
and N-phenyl-2-(dicyclohexylphosphanyl)pyrrole [21], [Ru(p-cym-
ene)Cl2]2 with DPEPhos 60 as ligand [22] and [IrCpCl2]2 together Alumina-entrapped Ag catalysts were prepared with Ag load-
with a base [23]. The high cost of the stabilizing ligands coupled ings from 0.29 to 13 wt.%. In a typical synthesis for a 2.4 wt.% Ag/
with difficulties in the recovery and reuse is the drawback of these Al2O3 sample, AgNO3 (63 mg/0.37 mmol) and aluminum tri-sec-
homogeneous systems. Hence, the application of efficient, easily butoxide (8.32 ml/32.4 mmol) were dissolved in 2-butanol
recoverable and recyclable heterogeneous catalysts would be (1.98 ml). The resulting gel was stirred at 100 °C for 3 h to give a
clearly advantageous. One of the earliest examples of a heteroge- yellow suspension, following by the dropwise addition of deion-
neous catalyst for N-alkylation is silica gel [29]. Since then, c-alu- ized water (4.5 ml). After stirring for another hour at 100 °C, the
mina [30], solid acids [31,32] and transition metal-based catalysts suspension was cooled to room temperature and filtered. The solid
such as Pd [33,34], Pt [35,36], Ru [37,38] and Ir [39] have been re- catalyst was washed with acetone, dried at 100 °C and calcined for
ported. Concerns about the cost and availability of Pt-group metals 1 h at 600 °C (10 °C min 1 ramp). Prior to use, the catalyst was acti-
have spurred the search for alternative metal catalysts. He et al. [40] vated in flowing H2 at 300 °C for 30 min. The same procedure was
used Cu(OH)x/Al2O3 and Cu(OH)x/TiO2 for the N-alkylation of used to prepare a 2.4 wt.% Cu/Al2O3 catalyst except that copper
amines with activated benzylic alcohols. Although no cocatalyst acetate was used as precursor for the metal.
was required, treatment of the precursor catalyst with NaOH was For comparative purpose, a 2.4 wt.% Ag sample was also synthe-
necessary for activity. Magnetite Fe3O4 also catalyzed N-alkylation, sized by impregnating 0.976 g of commercial c-Al2O3 (Merck,
but only between activated substrates, for example, aromatic 150 m2 g 1) with a 15 ml solution of silver nitrate (37 mg). The
amines and benzylic alcohols [41]. Supported iron oxide nanoparti- suspension was stirred for 4 h at room temperature before increas-
cles are active when used with a base, for example, 1,4-diaza- ing the temperature to 90 °C to evaporate the water. After drying
bicyclo[2.2.2]octane (DABCO) or potassium or sodium tert-butoxide overnight at 100 °C, the sample, 2.4 wt.% Ag/Al2O3–I, was calcined
under microwave irradiation where the temperatures are in the and activated as for the alumina-entrapped Ag samples. The metal
range of 150–170 °C [42]. In general, the activity of these base metal content was determined by ICP-AES.
catalysts is lower than that of precious metals.
Silver is a feasible choice as it is less expensive than metals like
2.2. Procedure for N-alkylation
Au, Pt, Rh, Ir or Ru. However, due to its lower activity for C–H bond
activation, it is currently not as widely used in catalysis as the plat-
Typically, the catalytic N-alkylation of benzyl alcohol and ani-
inum-group metals. Shimizu et al. [43] showed that alumina-sup-
line was carried out using 1.5 mmol aniline and 3 mmol benzyl
ported silver in combination with FeCl3 as a Lewis acidic cocatalyst
alcohol in 5 ml p-xylene. The reaction mixture was placed in a
formed an active system for the N-benzylation of anilines with
25-ml round bottom flask equipped with a condenser. After heat-
alcohols at 140 °C. The high catalytic activity with turnover num-
ing to 120 °C under N2, 0.165 g catalyst and 0.1 g base promoter
ber (TON) of 23 was attributed to a cooperation between coordin-
were added. Aliquots were removed at regular time intervals and
atively unsaturated sites on the very small silver particles
analyzed by gas chromatography. The products were identified
(diameter below 2 nm), acid/base sites on the alumina support
by gas chromatography–mass spectrometry (GC–MS).
and the Lewis acidic Fe(III) cation. The same Ag/Al2O3 catalyst
For lower boiling substrates, the reaction was carried out under
was also active in the one-step synthesis of N-substituted anilines
He in a 50-ml Teflon-lined autoclave (Parr) equipped with mag-
from nitroaromatics and benzyl alcohols under 0.2 MPa hydrogen
netic stirring and sampling port. The reaction mixture consisted
pressure [44]. However, more base (20 mol.% KF) and a reaction
of 3 mmol alcohol, 1.5 mmol amine and 5 ml p-xylene as solvent
temperature of 155 °C were needed. Recently, Shimizu et al. [45]
together with 0.165 g catalyst and 0.1 g Cs2CO3.
investigated alumina-supported copper–silver bimetallic catalysts
for the N-alkylation reaction. They found that at 5 atom% Ag, this
catalyst was highly efficient in the alkylation of anilines and ali- 2.3. Characterization
phatic amines with benzylic and aliphatic alcohols. The bimetallic
structure was proposed to be crucial in the promotion of alcohol The textural properties of the catalysts were determined by
dehydrogenation as well as hydride transfer to the imines. The nitrogen adsorption (Micromeritics Tristar 2000). Prior to the mea-
reaction was carried out without any cocatalyst although a high surements, the each sample was degassed under N2 at 300 °C for
reaction temperature of 155 °C was required. Cui et al. [46] com- 4 h. Powder X-ray diffraction (XRD) was measured with a Siemens
bined the dehydrogenation ability of silver with the nucleophilic D5005 diffractometer equipped with variable slits and a copper an-
addition property of molybdenum. The hybrid oxide, Ag6Mo10O33, ode (Cu Ka k 1.54056 Å). The diffractograms were recorded over a
together with K2CO3 or potassium tert-butoxide gave product 2h range from 30° to 95°. The samples were examined by transmis-
yields of 80–90% in the alkylation of amines, carboxamides, sulfon- sion electron microscopy (TEM, JEOL 3010) operated at 200 kV. The
amides and ketones with aliphatic alcohols. finely ground sample was suspended in 2-propanol, and a drop
The previous studies on supported silver catalysts [43–45] show was placed onto a carbon-coated copper grid and dried before
that a close interaction between the support and the metal catalyst TEM measurements.
132 H. Liu et al. / Journal of Catalysis 292 (2012) 130–137

3. Results and discussion

3.1. Textural properties

The textural properties of the catalysts are summarized in Table

1. The Ag/Al2O3 samples prepared by in situ generation of the sup-
port have high surface areas in excess of 250 m2 g 1 even after cal-
cination at 600 °C. For example, the 2.4 wt.% Ag/Al2O3 sample has a
surface area of 281 m2 g 1, which is almost twice that of the cata-
lyst prepared by wet impregnation onto c-Al2O3. The N2-sorption
isotherms of Ag/Al2O3 samples exhibit type IV isotherms with a
hysteresis loop at P/P°  0.6–0.9, indicative of the presence of
mesopores (Fig. 1). As the Ag loading increased from 0.29 to
13 wt.%, the mean pore diameter decreased from 14.5 to 5.4 nm.
The highly porous nature of the samples is reflected in the large
pore volumes, which decreased from 2.2 to 0.54 cm3 g 1 with high-
er Ag content. This can be attributed to partial blockage of the
pores by the metal aggregates at higher metal loading.
For samples with 0.29–2.4 wt.% Ag, the powder X-ray diffracto-
grams showed only the broad peaks of c-Al2O3 support (Fig. 2). The
absence of reflexes due to Ag suggests that the metal is well dis-
persed on the support and/or that the crystalline size of Ag nano-
particles is below the detection limit of X-ray diffractometry,
typically less than 4 nm. Indeed, transmission electron micro-
graphs of 0.59 and 2.4 wt.% Ag/Al2O3 showed that the Ag nanopar-
ticles were smaller than 3 nm (Figs. S1 and S2). Only at higher Ag
loading, 6.2 and 13 wt.%, did the reflexes of crystalline Ag appear in
the X-ray diffractogram. This agrees well with a mean Ag particle
size of 4.5 nm observed in the transmission electron micrograph
of 6.2 wt.% Ag/Al2O3 (Fig. S3). In comparison with the alumina-en-
trapped Ag catalysts, the sample prepared by wet impregnation
(Fig. S4) had bigger Ag particles. Here, the (1 1 1) reflex of metallic
Ag could be clearly detected in the X-ray diffractogram even at
2.4 wt.% Ag (Fig. S5). Using the Scherrer equation, the Ag crystallite
size for 2.4 wt.% Ag/Al2O3–I was estimated to be 5.4 nm, which is
larger than the mean Ag particle size of 1.7 nm found in the alu-
mina-entrapped 2.4 wt.% Ag/Al2O3 sample. Fig. 1. (i) Nitrogen adsorption–desorption isotherms and (ii) pore size distribution
of Ag/c-Al2O3 with (a) 0.29, (b) 0.59, (c) 1.2, (d) 2.4, (e) 6.2 and (f) 13 wt.% Ag.

3.2. Catalytic activity

3.2.1. N-alkylation of aniline with benzyl alcohol use gave 29% conversion after 19 h (Table 2, entry 4). The interme-
The activity of the catalysts was tested for the N-alkylation of diate imine, N-benzylideneaniline 1, and the secondary amine, N-
aniline with benzyl alcohol (Scheme 2). Benzyl alcohol and aniline benzylaniline 2, were formed in almost equal amounts (51:49).
did not form any products in the absence of a catalyst (Table 2, The addition of a base promoter, Cs2CO3, led to a tremendous in-
entry 1). Similarly, c-Al2O3 synthesized from aluminum tri-sec- crease in the activity, so that >99% conversion was obtained (Table
butoxide and that obtained commercially showed no activity 2, entry 5). Equally striking is the very high selectivity, >99%, to N-
(Table 2, entries 2 and 3). However, 2.4 wt.% Ag/Al2O3 prepared benzylaniline 2. Besides water, the only other product formed was
from aluminum tri-sec-butoxide and hydrogen-activated before the imine 1. No tertiary amine 3 was observed. Activation of the

Table 1
Textural properties of the catalysts.

Catalyst S.A (m2/g) Pore vol. (cm3/g) Mean pore size (nm) Metal particle size (nm)
0.29% Ag/Al2O3 273 1.6 14.5 n.d.
0.59% Ag/Al2O3 360 2.2 17.4 0.99c
1.2% Ag/Al2O3 326 1.1 9.3 n.d.
2.4% Ag/Al2O3 281 0.73 7.7 1.7c
6.2% Ag/Al2O3 257 0.76 6.5 4.5c
13% Ag/Al2O3 262 0.54 5.4 16
2.4% Cu/Al2O3 319 0.34 4.0 n.d.
2.4% Ag/Al2O3–Ia 150 0.27 4.5 5.4
5% Pd/Al2O3b 300 0.33 3.5 n.d.
5% Ru/Al2O3b 112 0.24 5.6 n.d.
a
Wet impregnation on commercial c-Al2O3.
b
Commercial sample from Degussa.
c
From TEM measurements; all others from XRD measurements using the Scherrer equation; n.d.: not detected; particles below detection limit
of XRD (<4 nm).
 H. Liu et al. / Journal of Catalysis 292 (2012) 130–137 133

Table 2
N-alkylation of aniline with benzyl alcohol by various catalysts.

Catalyst Additive Conv. (%)a Selectivity (%)

1 2 3
1 None Cs2CO3 None – – –
2 Synthesized Al2O3 Cs2CO3 None – – –
3 Commercial Al2O3 Cs2CO3 None – – –
4 2.4% Ag/Al2O3 – 29 51 49 –
5 2.4% Ag/Al2O3 Cs2CO3 >99 1 99 –
6 2.4% Ag/Al2O3b Cs2CO3 61 56 44 –
7 2.4% Ag/Al2O3c Cs2CO3 >99 1 99 –
8 2.4% Ag/Al2O3–I Cs2CO3 45 96 4 –
9 2.4% Ag/Al2O3 Na2CO3 17 77 23 –
10 2.4% Ag/Al2O3 K2CO3 43 39 61 –
11 2.4% Ag/Al2O3 K3PO4 96 1 99 –
12 2.4% Cu/Al2O3 Cs2CO3 77 64 36 –
13 5% Pd/Al2O3 Cs2CO3 26 97 3 –
Fig. 2. XRD diffractograms of (a) c-Al2O3 and Al2O3-entrapped samples with (b)
14 5% Ru/Al2O3 Cs2CO3 75 90 10 –
0.29, (c) 0.59, (d) 1.2, (e) 2.4, (f) 6.2 and (g) 13 wt.% Ag.
Reaction conditions: benzyl alcohol (3 mmol), aniline (1.5 mmol), catalyst
(165 mg), promoter (100 mg), xylene (5 ml), 120 °C, under N2.
a
After 19 h.
b
catalyst by pretreatment in a hydrogen flow is essential; without Without H2 activation.
c
this activation step, the conversion was only 61% and the selectiv- Treated in flowing N2 (300 °C/3 h) after H2 activation.

ity to N-benzylaniline 2 was reduced to 46% with N-benzylide-

neaniline 1 being the predominant product (Table 2, entry 6).
The hydrogen pretreatment results in the formation of active
metallic silver particles and the removal of any adsorbed oxygen.
In order to verify whether the hydrogen incorporated into N-ben-
zylaniline 2 comes from hydrogen adsorbed at the metal surface
during the pretreatment or from the dehydrogenation of benzyl
alcohol, the reduced catalyst was flushed with N2 at 300 °C for
3 h. The activity and selectivity of the reaction was unaffected by
this treatment (Table 2, entry 7), indicating that the hydrogen
incorporated into N-benzylaniline 2 comes indeed from the dehy-
drogenation of benzyl alcohol. Comparing the activity of Al2O3-en-
trapped Ag catalyst with that prepared by wet impregnation, it was
found that the conversion after 19 h over the impregnated catalyst,
2.4% Ag/Al2O3-I, was only 45% (Table 2, entry 8), while 2.4% Ag/
Al2O3 had >99% conversion. Furthermore, over 2.4% Ag/Al2O3-I,
the major product was N-benzylideneaniline 1 (96% selectivity)
and not N-benzylaniline 2. The predominant formation of the
imine may be due to a poorer hydrogenation capability of the sam-
Fig. 3. Initial rate () and TOF based on total metal (h) for N-alkylation of aniline by
ple, possibly due to the lower reactivity of the atoms at the surface benzyl alcohol over Al2O3-entrapped Ag catalysts.
of bigger Ag clusters (mean size 5.4 nm) as compared to that for
2.4% Ag/Al2O3 (mean size 1.7 nm). Shimizu et al. [43] also reported
a drop in turnover frequency and amine selectivity for Ag clusters The promotional effect of a base in the dehydrogenation of alcohols
above 2 nm. to form ketones and aldehydes has been observed in both homoge-
As the addition of Cs2CO3 increased the activity significantly, neous and heterogeneous catalytic systems [16,51,52]. It is pro-
other basic promoters were investigated (Table 2, entries 9–11). posed that the base helps in the deprotonation of the primary
The results show that K3PO4 is also an effective promoter with alcohol to form an alkoxide and also facilitates the removal of a hy-
96% conversion of aniline and >99% selectivity to benzylaniline 2. dride from it.

Scheme 2. N-alkylation of aniline with benzyl alcohol.

134 H. Liu et al. / Journal of Catalysis 292 (2012) 130–137

The catalytic activity of 2.4 wt.% Ag/Al2O3 in the presence of the rate and based on the total Ag content, decreased by more than
promoter is higher than that of 5% Ru/Al2O3 and 5% Pd/Al2O3. After an order of magnitude, from 13 to 1.2 h 1, as the Ag loading in-
19 h, full conversion was obtained for 2.4 wt.% Ag/Al2O3, but the creased from 0.29 to 13 wt.%. Similarly, the TOF based on the num-
conversion was only 26% and 75% for 5% Pd/Al2O3 and 5% Ru/ ber of surface Ag atoms also decreased, from 15 to 7.5 h 1 for the
Al2O3, respectively. Comparing the activity on a molar basis, the 0.59–13 wt.% Ag/Al2O3 (Fig. S8). The number of surface Ag atoms
intrinsic activity decreased from 41, 13 and 5 mol aniline was estimated based on the mean diameter of the silver crystallites
converted/mol metal for the alumina-supported Ag, Ru and Pd obtained from TEM and XRD measurements [53,54]. The TEM
catalysts, respectively. For the Ru and Pd catalysts, the main micrographs show that as the Ag loading increased from 0.59 to
product was the secondary imine 1, showing that these catalysts 6.2 wt.%, the mean crystallite size increased from 1.0 to 4.5 nm.
were not as efficient in hydrogenating the intermediate imine to Smaller crystallites offer a higher density of coordinatively unsatu-
N-benzylaniline as Ag. Surprisingly, the Al2O3-entrapped copper rated sites. Such edge and corner atoms are often the active sites
catalyst (Figs. S6 and S7) was more active (18 mol/mol Cu) than for reaction due to their higher reactivity. Furthermore, in the big-
the supported Ru or Pd catalysts. However, at 2.4 wt.% loading, ger crystallites, the number of Ag atoms adjacent to the support is
the selectivity toward the benzylaniline 2 was only 36% as com- reduced. The interaction of Ag with the acid–base sites of alumina,
pared to >99% for the silver-based catalyst. and the promoter plays a significant role in the catalytic cycle [43].
To optimize the N-alkylation of aniline by benzyl alcohol, the This explains the decrease in TOF with the increasing particle size,
effect of silver loading on the Ag/Al2O3 catalyst was studied. The even in the case that the TOF is normalized to surface atoms only.
initial reaction rate (determined from conversion after 1 h) in- The activity of our catalysts with 0.59–6.2 wt.% Ag corresponds
creased steeply up to 2.4 wt.% Ag before decreasing slightly to turnover numbers (TONs) of 16–75. This compares well with the
(Fig. 3). The turnover frequency (TOF), calculated from the initial reported turnover numbers of 18 and 23 for 5 mol.% Ru(OH)x/Al2O3

Table 3
N-Alkylation of amines with alcohols.

Amine Alcohol Products Conv. Sel. (%)

(%) Amine Imine Amide
1 NH2 OH >99 99 1 –
NH

95a 5
97b 3
2 NH2 OH 98 88 12 –
NH

3 NH2 OH 54 16 84 –
NH

4 NH2 CH3OH NH <1 – – –

5 NH2 OH H 5 >99 <1 –

N

6 NH2 OH H 25 80 20 –
N

7 NH2 OH H 69 >99 <1 –

N

8 NH2 CH3(CH2)17OH H <1 – – –

N CH
(CH2)17 3

9 NH2 15 <1 >99 –

N
OH
10 OH 9 7 93 –
NH2 N

11 NH2 OH 95 <1 >99 –

N

12 H OH 51 24 – 76
N N
N O
13 H OH 18 22 – 78
N
N
N O

Reaction condition: alcohol (3 mmol), amine (1.5 mmol), 2.4 wt.% Ag/Al2O3 (165 mg), Cs2CO3 (100 mg), xylene (5 ml), 120 °C, entry 1: under N2, entries 2–13: autogenous
pressure plus 1 atm He in autoclave.
a
First time recycled.
b
Second time recycled.
 H. Liu et al. / Journal of Catalysis 292 (2012) 130–137 135

[37] and 5 wt.% Ag/Al2O3 [43], respectively. Furthermore, the reac- with large amounts of 2-propanol, followed by acetone and reacti-
tion temperatures for these catalysts were higher (132 and 140 °C) vated in H2 at 300 °C. The rate of reaction was very similar to the
than the 120 °C used in this study. fresh catalyst, and N-benzylaniline was formed with 95–97% selec-
tivity at >99% conversion (Table 3, entry 1).
3.2.2. Other substrates
The N-alkylation of other substrates was investigated using the 3.2.4. Mechanistic studies
heterogeneous 2.4 wt.% Ag/Al2O3 catalyst in an autoclave at autog- In the N-alkylation of aniline with benzyl alcohol catalyzed by
enous pressure (plus 1 atm He). The rate of N-alkylation of aniline 2.4 wt.% Ag/Al2O3, only two products, N-benzylideneaniline 1 and
increased with the chain length of linear alkyl alcohols (Table 3, N-benzylaniline 2, were detected. The reaction profile shows the
entries 4–7). In all cases, the main product was the secondary consecutive transformation of aniline first to 1 and then to 2
amine. With methanol, the reaction was sluggish, and less than (Fig. 5). The concentration of 1 increased to a maximum at 4 h be-
1% N-methylaniline was formed. With 1-butanol and 1-pentanol, fore decreasing with time as it transformed to 2. From the activity
5% conversion and 25% conversion were observed, respectively, results with and without a base (Table 2), we postulate that a cru-
after 19 h. 1-Hexanol (Table 3, entry 7) reacted with aniline to give cial step is the formation of benzaldehyde from benzyl alcohol. We
a satisfactory yield of 69% N-hexylaniline. However, no reaction oc- have previously reported that K3PO4 is a highly efficient solid base
curred when long-chained 1-octadecanol was used as the alkylat- catalyst in the catalytic transfer hydrogenation of aldehydes using
ing agent (Table 3, entry 8). Possibly, the long C18 chain poses a alcohols without needing a metal catalyst [55]. The low reaction
steric hindrance in the transformation steps to the amine, in addi- rate when the catalyst is present without the base (Table 2, entry
tion to diffusion limitations due to its bulkier size. Indeed, the reac- 4) as compared to the rate enhancement when present together
tion between cyclohexylmethanol and aniline gave only 15% (Table 2, entries 5 and 11) shows the synergistic effect of Ag/
conversion after 19 h. The predominant product was the imine, Al2O3 and Cs2CO3 (or K3PO4). However, without any hydrogen
N-(cyclohexylmethylene)aniline (Table 3, entry 9) formed with acceptor, for example, oxygen or imine, the presence of the catalyst
the elimination of an equivalent of H2. Alkylation of secondary and base promoter is not enough to catalyze the dehydrogenation
amines, 1-phenylethanamine and cyclohexylamine, with benzyl of benzyl alcohol to benzaldehyde. A low yield of <2% was recorded
alcohol also yielded imines as the major product (Table 3, entries (Table 4, entry 1). The addition of a hydrogen acceptor such as N-
10 and 11). These results suggest that hydrogenation of the imines benzylideneaniline 1 is necessary for dehydrogenation to occur,
could be rate-determining due to steric hindrance. Interestingly, and benzaldehyde was formed in 15% yield (Table 4, entry 2).
the secondary amines, piperidine and pyrrolidine (Table 3, entries The coupling of benzaldehyde with aniline yielded 100% N-benzy-
12 and 13) formed tertiary amines as well as amides. The latter lideneaniline 1 within 1 h without needing any catalyst (Table 4,
was the major product with 76–78% selectivity. The hemiaminal entry 3). Using gaseous hydrogen, N-benzylideneaniline 1 was
formed from the secondary amine and the carbonyl species prefer- readily reduced to N-benzylaniline 2 with a yield of 80% (Table 4,
entially undergoes dehydrogenation rather than dehydration. entry 4). In comparison, the catalytic transfer hydrogenation of 1
with benzyl alcohol proceeded efficiently over the Ag/Al2O3 cata-
3.2.3. Heterogeneity of reaction and reuse of catalyst lyst to give 61% yield of 2 (Table 4, entry 2). An inert atmosphere
In order to test for leaching and to confirm that the reaction is is essential to ensure that the reaction can proceed to the desired
indeed heterogeneously catalyzed, the reaction was allowed to product 2 because under aerobic conditions, aniline and benzyl
run initially in the presence of 2.4% Ag/Al2O3 catalyst (Fig. 4). After alcohol react to give exclusively 1 (Table 4, entry 5).
2 h, the catalyst was removed from the reaction mixture by centri- Based on these observations, we propose that the reaction pro-
fugation. The clear supernatant was again brought to reaction con- ceeds with the transitory formation of the Ag hydride species
ditions. Over a period of next 16 h, the conversion remained (Scheme 3). The role of the base is connected with the dehydroge-
unchanged and no further reaction was observed. These experi- nation step to form the carbonyl compound since the reaction of
ments confirm that the active phase did not leach from the catalyst the imine to amine proceeds smoothly without base. Possibly,
under the reaction conditions. the base aids in deprotonation of the alcohol to form an alkoxide
The activity of 2.4% Ag/Al2O3 was tested for reuse. Following the at a Lewis acid site of the alumina support. The removal of a
batch reaction, the catalyst was recovered by filtration, washed

Fig. 5. Reaction profile for N-alkylation of aniline by benzyl alcohol over 2.4 wt.%
Ag/Al2O3. () Aniline, (d) N-benzylideneaniline and (h) N-benzylaniline. Reaction
Fig. 4. N-alkylation of aniline with benzyl alcohol in the presence of 2.4 wt.% Ag/ conditions: benzyl alcohol (3 mmol), aniline (1.5 mmol), xylene (5 ml), catalyst
Al2O3 (N) and after catalyst removal at 2 h (e). (165 mg), Cs2CO3 (100 mg), 120 °C.
136 H. Liu et al. / Journal of Catalysis 292 (2012) 130–137

Table 4
Reaction studies and conditions.

Reaction Conditions
1 O He (1 bar) 2.4 wt.% Ag/Al2O3 (165 mg), Cs2CO3 (100 mg) 120 °C, 19 h

OH H + H2

3 mmol < 2 % yield

2 He (1 bar) 2.4 wt.% Ag/Al2O3 (165 mg), Cs2CO3 (100 mg) 120 °C, 19 h
H O
N N
OH + H
+

1 mmol 4 mmol 61 % 15 %
3 O He (1 bar) 120 °C, 1 h
NH2 N
H + + OH

3 mmol 1.5 mmol 100 % 24 %

4 H2 (1 bar), 2.4 wt.% Ag/Al2O3 (165 mg) 120 °C, 19 h
H
N H2 N

80 %
5 Open to air, 2.4 wt.% Ag/Al2O3 (165 mg), Cs2CO3 (100 mg), 120 °C, 25 h
NH2
OH N + H2O
+

1.5 mmole 3 mmole 100 %

4. Conclusion

Silver-based catalysts were generated by in situ precipitation of

silver particles in an aluminum oxyhydroxide support followed by
calcination at 600 °C. The resulting Ag/Al2O3 samples were active
for the N-alkylation of amines with alcohols with 2.4 wt.% Ag being
optimal. The higher activity compared to samples prepared by wet
impregnation shows that an intimate mixing of silver and the
support is important. The addition of Cs2CO3 or K3PO4 as a base
promoter improved the activity and selectivity to amines. N-Ben-
zylaniline was formed in >99% yield from aniline and benzyl
alcohol. The base promoter is necessary for the efficient dehydro-
genation of the alcohol to the corresponding carbonyl compound
that then reacts spontaneously with the primary or secondary
amine to form the imine. Hydrogenation of the imine to the desired
amine takes place under an inert atmosphere using the self-gener-
ated hydrogen from the alcohol dehydrogenation step. In addition
to the tertiary amines, amides were formed when piperidine and
pyrrolidine were reacted with benzyl alcohols. The heterogeneous
Ag/Al2O3 can be recycled and reused without loss of activity and
selectivity.

Acknowledgment

Scheme 3. Proposed reaction steps for the N-alkylation of primary and secondary Financial support from National University of Singapore under
amines by alcohol catalyzed by Ag/Al2O3. Grant number R-143-000-374-112 is gratefully acknowledged.

hydride from the alkoxide by the Ag catalyst leads to the formation Appendix A. Supplementary material
of the carbonyl compound. The carbonyl compound readily reacts
with a primary amine to form the corresponding imine. Finally, the Supplementary data associated with this article can be found, in
hydride and a proton from the alumina are transferred back to the the online version, at http://dx.doi.org/10.1016/j.jcat.2012.05.007.
imine to afford the corresponding secondary amine. Starting with a
secondary amine, both the amine and amide can form. Due to sta- References
bilization by the electron donating alkyl substituents of the sec-
ondary amine, the hemiaminal intermediate has a longer life [1] A. Seayad, M. Ahmed, H. Klein, R. Jackstell, T. Gross, M. Beller, Science 297
time than that formed from a primary amine. Hence, it can under- (2002) 1676.
[2] M.B. Smith, J. March, March’s Advanced Organic Chemistry: Reactions,
go further hydride elimination forming the amide. Imine-like inter- Mechanisms, and Structure, Wiley, Hoboken, NJ, 2007. p. 711.
mediates are readily hydrogenated to the amine. [3] B. Schlummer, U. Scholz, Adv. Synth. Catal. 346 (2004) 1599.
 H. Liu et al. / Journal of Catalysis 292 (2012) 130–137 137

[4] D.A. Surry, S.L. Buchwald, Angew. Chem., Int. Ed. 47 (2008) 6338. [30] F. Valot, F. Fache, R. Jacquot, M. Spagnol, M. Lemaire, Tetrahedron Lett. 40
[5] R.N. Salvatore, C.H. Yoon, K.W. Jung, Tetrahedron 57 (2001) 7785. (1999) 3689.
[6] M. Beller, A. Tillack, J. Seayad, C. Bolm, Transition Metals for Organic Synthesis: [31] B. Das, P.R. Reddy, C. Sudhakar, M. Lingaiah, Tetrahedron Lett. 52 (2011) 3521.
Building Blocks and Fine Chemicals, John Wiley & Sons, Canada, 2004. 406. [32] M. Nehate, V.V. Bokade, Appl. Clay Sci. 44 (2009) 255.
[7] J. Pasek, P. Kondelik, P. Richter, Ind. Eng. Chem. Prod. Res. Dev. 11 (1972) 333. [33] Y. Zhang, X. Qi, X. Cui, F. Shi, Y. Deng, Tetrahedron Lett. 52 (2011) 1334.
[8] M.G. Edwards, J.M.J. Williams, Angew. Chem., Int. Ed. 41 (2002) 4740. [34] A. Corma, T. Ródenas, M.J. Sabater, Chem. Eur. J. 16 (2010) 254.
[9] M.H.S.A. Hamid, P.A. Slatford, J.M.J. Williams, Adv. Synth. Catal. 349 (2007) [35] H. Hamada, Y. Kuwahara, T. Sato, K. Wakabayashi, Bull. Chem. Soc. Jpn. 60
1555. (1987) 55.
[10] R. Martinez, D.J. Ramón, M. Yus, Tetrahedron 62 (2006) 8988. [36] J.T. Richardson, W.C. Lu, J. Catal. 42 (1976) 275.
[11] G.E. Dobereiner, R.H. Crabtree, Chem. Rev. 110 (2010) 681. [37] J.W. Kim, K. Yamaguchi, N. Mizuno, J. Catal. 263 (2009) 205.
[12] G. Guillena, D.J. Ramón, M. Yus, Angew. Chem. Int. Ed. 46 (2007) 2358. [38] K. Yamaguchi, J. He, T. Oishi, N. Mizuno, Chem. Eur. J. 16 (2010) 7199.
[13] S. Bähn, S. Imm, L. Neubert, M. Zhang, H. Neumann, M. Beller, ChemCatChem 3 [39] H. Ohta, Y. Yuyama, Y. Uozumi, Y.M.A. Yamade, Org. Lett. 13 (2011) 3892.
(2011) 1853. [40] J. He, K. Yamaguchi, N. Mizuno, Chem. Lett. 39 (2010) 1182.
[14] R. Grigg, T.R.B. Mitchell, S. Sutthivaiyakit, N. Tongpenyai, Chem. Commun. [41] R. Martínez, D.J. Ramón, M. Yus, Org. Biomol. Chem. 7 (2009) 2176.
(1981) 611. [42] C. Gonzalez-Arellano, K. Yoshida, R. Luque, P.L. Gai, Green Chem. 12 (2010)
[15] Y. Watanabe, Y. Tsuji, Y. Ohsugi, Tetrahedron Lett. 22 (1981) 2667. 1281.
[16] A. Martínez-Asencio, D.J. Ramón, M. Yus, Tetrahedron 67 (2011) 3140. [43] K. Shimizu, M. Nishimura, A. Satsuma, ChemCatChem 1 (2009) 497.
[17] T.D. Nixon, M.K. Whittlesey, J.M.J. Williams, Dalton Trans. (2009) 753. [44] K. Shimizu, K. Shimura, M. Nishimura, A. Satsuma, ChemCatChem 3 (2011)
[18] Y. Watanabe, Y. Tsuji, H. Ige, Y. Ohsugi, T. Ohta, J. Org. Chem. 49 (1984) 3359. 1755.
[19] A. Arcelli, B.T. Khai, G. Porzi, J. Organomet. Chem. 235 (1982) 93. [45] K. Shimizu, K. Shimura, M. Nishimura, A. Satsuma, RSC Adv. 1 (2011) 1310.
[20] G. Guillena, D.J. Ramón, M. Yus, Chem. Rev. 110 (2010) 1611. [46] X. Cui, Y. Zhang, F. Shi, Y. Deng, Chem. Eur. J. 17 (2011) 1021.
[21] A. Tillack, D. Hollmann, K. Mevius, D. Michalik, S. Bähn, M. Beller, Eur. J. Org. [47] M.S. Kwon, N. Kim, S.H. Seo, I.S. Park, R.K. Cheedrala, J. Park, Angew. Chem. Int.
Chem. (2008) 4745. Ed. 44 (2005) 6913.
[22] A.J.A. Watson, A.C. Maxwell, J.M.J. Williams, J. Org. Chem. 76 (2011) 2328. [48] I.S. Park, M.S. Kwon, N. Kim, J.S. Lee, K.Y. Kang, J. Park, Chem. Commun. (2005)
[23] K.-i. Fujita, Y. Enoki, R. Yamaguchi, Tetrahedron 64 (2008) 1943. 5667.
[24] Y. Ishii, S. Sakaguchi, Bull. Chem. Soc. Jpn. 77 (2004) 909. [49] W.-H. Kim, I.S. Park, J. Park, Org. Lett. 8 (2006) 2543.
[25] K.-i. Fujita, Z. Li, N. Ozeki, R. Yamaguchi, Tetrahedron Lett. 44 (2003) 2687. [50] J.H. Choi, N. Kim, Y.J. Shin, J.H. Park, J. Park, Tetrahedron Lett. 45 (2004) 4607.
[26] D. Gnanamgari, E.L.O. Sauer, N.D. Schley, C. Butler, C.D. Incarvito, R.H. Crabtree, [51] K. Shimizu, R. Sato, A. Satsuma, Angew. Chem. Int. Ed. 48 (2009) 3982.
Organometallics 28 (2009) 321. [52] L. Liu, L. Ji, Y. Wei, Catal. Commun. 9 (2008) 1379.
[27] O. Saidi, A.J. Blacker, G.W. Lamb, S.P. Marsden, J.E. Taylor, J.M.J. Williams, Org. [53] R. Van Hardeveld, F. Hartog, Surf. Sci. 15 (1969) 189.
Process Res. Dev. 14 (2010) 1046. [54] A. Abd, A. Corma, H. García, Chem. Eur. J. 14 (2008) 212.
[28] R. Kawahara, K-i. Fujita, R. Yamaguchi, Adv. Synth. Catal. 353 (2011) 1161. [55] R. Radhakrishnan, D.M. Do, S. Jaenicke, Y. Sasson, G.K. Chuah, ACS Catal. 1
[29] A.B. Brown, E.E. Reid, J. Am. Chem. Soc. 46 (1924) 1836. (2011) 1631.