Reductive amination of 2-propanol to monoisopropylamine over Co/γ-Al2O3 catalysts

Article abstract of DOI:10.1016/j.apcata.2012.01.011

Co/γ-Al₂O₃ catalysts with 4-27 wt% cobalt loadings were prepared by incipient-wetness impregnation and used to catalyze the synthesis of monoisopropylamine by the reductive amination of 2-propanol in the presence of hydrogen and ammonia. The catalysts were characterized by X-ray diffraction, H₂-temperature programmed reduction, N ₂-sorption, and H₂-chemisorption. 23 wt% Co loading resulted in the highest catalytic activity and a long-term stability of up to 100 h on stream. 2-Propanol conversion was related to the exposed metal surface area and the number of exposed cobalt atoms. In the absence of hydrogen, the catalyst was progressively deactivated; its initial activity and selectivity were completely recovered upon re-exposure to hydrogen. The deactivation was due to the formation of metal nitride caused by the strong adsorption of ammonia on the surface of the metal phase. Excess hydrogen hindered the phase transition to metal nitride, preventing deactivation.

Full text of DOI:10.1016/j.apcata.2012.01.011

Applied Catalysis A: General 417–418 (2012) 313–319
Contents lists available at SciVerse ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Reductive amination of 2-propanol to monoisopropylamine over Co/␥-Al2O3
catalysts
Jun Hee Choa, Jung Hyun Parka, Tae-Sun Changb, Gon Seoc, Chae-Ho Shina,∗
a Department of Chemical Engineering, Chungbuk National University, Chungbuk 361-763, Republic of Korea
b Environment & Resources Research Center, Korea Research Institute of Chemical Technology, Deajeon 305-353, Republic of Korea
c School of Applied Chemical Engineering and the Institute for Catalysis Research, Chonnam National University, Gwangju 500-757, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
Article history:
Co/␥-Al2O3 catalysts with 4–27 wt% cobalt loadings were prepared by incipient-wetness impregnation
and used to catalyze the synthesis of monoisopropylamine by the reductive amination of 2-propanol
in the presence of hydrogen and ammonia. The catalysts were characterized by X-ray diffraction, H2-
temperature programmed reduction, N2-sorption, and H2-chemisorption. 23 wt% Co loading resulted in
the highest catalytic activity and a long-term stability of up to 100 h on stream. 2-Propanol conversion
was related to the exposed metal surface area and the number of exposed cobalt atoms. In the absence of
hydrogen, the catalyst was progressively deactivated; its initial activity and selectivity were completely
recovered upon re-exposure to hydrogen. The deactivation was due to the formation of metal nitride
caused by the strong adsorption of ammonia on the surface of the metal phase. Excess hydrogen hindered
the phase transition to metal nitride, preventing deactivation.
Received 10 November 2011
Received in revised form 6 January 2012
Accepted 7 January 2012
Available online 16 January 2012
Keywords:
2-Propanol
Acetone
Reductive amination
Cobalt
© 2012 Elsevier B.V. All rights reserved.
Monoisopropylamine
1. Introduction
hydrogenation and dehydrogenation reactions or the oxidized state
for oxidation [17–21]. Gardner et al. [22] reported cobalt catalysts
The reductive amination of aliphatic alcohols is used to produce
alkylamines that can be used as intermediates for the synthe-
sis of herbicides, insecticides, pharmaceutical chemicals, corrosion
inhibitors, plastics, and rubber chemicals [1–5]. The process can
proceed by hydrogenation–dehydrogenation over nickel, cobalt,
copper, and solid acid catalysts; it is metal-catalyzed, with alcohol
conversion being proportional to the exposed metal surface area
[6–10].
Monoisopropylamine (MIPA) is generally synthesized by con-
tacting 2-propanol and ammonia at high pressure (≈20 bar) [11,12].
The reductive amination of 2-propanol over supported metal cata-
lysts is complex and involves dehydrogenation, condensation, and
hydrogenation steps (Scheme 1), with the dehydrogenation of 2-
propanol to acetone being the rate determining step. The first and
last redox processes are catalyzed by the metal; the reaction of the
intermediate carbonyl compound with NH3 to form an imine can
be accelerated by acid/base catalysis [13–16].
with higher activity than supported nickel catalysts for reductive
amination with high metal loadings and high space velocity at rela-
tively low temperature. However, conversion and yield are usually
low in the synthesis of alkylamines from alcohols and ammonia at
atmospheric pressure.
In batch-type reductive amination, industrial separation and
recycling of catalyst and by-product are difficult. The reductive ami-
nation of alcohols has been widely investigated. Dobson et al. [23]
reported the reductive amination of 2-propanol over supported
nickel–palladium–ruthenium catalysts at high pressures (≈50 bar)
in an autoclave. However, most studies were focused on methanol,
ethanol and cyclohexanol [5,24–26]. There are few reports regard-
ing the influence of reaction parameters on the reductive amination
of 2-propanol over supported metal catalysts.
This work reports the catalytic properties of Co/␥-Al2O3 cata-
lysts for the reductive amination of 2-propanol in the presence of
ammonia and hydrogen at atmospheric pressure. 2-Propanol con-
version and selectivities to MIPA, acetone, diisopropyamine (DIPA),
and diisopropylether (DIPE) were evaluated on Co/␥-Al2O3 cata-
lysts with various Co loadings. Various hydrogen and ammonia
feed compositions and reaction temperatures were tested. Cal-
cined and reduced catalysts were characterized by X-ray diffraction
(XRD), H2-temperature programmed reduction (H2-TPR), NH3-
temperature programmed desorption (NH3-TPD), N2-sorption, and
H2-chemisorption.
Cobalt-based catalysts are industrially important in a variety
of reactions. The redox properties of cobalt species allow their
use in either the reduced state for Fischer–Tropsch synthesis and
∗
Corresponding author. Tel.: +82 43 261 2376; fax: +82 43 269 2370.
E-mail address: chshin@chungbuk.ac.kr (C.-H. Shin).
0926-860X/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2012.01.011

314
J.H. Cho et al. / Applied Catalysis A: General 417–418 (2012) 313–319
Scheme 1. Reaction pathways in the reductive amination of 2-propanol.
vacuum pump providing 10−6 Torr. Prior to adsorption exper-
2. Experimental
iments, each 0.3 g sample was reduced at 600 ◦C for 3 h. H2
chemisorption uptakes were separately determined as the differ-
ence of two successive isotherm measurements. Metal dispersion
and cobalt metal surface area were calculated assuming a H/Co
stoichiometry of 1. The degree of reduction was determined by
O2 titration and the sizes of the cobalt particles were corrected
considering the degree of reduction, which was determined by
the following equation: [the amount of O2 consumption (mmol
O2; 3Co + 2O2 → Co3O4)]/[the theoretical amount of H2 consump-
tion with the assumption of fully reduced cobalt oxides (mmol H2;
Co3O4 + 4H2 → 3Co + 4H2O)] × 100) [17,27].
2.1. Catalyst preparation
Co/␥-Al2O3 catalysts with 4–27 wt% cobalt loadings were pre-
pared by incipient wetness impregnation using Co(NO)3·6H2O
(Sigma–Aldrich) solution on commercial ␥-Al2O3 (Procatalyse,
194 m2 g−1). The impregnated catalysts were dried at 100 ◦C
overnight and subsequently calcined at 500 ◦C for 2 h in a muffle
furnace under flowing air (200 cm3 min−1). The final catalysts are
labeled Co(x)/Al2O3, with x denoting the weight percent of Co metal
(4, 11, 17, 23 or 27 wt%).
To identify the species adsorbed over the catalysts, the evolu-
tion of NH3 and H2 were examined during TPD/TPR using a Balzers
QMS200 quadruple mass spectrometer (QMS). The QMS signals
of •NH3 (m/e = 17) and •H2 (m/e = 2) were recorded. Prior to mea-
surement, samples were reduced under a 50 cm3 min−1 H2 flow at
2.2. Catalyst characterization
The catalysts were degassed for 6 h at 250 ◦C before their
BET surface areas, pore volumes and pore size distributions were
determined by N2 physisorption at −196 ◦C using Micromeritics
ASAP2020 apparatus. Surface areas were calculated over a 0.01–0.2
relative pressure range. Pore size distributions were calculated
from the desorption branch using BJH formulism. X-ray diffrac-
tion (XRD) patterns were recorded on a Rigaku D/MAX 2500H
diffractometer using Cu K␣ radiation to identify the phases of Co/␥-
Al2O3. Co particle sizes were determined from the broadening of
the diffraction peaks using Scherrer’s equation. H2-temperature-
programmed reduction (H2-TPR) using a Micromeritics AutoChem
II 2920 instrument determined the reducibility of cobalt oxides.
Before measurement, each 0.1 g sample was pretreated under flow-
ing He (50 cm3 min−1) up to 400 ◦C and held for 1 h to remove
adsorbed water and other contaminants before being cooled to
30 ◦C under flowing He. Reducing gas containing 5% H2/Ar mix-
ture (50 cm3 min−1) was then passed with 15 ◦C min−1 heating to
800 ◦C. A thermal conductivity detector determined the amount of
H2 consumed.
600 ◦C for 3 h with heating at 10 ◦C min−1
.
2.3. Catalytic activity tests
All catalytic experiments were performed at atmospheric pres-
sure under a continuous flow in a fixed bed microreactor. Prior
to testing, the catalysts were activated at 600 ◦C for 3 h under a
50 cm3 min−1 flow of purified H2 and kept at the desired tem-
perature to establish a standard operating procedure, allowing
time for the product distribution to stabilize. Standard reac-
tion procedures were considered under the following conditions:
catalyst = 0.1 g; T = 210 ◦C; WHSV = 4.29 h−1, feed composition of 2-
propanol/NH3/H2/N2 = 1/4/6/22.8. Molar ratios of H2/2-propanol
and NH3/2-propanol were varied in the ranges 2–12, and 2–16,
respectively, and reaction temperatures of 170–250 ◦C were tested.
The total flow rate was fixed at 90 cm3 min−1 using nitrogen dilu-
ents to maintain constant partial pressures of the other reactants.
The partial pressure of 2-propanol was constant 3 kPa. The reac-
tion products were analyzed on-line using a Chrompack-CP-9001
The dispersion of cobalt and metal surface area was mea-
sured by H2 chemisorption at 100 ◦C under static conditions using
a Micromeritics ASAP 2020C instrument equipped with a high

J.H. Cho et al. / Applied Catalysis A: General 417–418 (2012) 313–319
315
Table 1
Physical properties of Co/␥-Al2O3 catalysts.
Co loading
(wt%)
Surface area
Pore volume
Pore size
(nm)
(m2 g−1
)
(cm3 g−1
)
4
11
17
23
27
163
144
137
124
119
0.60
0.50
0.40
0.38
0.34
15
14
12
12
11
gas chromatograph equipped with a CP-Volamine capillary column
(60 m × 0.32 mm) and a flame ionization detector.
The conversion of 2-propanol and selectivity to MIPA was
defined as follows:
Conversion (%)
2-propanolfeed (mol) − 2-propanolunreacted (mol)
=
× 100
× 100
2-propanolfeed (mol)
Selectivity (%)
MIPAproduced (mol)
=
2-prppanolfeed (mol) − 2-propanolunreacted (mol)
3. Results and discussion
3.1. Characterization
The physical properties of Co/␥-Al2O3 catalysts calcined are
summarized in Table 1. The BET surface areas of the catalysts
decreased with increasing cobalt loading. The specific surface area,
pore volume and pore size were progressively decreased as the
cobalt loading increased from 4 to 27 wt%. This may be due to par-
tial blockage of the Al2O3 pores by impregnation and successive
thermal treatment of corresponding catalysts.
The XRD patterns of the various Co/␥-Al2O3 catalysts calcined
at 500 ◦C and reduced at 600 ◦C (Fig. 1) showed diffraction peaks
at 2ꢀ = 37.6◦, 45.8◦ and 66.8◦ from the ␥-Al2O3 (JCPDS 29-0063).
The calcined catalysts showed a characteristic reflection peak at
2ꢀ = 36.8◦ from Co3O4 phase (JCPDS 42-1467). The XRD patterns
show no new crystalline compounds formed between the cobalt
oxide and the ␥-Al2O3. The reduced Co/␥-Al2O3 catalysts showed
characteristics peaks of metallic Co (JCPDS 15-0806) and ␥-Al2O3.
The sizes of the metallic Co particles (Table 2) were determined
from the broadening of the diffraction peaks using Scherrer’s equa-
tion. The calcined Co3O4/␥-Al2O3 showed Co3O4 particles that
increased from 8 to 20 nm with increasing Co loading. High-
temperature reduction at 600 ◦C resulted in smaller metallic Co
particles, 8–15 nm. Reduced Co(11)/Al2O3 exhibited the smallest Co
particles of 8.1 nm and reduced Co(27)/Al2O3 contained the largest
Co particles of 14.8 nm. The Co species were more likely to sinter
to yield larger Co particles at increased cobalt loading. This result
appears to agree with the data on the degree of metal dispersion
(Table 2).
Fig. 1. XRD patterns of Co/␥-Al2O3 catalysts with various cobalt contents: (a) cal-
cined at 500 ◦C for 2 h, and (b) reduced at 600 ◦C for 3 h.
Co(27)/AlO3
2
Co(23)/AlO3
2
Co(17)/AlO3
2
Co(11)/AlO3
2
H2-TPR profiles of the various catalysts calcined at 500 ◦C
(Fig. 2) show two main broad reduction peaks at 300–500 ◦C
and 600–750 ◦C. The reduction peaks for the bulk Co3O4 were
assigned to the two-step reduction of Co3O4 to CoO to Co0 [5,28],
with the first being attributed to the reduction of bulk Co3O4
(Co3+ → Co2+ → Co0) and the higher temperature peak correspond-
ing to the reduction of cobalt aluminate (CoO Al2O3) formed on
the catalyst surface layer [29]. Alumina supported cobalt catalysts
have been reported to form cobalt aluminate that can hinder the
complete reduction of the cobalt species [30]. Therefore, the second
Co(4)/AlO3
2
200
300
400
500
600
700
800
Temperature (oC)
Fig. 2. H2-TPR profiles of Co/␥-Al2O3 catalysts calcined at 500 ◦C for 2 h.

316
J.H. Cho et al. / Applied Catalysis A: General 417–418 (2012) 313–319
peak can be assigned to the cobalt oxide layer contacting the alu-
mina surface. Increasing cobalt loading increased the intensity and
area of the first reduction peak and shifted it to lower temperatures,
indicating that the decrease of interactions between the Co3O4 and
the ␥-Al2O3 support [31]. The increasing intensity suggests the
more facile reduction of larger cobalt oxide particles within the
size range of 8.1–19.5 nm (XRD). Therefore, particle size and cobalt
loading affected the degree of reduction due to differences in their
interactions with alumina.
Measurement of H2 chemisorption on the reduced catalysts
allowed quantification of the number of cobalt metal sites, cobalt
dispersion and average metal particle size. The results of H2
chemisorption measurement were corrected by the degree of
reduction measured by O2 titration (Table 2). The cobalt parti-
cles, when corrected by considering the degree of reduction, were
smallest (ca. 12.5 nm) in Co(11)/Al2O3. Particle size increased from
12.5 to 21.5 nm with increasing cobalt loading. The cobalt particle
sizes were in good agreement with the results of the XRD analy-
sis, which showed the smallest Co particles of 8.1 nm exhibited by
Co(11)/␥-Al2O3 and the largest of 14.8 nm shown by Co(27)/Al2O3.
Comparison of the O2 titrations of cobalt oxide on alumina sup-
ports shows increasing reduction with increasing cobalt loading, in
agreement with the H2-TPR results that showed shifting to lower
temperatures of the first peaks corresponding to the reduction of
cobalt oxide. The low degree of reduction of the Co/␥-Al2O3 catalyst
was attributed to the small cobalt oxide particles that could eas-
ily form cobalt–aluminate during calcination or reduction [32,33].
The formation of cobalt–aluminate was difficult and impractical to
reduce at the temperatures used here and was therefore inactive
towards the reductive amination. The reduced metal−s1urface area of
cobalt was largest for Co(23)/Al2O3 at ca. 2.2 m2 gcatalyst and was
correlated with 2-propanol conversion (Table 2).
3.2. Reductive amination of 2-propanol
The conversions of 2-propanol and selectivities to MIPA, ace-
tone, DIPA, and DIPE in the reductive amination of 2-propanol on
Co/␥-Al2O3 catalysts were tested at cobalt loadings of 4–27 wt%
(Table 3). The 2-propanol conversions over these catalysts ranged
from 32.4 to 81.5%. The conversion increased with the increase
of cobalt loading, but after reaching 23 wt% Co the conversion
decreased. Although the conversion is not directly proportional
to the metallic surface area of Co/␥-Al2O3 catalyst the observed
variation in the conversion can be correlated to the difference
in the reduced metallic surface area of the catalysts. This was
likely due to variations of reducibility and cobalt metal disper-
sion accompanying the higher metallic surface areas [34], which
was shown by the H2-chemisorption results (Table 2). The sum of
the selectivities to MIPA and acetone as main products was almost
constant at ca. 97% under the tested operating conditions. Small
quantities of DIPA were formed by the consecutive reaction of
MIPA with 2-propanol. Also, DIPE was formed by the condensa-
tion/dehydration of 2-propanol. The selectivity pattern was similar
to those over other catalysts for reductive amination [35]. Various
by-products which could be formed by the condensation, decar-
bonylation, disproportionation and hydrogenolysis of 2-propanol
under the tested operating conditions were C1 C2 compounds,
DIPE and DIPA. However, the quantities formed here were negli-
gible during the reductive amination of 2-propanol.
The effects of hydrogen partial pressure were examined at
210 ◦C with a space velocity of 4.29 h−1 and R = [2-propanol]:[NH3]
molar ratio of 1:4 in the reductive amination of 2-propanol over
Co(23)/Al2O3 catalyst (Fig. 3). The total flow rate was 90 cm3 min−1
.
To maintain constant partial pressures of the other reactants, nitro-
gen was used as diluent. At H2 partial pressures below 18 kPa,

J.H. Cho et al. / Applied Catalysis A: General 417–418 (2012) 313–319
317
Table 3
Effect of cobalt content in the reductive amination of 2-propanol over Co/␥-Al2O3.a
Conversionb (%)
Selectivity (%)
MIPA
Catalyst
MIPA yield (%)
Acetone
DIPAc
DIPEd
Co(4)/Al2O3
Co(11)/Al2O3
Co(17)/Al2O3
Co(23)/Al2O3
Co(27)/Al2O3
32.4
76.4
78.1
81.5
80.3
76.1
72.5
71.2
71.7
71.7
21.8
25.0
25.9
25.9
25.6
–
2.1
2.3
2.3
2.3
2.5
24.7
55.4
55.6
58.4
57.6
0.2
0.4
0.2
0.2
a
Reaction conditions: T = 210 ◦C, WHSV = 4.29 h−1; feed compositions of 2-propanol/NH3/H2/N2 (mol ratio) = 1/4/6/22.8.
Conversion and selectivities to MIPA, acetone, DIPA, and DIPE obtained at 5 min on stream.
DIPA, diisopropylamine.
DIPE, disopropyl ether.
b
c
d
conversion increased rapidly with increasing H2 partial pressure;
ca. 79%, at 190 ◦C. Decreasing amine selectivity was observed at
high temperature due to strong enhancement of side reactions,
i.e. the dehydrogenation of corresponding alcohols and the dispro-
portion of reactant and product amines [5,34,37–39]. As discussed
above, acetone is the first product of 2-propanol dehydrogenation
as shown in Scheme 1. High performance for the reductive ami-
nation of acetone to MIPA would be possible at lower reaction
temperature than that of 2-propanol. The water vapor as another
compound produced during the dehydroamination of acetone was
above this, increasing H2 partial pressure resulted in a slight
increase of conversion. Decreasing H2 partial pressure progres-
sively deactivated the catalyst, possibly due to the formation
of cobalt nitride by strongly adsorbed nitrogen species on the
cobalt metal (3Co + NH3 → Co3N + 1.5H2) [5,36]. Excess hydrogen
efficiently hindered the phase transition of the catalyst to metal
nitride during the reaction and prevented the deactivation of the
catalyst. MIPA selectivity increased progressively with increasing
H2 partial pressure and selectivities to acetone and DIPE decreased.
This is consistent with the postulated reaction mechanism whereby
the imine intermediate must first be hydrogenated to form amine
before further reactions can occur [5].
The effects of ammonia partial pressure on the reductive ami-
nation of 2-propanol over Co(23)/␥-Al2O3 was investigated at
210 ◦C with a space velocity of 4.29 h−1 and R = [2-propanol]:[H2]
molar ratio of 1:8 (Fig. 4). 2-Propanol conversion was almost con-
stant with increasing ammonia partial pressure. MIPA selectivity
increased progressively with increasing ammonia partial pressure
and acetone selectivity decreased. Excess ammonia had a positive
effect on MIPA selectivity by the reductive amination of acetone,
limiting the production of DIPA, DIPE and acetone [37–39].
100
80
60
40
20
0
100
80
Conversion
60
40
Selectivity
MIPA
Acetone
DIPA
DIPE
20
0
Reaction temperatures of 170–250 ◦C were adopted at 2-
propanol WHSV of 4.29 h−1 and 2-propanol:NH3:H2 molar ratio
of 1:6:12 (Fig. 5). The total flow rate of 90 cm3 min−1 was main-
tained by controlling the nitrogen flow. The 2-propanol conversion
increased with increasing reaction temperature. The MIPA selec-
tivity decreased with the increase of the reaction temperature
as opposed to the conversion and selectivity to acetone. At high
temperature, the formation of acetone was favored by the dehy-
drogenation of 2-propanol. The maximum yield of MIPA was
6
12
18
24
30
36
Ammonia partial pressure (kPa)
Fig. 4. Influence of ammonia partial pressure on the reductive amination of 2-
propanol over Co(23)/Al2O3. Reaction conditions: T = 210 ◦C, WHSV (h−1) = 4.29; feed
composition of 2-propanol/H2 (mol%) = 1/8.
Conversion
Selectivity: MIPA,
MIPA yield
Acetone ,
DIPA ,
DIPE
100
90
100
80
100
80
100
80
Conversion
60
80
Selectivity
60
40
MIPA
60
Acetone
DIPA
70
20
0
DIPE
20
0
40
60
160
180
200
220
240
260
6
12
18
24
30
36
42
48
Temperature (oC)
Hydrogen partial pressure (kPa)
Fig. 3. Effects of hydrogen partial pressure on the reductive amination of 2-propanol
over Co(23)/Al2O3. Reaction conditions: T = 210 ◦C, WHSV (h−1) = 4.29; feed compo-
sitions of 2-propanol/NH3 (mol%) = 1/4.
Fig. 5. Effects of reaction temperature on the reductive amination of 2-propanol
over Co(23)/Al2O3. Reaction conditions: WHSV (h−1) = 4.29; feed composition of 2-
propanol/NH3/H2/N2 (mol%) = 1/6/12/14.8.

318
J.H. Cho et al. / Applied Catalysis A: General 417–418 (2012) 313–319
Conversion
Selectivity: MIPA,
Acetone,
C -C
1 2
100
80
60
40
20
0
100
80
60
40
20
0
H2
+
N2
NH3
+
N2
NH3
+
H2
+
N2
NH3
+
H2
+
N2
NH3
+
H2
+
N2
0
2
4
6
8
10
12
14
16
Time on stream (h)
Fig. 6. Evolution of 2-propanol conversion and MIPA selectivity as a function of
time on stream in the presence or absence of hydrogen for the reductive amination
of 2-propanol over Co(23)/Al2O3. During the reaction, the flow of the N2 + H2 mix-
ture was changed to pure N2 and then returned to its initial composition. Reaction
conditions: T = 190 ◦C, WHSV = 4.29 h−1; feed composition of 2-propanol/NH3/H2/N2
(mol%) = 1/6/12/14.8, 1/6/0/26.8.
not examined in this study. However, only slight influence of water
vapor in yield of amine product has been reported on the catalytic
amination of 1-methoxy-2-propanol over silica supported catalyst
[37].
To observe the effect of deactivation, the evolution of 2-propanol
conversion and selectivities to MIPA, acetone and C1 C2 prod-
ucts were investigated over Co(23)/Al2O3 catalyst with respect to
time on stream in the presence or absence of NH3 or H2 (Fig. 6).
During the reaction, the flow of NH3 or H2 was stopped and
returned to the initial flow composition. When amination was per-
formed with initial feed compositions of 2-propanol/NH3/H2/N2
(mol ratio) = 1/6/12/14.8, 2-propanol conversion and selectivities to
MIPA and acetone were 90%, 88%, and 11%, respectively. Removing
the H2 flow from the reactant stream rapidly deactivated the cata-
lyst and no formation of MIPA was observed within 1 h on stream.
However, the initial activity and selectivities to MIPA and acetone
were completely recovered upon re-exposure to feed containing
hydrogen. Also, no deactivation of the catalyst was shown. Verhaak
et al. [40] observed the rapid deactivation of supported nickel cata-
lyst in the disproportionation of n-propylamine when hydrogen in
the feed was replaced by nitrogen. In this study similar result was
observed. The conversion and selectivity were totally regenerated
when the gas feed containing hydrogen was replaced once again.
They ascribed this deactivation to the formation of metal nitride
during the reaction in nitrogen. The formation of metal nitride
can deactivate metallic catalysts used for reductive amination [41].
Feed containing excess hydrogen could efficiently hinder the phase
transition of the metallic catalyst to nitride during the reaction, pre-
venting deactivation. A reactant flow without NH3 resulted in the
synthesis of acetone as the main product of the dehydrogenation
of 2-propanol in the presence of H2; propylene was detected as a
minor product and negligible amounts of methane and ethylene
were also observed from the hydrogenolysis of C3 compounds. The
only alumina support gave products of propylene and DIPE, which
formed on weakly acidic sites [42,43]. After adding NH3 to the reac-
tants, the initial conversion of 2-propanol and selectivities to MIPA
and acetone over Co/Al2O3 catalyst were completely recovered.
To assess the adsorbed species during the amination reaction
and the origin of deactivation, NH3-TPD and simultaneous NH3-
TPD and H2-TPR experiments were conducted using a quadruple
mass spectrometer. Argon and a mixture of hydrogen and argon
Fig. 7. Evolution of NH3 and H2 during TPD/TPR for Co(23)/Al2O3 catalyst under
flowing (a) 5% H2/Ar, (b) Ar, and (c) 5% H2/Ar. The samples reduced were pretreated
under flowing ((a) and (b)) 20% NH3/N2 or (c) 4% NH3/16% H2/N2 at 190 ◦C for 12 h.
were used. Prior to each experiment, samples were reduced under
a H2 flow at 600 ◦C for 3 h with heating at 10 ◦C min−1. Fig. 7 shows
the evolution of NH3 and H2 during TPD/TPR using Co(23)/Al2O3
catalyst under H2/Ar and Ar flows. The reduced samples were pre-
treated under flowing 20% NH3/N2 or 4% NH3/16% H2/N2 at 190 ◦C
for 12 h. Pure Ar was then passed through the reactor for 15 min.
NH3 desorption and H2 consumption were recorded under flowing
5% H2/Ar for the sample pretreated with 20% NH3/N2 (Fig. 7(a)). The
NH3-TPD profiles of the sample pretreated with NH3/Ar show two
main peaks at 110 and 170 ◦C and broad peaks up to 400 ◦C. The first
peak originated from the desorption of weakly adsorbed ammonia
on the surface of the metal phase and alumina support; the second
one corresponds to the desorption of strongly adsorbed ammonia

J.H. Cho et al. / Applied Catalysis A: General 417–418 (2012) 313–319
319
100
80
60
40
20
0
100
80
60
40
20
0
Acknowledgement
This work was financially supported by a grant from the Indus-
trial Source Technology Development Programs (2008-10031908)
of the Ministry of Knowledge Economy (MKE) of Korea.
Conversion
MIPA yield
References
Selectivity
MIPA
[1] K.S. Hayes, Appl. Catal. A: Gen. 221 (2001) 187–195.
[2] R. Vultier, A. Baiker, A. Wokaun, Appl. Catal. 30 (1987) 167–176.
[3] A. Fischer, T. Mallat, A. Baiker, Angew. Chem. Int. Ed. 38 (1999) 351–354.
[4] M. Pérez-Mendoza, M. Domingo-García, F.J. López-Garzón, Appl. Catal. A: Gen.
224 (2002) 239–253.
[5] G. Sewell, C. O’Connor, E. Steen, Appl. Catal. A: Gen. 125 (1995) 99–112.
[6] J. Becker, J.P.M. Niederer, M. Keller, W.F. Hölderich, Appl. Catal. A: Gen. 197
(2000) 229–238.
Acetone
DIPA
DIPE
0
20
40
60
80
100
[7] G. Boettger, H., Corr, H., Hoffmann, H., Toussaint, S. Winderl, US Patent,
4,014,933 (1977).
Time on stream (%)
[8] H.Y. Jeon, C.-H. Shin, H.J. Jung, S.B. Hong, Appl. Catal. A: Gen. 305 (2006) 70–78.
[9] K.V.R. Chary, K.K. Seel, D. Naresh, P. Ramakanth, Catal. Commun. 9 (2008) 75–81.
[10] G. Sewell, C. O’Connor, E. Steen, J. Catal. 167 (1997) 513–521.
[11] G.A. Vedage, L.A. Emig, H.X. Li, J.N. Armor, US Patent 5,917,092 (1999).
[12] G.A. Vedage, K.S. Hayes, M., Leeaphon, J.N. Armor, US Patent 5,932,769 (1999).
[13] A. Fischer, T. Mallat, A. Baiker, Catal. Today 37 (1997) 167–189.
[14] C. Dume, W.F. Holderich, Appl. Catal. A: Gen. 183 (1999) 167–176.
[15] M.E. Dominea, M.C.H. Sotoa, Y. Perez, Catal. Today 159 (2011) 2–11.
[16] H. Kimura, K. Ishikawa, K. Nishino, S. Nomura, Appl. Catal. A: Gen. 286 (2005)
120–127.
[17] C.U. Kim, Y.S. Kim, H.J. Chae, K.E. Jeong, S.Y. Jeong, K.W. Jun, K.Y. Lee, Korean J.
Chem. Eng. 27 (2010) 777–784.
[18] S. Zafeiratos, T. Dintzer, D. Teschner, R. Blume, M. Havecker, A. Knop-Gericke,
R. Schlogl, J. Catal. 269 (2010) 309–317.
[19] D.S. Kim, Y.H. Kim, J.E. Yie, E.D. Park, Korean J. Chem. Eng. 27 (2010) 822–827.
[20] S.H. Kang, K.J. Woo, J.W. Bae, K.W. Jun, Y. Kang, Korean J. Chem. Eng. 26 (2009)
1533–1538.
[21] P. Khemthong, W. Klysubun, S. Prayoonpokarach, F. Roessner, J. Wittayakun, J.
Ind. Eng. Chem. 16 (2010) 531–538.
[22] D.A. Gardner, R.T. Clark, US Patent 4,255,357 (1981).
[23] I.D. Dobson, W.A. Lidy, P.S. Williams, US Patent 4,912,260 (1990).
[24] B. Tijsebaert, B. Yilmaz, U. Muller, H. Gies, W. Zhang, X. Bao, F.S. Xiao, T. Tatsumi,
D.D. Vos, J. Catal. 278 (2011) 246–252.
[25] D.R. Corbin, S. Schwarz, G.C. Sonnichsen, Catal. Today 37 (1997) 71–102.
[26] J.T. Richardson, W.C. Lu, J. Catal. 42 (1976) 275–281.
[27] J.W. Bae, S.M. Kim, S.H. Kang, K.V.R. Chary, Y.J. Lee, H.J. Kim, K.W. Jun, J. Mol.
Catal. A 311 (2009) 7–16.
[28] M.M. Yung, E.M. Holmgreen, U. Ozkan, J. Catal. 247 (2007) 356–367.
[29] G. Jacobs, Y. Ji, B.H. Davis, D. Cronauer, A.J. Kropf, C.L. Marshall, Appl. Catal. A:
Gen. 333 (2007) 177–191.
[30] A. Tavasoli, R.M.M. Abbaslou, M. Trepanier, A.K. Dalai, Appl. Catal. A: Gen. 345
(2008) 134–142.
Fig. 8. Long-term stability of Co(23)/Al2O3 for reductive amination. Reaction con-
ditions: T = 190 ◦C, WHSV = 4.29 h−1; feed composition of 2-propanol/NH3/H2/N2
(mol%) = 1/6/12/14.8.
on the cobalt, which may be due to the formation of surface cobalt
nitride [44]. The hydrogen consumption peak at 160 ◦C indicates the
removal of strongly adsorbed nitrogen-containing surface species
in the presence of hydrogen. To investigate the role of hydrogen
in the amination reaction, NH3-TPD was performed under flow-
ing Ar in the absence of hydrogen (Fig. 7(b)). When only Ar was
used during TPD, the second peak was smaller, indicating the facile
removal of strongly adsorbed nitrogen-containing surface species
in the presence of hydrogen. H2 evolution was not detected during
NH3-TPD. The evolution of NH3 or H2 was not observed over the
sample pretreated with 4% NH3/16% H2/N2, indicating that hydro-
gen was linked to the prevention of catalyst deactivation due to
surface nitride formation [5,40,41,44,45].
The long term stability of the Co(23)/Al2O3 catalyst during
the reductive amination of 2-propanol was assessed (Fig. 8). The
Co(23)/␥-Al2O3 catalyst which showed the best catalytic activity
among the catalysts studied here exhibited constant conversion
and selectivities to MIPA and acetone up to 100 h on stream. The
Co/␥-Al2O3 catalyst showed good catalytic activity and stability,
making its use economically viable.
[31] Y. Ji, Z. Zhao, A. Duan, G. Jiang, J. Liu, J. Phys. Chem. C 113 (2009) 7186–7199.
[32] L. Zhang, L. Dong, W. Yu, L. Liu, Y. Deng, B. Liu, H. Wana, F. Gao, K. Sun, L. Dong,
J. Colloid Interface Sci. 355 (2011) 464–471.
[33] J.W. Bae, S.M. Kim, Y. Jo Lee, M.J. Lee, K.W. Jun, Catal. Commun. 10 (2009)
1358–1362.
[34] A.K. Rausch, E. Steen, F. Roessner, J. Catal. 253 (2008) 111–118.
[35] V. Zamlynny, L. Kubelkova, E. Baburek, K. Jiratova, J. Novakova, Appl. Catal. A:
Gen. 169 (1998) 119–125.
4. Conclusions
The Co/␥-Al2O3 catalysts with 4–27 wt% cobalt loadings
were prepared by incipient-wetness impregnation. Particle sizes
increased from 12.5 to 21.5 nm with increasing cobalt content.
Larger cobalt oxide particles could be more easily reduced due to
weaker interactions with the support. The highly reduced cobalt
metal surface area could be correlated with the enhancement
of 2-propanol conversion. Excess ammonia enhanced 2-propanol
conversion and MIPA selectivity. Excess hydrogen hindered the
phase transition of the catalyst to metal nitride during reaction and
the deactivation of the catalyst. The Co/␥-Al2O3 catalyst showed
good catalytic performance and its catalytic activities were stable in
the presence of excess hydrogen and ammonia under atmospheric
pressure.
[36] A. Baiker, I. Monti, Y. Songfan, J. Catal. 88 (1984) 81–88.
[37] V.A. Bassili, A. Baiker, Appl. Catal. 65 (1990) 293–308.
[38] S.R. Kirumakki, M. Papadaki, K.V.R. Chary, N. Nagarajua, J. Mol. Catal. A 321
(2010) 15–21.
[39] A. Fischer, M. Maciejewski, T. Bu˝ rgi, T. Mallat, A. Baiker, J. Catal. 183 (1999)
373–383.
[40] M.J.F.M. Verhaak, A.J. van Dillen, J.W. Geus, Appl. Catal. A: Gen. 109 (1994)
263–275.
[41] M.J.F.M. Verhaak, A.J. van Dillen, J.W. Geus, J. Catal. 143 (1993) 187–200.
[42] C.R. Narayanan, S. Srinivasan, A.K. Datye, R. Gorte, A. Biaglow, J. Catal. 138 (1992)
659–674.
[43] S. Srinivasan, C.R. Narayanan, A. Biaglow, R. Gorte, A.K. Datye, Appl. Catal. A:
Gen. 132 (1995) 271–287.
[44] A. Baiker, J. Kijenski, Catal. Rev. Sci. Eng. 27 (1985) 653–697.
[45] J. Kritzenberger, E. Jobson, A. Wokaun, A. Baik un, A. Baiker, Catal. Lett. 5 (1990)
73–80.