Reductive amination of isopropanol to monoisopropylamine over Ni-Fe/Γ-Al2O3 catalysts: Synergetic effect of Ni-Fe alloy formation

Article abstract of DOI:10.1016/j.apcata.2017.05.003

We herein reported the syntheses of 17?wt% Ni-xFe/γ-Al₂O₃ catalysts containing different Fe/Ni molar ratios (x?=?0–0.7) via an incipient wetness impregnation method, followed by subsequent investigation of their catalytic properties in the reductive amination of isopropanol (IPA). We attempted to reveal the synergetic effect of bimetallic Ni-Fe on γ-Al₂O₃ support through characterization of the catalysts by N₂-sorption, X-ray diffraction, H₂-temperature-programmed reduction, and X-ray photoelectron spectroscopy. We found that conversion and selectivity to monoisopropylamine (MIPA) through the reductive amination of IPA was influenced by the reducibility of the catalysts. In addition, the surface Ni⁰ content increased upon the addition of Fe species, with a maximum Ni⁰ content being reached at a Fe/Ni molar ratio of 0.3. This molar ratio also yielded the highest conversion and MIPA selectivity, and no deactivation was observed for 100?h on stream in the presence of hydrogen gas. In addition, reaction parameters, such as the partial pressures of H₂ and NH₃, the space velocity, and the reaction temperature were also examined to optimize the experimental conditions.

Full text of DOI:10.1016/j.apcata.2017.05.003

Accepted Manuscript
Title: Reductive Amination of iso-Propanol to
Monoisopropylamine over Ni-Fe/␥-Al O Catalysts:
2
3
Synergetic Effect of Ni-Fe Alloy Formation
Authors: Eunpyo Hong, Seonu Bang, Jun Hee Cho, Chae-Ho
Shin
PII:
S0926-860X(17)30193-X
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2017.05.003
APCATA 16225
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
16-3-2017
1-5-2017
5-5-2017
Please cite this article as: Eunpyo Hong, Seonu Bang, Jun Hee Cho, Chae-Ho
Shin, Reductive Amination of iso-Propanol to Monoisopropylamine over Ni-Fe/␥-
Al2O3 Catalysts: Synergetic Effect of Ni-Fe Alloy Formation, Applied Catalysis A,
Generalhttp://dx.doi.org/10.1016/j.apcata.2017.05.003
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Highlights
●
●
●
●
Fe species improves the reducibility of the Ni component
Ni-Fe alloy is formed during reduction process after calcination
0
Surface Ni content critically influences on reductive amination of iso-propanol
Ni-0.3Fe/Al O catalyst shows the best catalytic performance
2
3

Reductive Amination of iso-Propanol to Monoisopropylamine over Ni-Fe/γ-
Al O Catalysts: Synergetic Effect of Ni-Fe Alloy Formation
2
3
†
Eunpyo Hong, Seonu Bang, Jun Hee Cho, and Chae-Ho Shin
Department of Chemical Engineering, Chungbuk National University, Chungbuk 28644, Republic of
Korea
†
To whom correspondence should be addressed.
Phone: +82-43-261-2376. Fax: +82-43-269-2370.
E-mail: chshin@chungbuk.ac.kr

ABSTRACT
We herein reported the syntheses of 17 wt% Ni-xFe/γ-Al O catalysts containing different
2
3
Fe/Ni molar ratios (x = 0 – 0.7) via an incipient wetness impregnation method, followed by
subsequent investigation of their catalytic properties in the reductive amination of iso-
propanol (IPA). We attempted to reveal the synergetic effect of bimetallic Ni-Fe on γ-Al O
2
3
support through characterization of the catalysts by N -sorption, X-ray diffraction, H -
2
2
temperature-programmed reduction, and X-ray photoelectron spectroscopy. We found that
conversion and selectivity to monoisopropylamine (MIPA) through the reductive amination
0
of IPA was influenced by the reducibility of the catalysts. In addition, the surface Ni content
0
increased upon the addition of Fe species, with a maximum Ni content being reached at a
Fe/Ni molar ratio of 0.3. This molar ratio also yielded the highest conversion and MIPA
selectivity, and no deactivation was observed for 100 h on stream in the presence of hydrogen
gas. In addition, reaction parameters, such as the partial pressures of H and NH , the space
2
3
velocity, and the reaction temperature were also examined to optimize the experimental
conditions.
Keywords: reductive amination, bimetallic Ni-Fe, iso-propanol, monoisopropylamine,
bifunctional catalyst

1
. Introduction
The reductive amination of aliphatic alcohols or carbonyl compounds is one of the most
effective ways to produce alkyl amine groups. Alkyl amines are important chemical
compounds for use as valuable intermediates in a variety of applications including herbicides,
insecticides, pharmaceutical chemicals, corrosion inhibitors, water treatment chemicals, and
rubber chemicals [1-5]. Among the amine family, monoisopropylamine (MIPA) is employed
as versatile intermediate for a variety of applications. The main applications can be found in
the field of agricultural chemicals, and producing the surfactants and active pharmaceutical
ingredients.
To date, MIPA has been synthesized by the reductive amination of iso-propanol (IPA) or
acetone [1], where the reductive amination of IPA to MIPA over supported metal catalysts
involves dehydrogenation, condensation, and hydrogenation steps [3, 6]. Although the use of
catalysts containing precious metals, such as Pt and Pd, has been reported for reductive
amination processes [7], Ni- and Co-based catalysts [3, 6, 8-11] are generally employed for
the synthesis of primary amines, due to their good catalytic activity and low costs.
To further improve the catalytic activity of Ni-based catalysts, various optimization studies
have been carried out involving the addition of promoters. For example, Fe, a group VIII
metal, has been widely considered as an active material or a promoter in various catalytic
reactions, such as the Fischer-Tropsch synthesis [12], CO hydrogenation [13],
thioetherification [14], and the water-gas shift reaction [15] because of its high activity, good
redox properties, and low cost. Especially, bimetallic Ni-Fe/γ-Al O catalysts have exhibited
2
3
enhanced catalytic performances compared to monometallic Ni/γ-Al O due to weak
2
3
interactions between the metal and the support in the bimetallic Ni-Fe systems [16-18].
In our previous work, we examined the effects of changing the nickel loading and the

phase of the Al O support on the reductive amination of IPA over a Ni/Al O catalyst, and
2
3
2
3
proposed the use of 17 wt% Ni/γ-Al O as an optimal catalyst [4, 8]. In this study, a series of
2
3
Ni-Fe/γ-Al O catalysts containing various Fe loadings were prepared using a co-
2
3
impregnation method. To the best of our knowledge, no study has yet been reported the use of
a γ-Al O supported bimetallic Ni-Fe catalyst for the reductive amination of IPA. We
2
3
therefore aim to reveal the synergetic effect between Ni and Fe metals on the γ-Al O support
2
3
in the reductive amination reaction. For this purpose, characterizations of the synthesized Ni-
Fe/γ-Al O catalysts were performed by N -sorption, X-ray diffraction (XRD), H -
2
3
2
2
temperature-programmed reduction (H -TPR), and X-ray photoelectron spectroscopy (XPS)
2
as well as reductive amination reaction.
2
. Experimental procedure
2
.1. Catalyst Preparation
The Ni/γ-Al O and Ni-Fe/γ-Al O catalysts were synthesized by an incipient wetness
2
3
2
3
impregnation method using a precursor solution of nickel (Ni(NO ) ·6H O, Kanto) and iron
3
2
2
2
(
Fe(NO ) ·9H O, Kanto) and a commercial γ-Al O (Procatalyse, 194 m /g) support. The
3 3 2 2 3
impregnated catalysts were dried at 100 °C overnight, and then calcined at 500 °C for 2 h
under a constant flow of air. In the series of Ni-Fe/γ-Al O catalysts, the Ni loading was fixed
2
3
at 17 wt% (Ni/(Ni+Al O )), and the molar ratio of Fe/Ni was varied between 0 and 0.7. The
2
3
final catalysts will thus be referred to as Ni-xFe/Al O catalysts, where x denotes the molar
2
3
ratio of Fe/Ni, and the monometallic Ni loaded on γ-Al O will be referred to as Ni/Al O .
2
3
2
3
2
.2. Characterization
N -sorption isotherms were obtained on a Micromeritics ASAP2020. Prior to measurement
2

of the N -sorption isotherms, the samples were degassed at 250 °C for 6 h under vacuum. The
2
XRD patterns of all samples were recorded on a Rigaku D/MAX 2500H diffractometer using
Cu Kα radiation operating at 40 kV and 40 mA with a scan rate of 4 °/min. In addition,
samples reduced at 600 °C for 3 h under a stream of H (50 mL/min), followed by surface
2
passivation with 1.0% O /He for 0.5 h at room temperature (RT), were also characterized by
2
XRD. The H -TPR experiments were carried out using a Micromeritics AutoChem II 2920
2
instrument to determine the reducibility of the Ni-xFe/Al O catalysts. Prior to carrying out
2
3
the measurements, the sample (0.1 g) was pretreated under flowing He (50 mL/min) at
00 °C for 1 h to remove any adsorbed water and other contaminants. After cooling the
4
sample to 30 °C under a flow of He, it was exposed to 5% H /Ar mixture gas (50 mL/min),
2
and heated to 800 °C at a rate of 10 °C/min. A quadruple mass spectrometer (QMS, Balzers
QMS 200) was used to determine the amount of H consumption with a fragment of H (m/z
2
2
=
2). Furthermore, the surface oxidation states of the nickel species were identified by XPS
on a ESCALAB 210 Electron Spectrometer (Al Kα radiation; hν = 1486.6 eV). The XPS data
were calibrated using the binding energy of C1s (284.6 eV) as a standard. To determine the
peak areas and positions, the deconvolution of the Ni 2p3/2 spectra was performed using a
Gaussian–Lorentzian curve-fitting method after background subtraction by the Shirley
method.
2
.3. Catalytic Activity Tests
All catalytic experiments were performed at atmospheric pressure under a continuous flow.
Each catalyst sample (0.1 g) was loaded in a quartz reactor using quartz wool. Prior to each
test, the samples were reduced at 600 °C for 3 h under a flow of purified H (50 mL/min).
2
After successful activation step, the flow of H was maintained until the desired temperature
2
was reached. The standard reaction conditions for experiments with varying Fe content (Ni-

-1
-1
xFe/Al O catalysts) were as follows: T = 150 °C, space velocity (SV) = 500 mL∙min ∙g ,
2
3
total flow rate (at standard temperature and pressure, STP) = 50 mL/min, a molar feed ratio of
IPA/NH /H = 1/8/6 (with N balance), and a fixed IPA partial pressure of 3 kPa, where SV
3
2
2
was defined as total flow rate (mL/min) per unit mass (g) of catalyst. To optimize the reaction
conditions for reductive amination of IPA, variation in the H partial pressure (12–30 kPa),
2
-1
-1
NH partial pressure (6–24 kPa), SV (300–100 mL∙min ∙g ), and reaction temperature (130–
3
1
9
60 °C) were examined. The reaction products were analyzed on-line using a Chrompack CP-
001 gas chromatograph equipped with a CP-Volamine capillary column (60 m × 0.32 mm)
and a flame ionization detector.
The conversion of IPA and the selectivity were defined as follows:
퐹IPA 푟푒푎푐푡푒푑
퐶표푛푣푒푟푠푖표푛 (%) =
× 100 %
퐹IPA 푓푒푑
퐹ꢀ퐶ꢀ
푖 퐶표푚푝표푛푒푛푡 푠푒푙푒푐푡푖푣푖푡푦 (%) =
× 100 %
훴퐹 퐶
ꢀ
ꢀ
where FIPA and 퐹 are the molar flow rates of IPA and product i (MIPA, diisopropylamine
ꢀ
(
DIPA), and acetone), respectively, and C is the carbon number of each product, i.
i
3
. Results and discussion
3
.1. Characterization
Fig. 1(a) shows the N -sorption isotherms of the γ-Al O support and the Ni-xFe/Al O
3
2
2
3
2
catalysts calcined at 500 °C for 2 h under a flow of air. According to IUPAC nomenclature,
all isotherms could be classified as a type IV isotherm with a type H2 hysteresis loop,
indicating that calcined catalysts exhibited mesoporous structures. In addition, the hysteresis
of the γ-Al O support and the Ni/Al O catalysts were narrower than those of the bimetallic
2
3
2
3

Ni-Fe/Al O catalysts. Hence, a wider range of pore size distributions was developed upon
2
3
increasing the Fe content, as indicated in Fig. 1(b), where all samples including γ-Al O
2
3
support exhibit bimodal pore size distributions (13.5 nm for large pores and 7–10 nm for
smaller pores). In addition, with increasing Fe contents, the large pore content decreased due
to partial blockage of the γ-Al O support. Therefore, a decrease in the specific surface area
2
3
was observed upon increasing the Fe/Ni molar ratio to 0.3. However, above this point, the
specific surface area began to increase once again, likely due to the creation of inter- or intra-
pores by the addition of Ni and Fe species. Similar results could be also confirmed in the
literature [19]. Therefore, the specific surface areas of the Ni-xFe/Al O catalysts showed the
2
3
reversed volcano shaped curve as a function of the Fe/Ni molar ratio. However, all Ni-
xFe/Al O catalysts displayed lower specific surface areas than the γ-Al O support (194
2
3
2
3
2
m /g). The specific surface areas, total pore volumes, and average pore diameters of the Ni-
xFe/Al O catalysts are summarized in Table 1.
2
3
Fig. 2(a) shows the XRD patterns of Ni-xFe/Al O catalysts calcined at 500 °C for 2 h. The
2
3
diffraction peaks at 2θ = 45.8 and 66.7° corresponded with that of γ-Al O (JCPDS No. 29-
2
3
0
063), and peaks attributed to the Fe and Ni oxides, i.e., Fe O (JCPDS No. 25-1402) and
2 3
NiO (JCPDS No. 47-1049), were also detected. However, no distinguishable peak was
observed corresponding to the crystalline Ni-Fe alloy. As previously reported, the supported
Ni-Fe and unsupported Ni-Fe catalysts exhibited different crystalline structures, mixed oxide
phases (e.g., Ni(Fe)O , and NiFe O ) was detected after calcination under air flow in the
X
2
4
unsupported catalysts [18], and individual oxides were usually observed in the supported
catalysts due to the dispersion of the Ni and Fe precursors over the relatively large surface of
the support [14]. As such, the calcined samples were consecutively reduced using H gas, and
2
XRD patterns of the reduced catalysts were recorded (see Fig. 2(b)). In this case, a peak
corresponding to metallic Ni (JCPDS No. 04-0850) was newly detected, while peaks assigned

to NiO and Fe O disappeared. However, the complete elimination of the NiO and Fe O
3
2
3
2
peaks could not be confirmed by XRD, as these peaks partially overlapped with those of
metallic Ni, γ-Al O , and the others. In addition, the formation of Ni-Fe (JCPDS No. 03-1175)
2
3
and Ni rich Ni-Fe (JCPDS No. 65-3244) alloys was also confirmed following reduction.
These results indicated that the co-impregnated Ni and Fe species existed mainly on the
surface of the γ-Al O support as individual oxides following calcination, while Ni-Fe alloys
2
3
were formed during the reduction process. In addition, we carried out the TEM-EDS analysis
to obtain the specific morphology of Ni-Fe/Al O catalysts. However, no clear evidence for
2
3
Ni-Fe alloy formation was found from the TEM images. This is because particles were highly
agglomerated each other, and Ni and Fe particles were widely dispersed. However, it was
confirmed by EDS analysis that Ni and Fe particles were evenly distributed over Ni-Fe/Al O
2
3
catalysts (see the supplementary material, Fig. S1).
In order to analyze the reducibility and metal-support interactions of the samples, H -TPR
2
experiments were carried out using the Ni-xFe/Al O catalysts calcined at 500 °C for 2 h, as
2
3
shown in Fig. 3(a). In the H -TPR profile of the Ni/Al O catalyst, three clear reduction peaks
2
2
3
(i.e., α, β, and γ) with board distributions were observed at 350–500, 450–700, and 700–
8
50 °C, respectively [4, 8, 14]. The first peak (α, at ca. 430 °C) was due to the reduction of
bulk NiO or NiO weakly interacting with the support, while the second peak (β, at ca. 600 °C)
was attributed to the reduction of NiO that strongly interacted with the support, and the third
peak (γ, at above 700 °C) was assigned to the reduction of nickel aluminate. Due to the high
reduction temperature of nickel aluminate, it has been reported to hinder the complete
reduction of the nickel species [4, 8, 10, 14]. Furthermore, the observation of a peak
corresponding to nickel aluminate was only observed at low Fe/Ni molar ratios (i.e., ≤0.1).
Thus, it indicated that an increase in the Fe content successively restrain the formation of
nickel aluminate.

Upon the addition of Fe to the Ni/Al O catalysts, new signals (i.e., δ-1 and δ-2) in the H -
2
3
2
TPR profile began to appear at 350–370 and 390–410 °C, respectively. Thus, these peaks
were attributed to reduction of the Fe species, i.e., Fe O → Fe O → FeO. In addition, the
2
3
3
4
Fe/Al O catalyst showed broad reduction peak between 350 and 450 °C (see the
2
3
supplementary material, Fig. S2), which corresponds to previous literature observations.[14,
9, 20] For example, Liu et al. [20] carried out H -TPR measurements using an Fe O
1
2
2
3
catalyst, and observed that the reduction of Fe O to FeO occurred below 500 °C, but
2
3
0
complete reduction to Fe took place above 700 °C. Thus, it was a clear evidence that the
peaks at 350–410 °C corresponded to reduction of the Fe species. Indeed, as shown in Table 2,
the areas of the δ-1 and δ-2 peaks generally increased upon increasing the Fe content up to
Fe/Ni molar ratio of 0.5.
The positions and areas of all observed reduction peaks are also summarized in Table 2.
For the Ni-xFe/Al O catalysts, the positions of the δ-1, δ-2, α, and β peaks varied as a
2
3
function of the Fe/Ni molar ratio. Amongst the samples, Ni-0.3Fe/Al O exhibited the lowest
2
3
reduction temperature. A peak shift to lower temperatures in H -TPR profiles means the
2
enhanced reducibility through moderation of the interactions between the metal and the
support [11, 14, 21], and thus Ni-0.3Fe/Al O catalyst had the highest reducibility amongst
2
3
the samples. As previous reported, the quantity of metallic Ni species present in the catalyst
has a significant influence on the catalytic activity of the reductive amination reaction [4, 8].
This is because unreduced NiO cannot participate in the dehydrogenation-hydrogenation
process that takes place during reductive amination [22]. Therefore, the reduction behavior of
the NiO related peaks (i.e., α and β) was particularly important for determining the optimal
Fe/Ni composition in the bimetallic Ni-Fe system. Although the α peak temperature
corresponding to the Ni-0.1Fe/Al O was slightly higher than that of Ni/Al O , the NiO
2
3
2
3
species corresponding to the α peak can be completely reduced under the experimental

conditions employed herein (i.e., reduction at 600 °C for 3 h). Furthermore, the species
responsible for the NiO β signal can also be partially reduced at 600 °C. It was therefore
expected that the areas of the α and β reduction peaks were closely related to the surface
0
Ni /NiO composition following the reduction process. As such, Fig. 3(b) shows the areas of
the α and β peaks as a function of the Fe/Ni molar ratio in the Ni-xFe/Al O catalysts. The
2
3
sums of the α and β peak areas produced volcano shaped curves, with the highest areas being
observed for the Ni-0.3Fe/Al O catalyst, indicating that the reduction of NiO was promoted
2
3
by the presence of Fe species at this optimal composition. Furthermore, the H -TPR
2
experiments also provided valuable and quantitative information regarding the chemical state
of the Fe species. For example, in the Ni-0.1Fe/Al O catalyst, the Fe/Ni molar ratio was 0.1,
2
3
but the reduction quantities of (δ-1 + δ-2)/(α + β) were significantly smaller than 0.1.
Therefore, it indicates that the majority of Fe presents in the Ni-xFe/Al O catalysts as the
2
3
oxide form after reduction process, and it was further discussed in the context of the XPS
measurements.
Fig. 4(a) shows the Ni 2p XPS spectra of the various Ni-xFe/Al O catalysts. XPS was
2
3
conducted to examine the nature of the surface nickel species following reduction, and so all
samples were reduced at 600 °C for 3 h after calcination at 500 °C for 2 h. In the Ni 2p XPS
spectra, characteristic peaks with shakeup satellites were detected, which corresponded to Ni
0
2+
2
8
p3/2 and Ni 2p . For the Ni/Al O catalyst, the Ni and Ni (NiO) peaks were observed at
1/2 2 3
2
+
52.5 and 855.0 eV, respectively. In addition, the Ni peak for the NiAl O spinel structure
2
4
was confirmed at ca. 857.2 eV, and was only detected in the Ni/Al O and Ni-0.1Fe/Al O
3
2
3
2
catalysts. This correlates to the H -TPR results provided in Fig. 3. Furthermore, in the Fe
2
0
2+
promoted catalysts, the Ni and Ni peaks were shifted to slightly higher binding energies,
indicating that the addition of Fe species enhanced the electron density of the Ni species [14].
0
As shown in Table 3, the metallic Ni content reached a maximum at Ni-0.3Fe/Al O , but
2
3

decreased again at Fe/Ni molar ratios beyond 0.3. This tendency also correlated with the H2-
TPR results presented in Fig. 3, and these results clearly demonstrate that the reducibility of
the Ni-xFe/Al O catalysts increased with a promotional effect of Fe species up to an Fe/Ni
2
3
molar ratio of 0.3.
The Fe 2p spectra of the Ni-xFe/Al O catalysts are presented in Fig. 4(b), where two
2
3
major peaks with binding energies at ca. 710 and 723 eV were observed, corresponding to Fe
p3/2 and Fe 2p , respectively. The spectra in the Fe 2p region showed a relatively low
2
1/2
intensity due to the low content and sensitivity of the Fe species. In addition, although the Fe
0
2
p3/2 spectrum could be deconvoluted into three characteristic peaks corresponding to the Fe ,
2
+
3+
Fe , and Fe species [14, 23, 24], no relationship was observed between the Fe intensities
0
and the Fe/Ni molar ratio. However, it can be confirmed that the metallic Fe content was
always <15%, as shown in Table 3. These results indicated that although the presence of Fe
can promote reduction of the Ni species, the Fe species was mostly remained in their oxide
forms, i.e., Fe O , Fe O , and FeO, as confirmed by H -TPR measurements (Fig. 3 and
2
3
3
2
2
Table 2).
3
.2. Reductive Amination of IPA
Fig. 5 shows the effect of Fe loading on the reductive amination of IPA over Ni-xFe/Al O
2
3
-1
-1
catalysts (reaction conditions: T = 150 °C, SV = 500 mL∙min ∙g , and IPA/NH /H feed
3
2
composition (molar ratio) = 1/8/6). MIPA was the main product for all reactions, while DIPA
and acetone were produced as by-products. As can be deduced from the reductive amination
mechanism, acetone is formed initially by the dehydrogenation of IPA, while DIPA is
produced by the consecutive reaction of MIPA with IPA through condensation/dehydration. In
addition to MIPA, DIPA, and acetone, a number of other possible by-products can be
produced by condensation, decarbonylation, disproportionation, and hydrogenolysis,

including C –C compounds, di-isopropyl ether, and trimethylamine. However, the sum of the
1
3
selectivities of MIPA, DIPA, and acetone was almost above 99% under our experimental
conditions, thus indicating that the formation of other side products was negligible.
The reductive amination of IPA over the various Ni-xFe/Al O catalysts is shown in
2
3
Fig. 5(a), where the deactivation rate, defined as the difference in conversion between initial
and at 9 h divided by the initial conversion, was <1.5% for all catalysts examined herein. As
shown, the conversion exhibited a volcano shaped curve as a function of the Fe/Ni molar
ratio (Fig. 5(b)), and all Fe-containing catalysts showed better performance than the Ni/Al O
2
3
catalyst. More specifically, the Ni-0.3Fe/Al O catalyst exhibited the optimal catalytic
2
3
0
performance due to the presence of a highly exposed metallic Ni surface, as confirmed by
H -TPR (Fig. 3) and XPS (Fig. 4) measurements. Hence, the observed conversions well
2
0
correlated with the surface metallic Ni content as shown in Fig. 5(c). This result clearly
0
indicated that the highly exposed Ni species promotes the conversion of IPA in the reductive
amination reaction, as described in previous studies [4, 8, 22]. However, above the optimal
Fe/Ni ratio, the IPA conversion decreased, likely due to the negative effect of excess Fe
0
addition, which are mainly caused by lower activity of Fe species and a decrease in Ni
content [16].
In order to optimize the reaction parameters, the effects of NH partial pressure, H partial
3
2
pressure, space velocity (SV), and reaction temperature were totally investigated, and the
results are summarized in Table 4 and Fig. S3 – S6. Initially, the effect of H partial pressure
2
-1
-1
was examined at 150 °C with an SV of 500 mL∙min ∙g in the reductive amination over a
Ni-0.3Fe/Al O catalyst. The IPA and NH partial pressures were fixed at 3 and 24 kPa,
2
3
3
respectively. In this case, the IPA conversion decreased gradually upon increasing the H2
partial pressure. On the other hand, MIPA, DIPA, and acetone selectivities remained
relatively constant. As indicated in the table, a maximum MIPA yield of 78.0% was achieved

at an IPA/NH /H molar ratio of 1/8/4. We previously reported that γ-Al O supported Ni and
3
2
2
3
Co catalysts behaved differently in the conversion of IPA upon variation in the H partial
2
pressure [3, 8]. Upon increasing the H partial pressure, the Co-based catalyst exhibited
2
increased IPA conversion, whereas that of Ni/γ-Al O catalyst was rather decreased. Sewell et
2
3
al. [11] also reported similar results for Co/SiO and Ni/SiO catalysts in the reductive
2
2
amination of ethanol. This is because only a small amount of hydrogen is required to prevent
the metal nitride formation in the Ni-based catalyst, which can result in catalyst deactivation
during the reductive amination reaction [25, 26]. Thus, it was expected that the bimetallic Ni-
Fe catalyst could prevent metal nitride formation in the presence of small quantities of H2,
while an excess of H leads to negative effects caused by the competitive adsorption amongst
2
the IPA, H , and NH .
2
3
The effect of NH partial pressure was then investigated at 150 °C with an SV of 500
3
-1
-1
mL∙min ∙g and at the optimal IPA/H molar ratio of 1:4. In this case, an increase in IPA
2
conversion was observed with increasing NH partial pressure. MIPA selectivity was also
3
proportional to the NH partial pressure, while that of acetone was inversely proportional. In
3
terms of the reaction mechanism, acetone is initially formed by the dehydrogenation of IPA,
and this acetone is subsequently reacted with NH to form MIPA [3, 4, 8]. An excess of NH
3
3
therefore exhibited a positive effect on MIPA production, as the reaction of the acetone
intermediate with NH was promoted. In this context, a decreased contact time caused by
3
increasing the SV promoted the formation of acetone, while a low SV resulted in increased
MIPA selectivity and IPA conversion.
As the reaction temperature was increased from 130 to 150 °C, the IPA conversion
progressively increased from 76.8 to 96.5% due to favorable reaction kinetics. However, an
-1
-1
optimal MIPA yield of 79.3% was obtained at 150 °C (300 mL∙min ∙g SV and IPA/NH /H
3
2
=
1/8/4), and the yield decreased once again above this temperature. It was due to strong

enhancement of the side reactions at high temperatures. More specifically, the DIPA
formation which was produced from MIPA through the consecutive reaction was more
favorably formed at higher temperatures as shown in Table 4.
Fig. 6 shows the H on/off experiments for the reductive amination of IPA using a Ni-
2
0
.3Fe/Al O catalyst. For the first 4 h, the conversion and selectivities remained constant
2 3
-1
-1
under a temperature of 150 °C, SV of 300 mL∙min ∙g , and IPA/NH /H feed composition of
3
2
1
/8/4. Under these conditions, the initial conversion and selectivities towards MIPA and
acetone were 95.3%, 86.7%, and 4.2%, respectively. After 4 h on stream, the flow of mixed
N and H gas mixture was switched to a flow of pure N for 6 h. During this period, the IPA
2
2
2
conversion decreased gradually, while the MIPA selectivity dropped to ca. 70%, and the
quantity of diminished MIPA was compensated by the increased acetone content. It therefore
appeared that excess H effectively hindered the phase transition of the metal to the metal
2
nitride, and thus deactivation could be prevented under excess H conditions [3, 4, 8, 25, 26].
2
Thus, the Ni-0.3Fe/Al O catalyst was rapidly deactivated due to metal nitride formation
2
3
under feed flow in the absence of H . However, upon re-exposure to a feed stream containing
2
H gas, the initial conversion and selectivity were completely recovered, and deactivation was
2
not observed.
Finally, the long-term stability of the Ni-0.3Fe/Al O catalyst in the reductive amination
2
3
reaction was studied as shown in Fig. 7. The Ni-0.3Fe catalyst exhibited the best performance
amongst the catalysts examined herein, and it exhibited a relatively constant conversion and
-1
-1
selectivity to MIPA 100 h on stream (T = 150 °C, SV = 300 mL∙min ∙g , and IPA/NH /H
3
2
feed composition (molar ratio) = 1/8/4). In our group, we have conducted previous studies on
reductive amination of the IPA using Ni- and Co- based catalysts [3, 4, 8]. Although the
experimental conditions are somewhat different, the catalysts in this study exhibited higher
catalytic activity at lower reaction temperatures (conversion at 95.6% at 150 °C) compared to

previous studies (Co(23 wt%)/γ-Al O [3]: conversion ca. 95% at 210 °C, Ni(17 wt%)/η-
2
3
Al O [4]: conversion ca. 78% at 160 °C, and Ni(17 wt%)/γ-Al O [8]: conversion ca. 88% at
2
3
2
3
1
70 °C). This result suggested that the addition of Fe species to Ni/Al O catalysts may result
2 3
in increased reducibility of NiO species. In this study, to demonstrate this phenomenon, H2-
TPR analysis was performed, and it was confirmed that reduction temperatures of NiO
related peaks were the lowest at Ni-0.3Fe/Al O catalyst. In addition, it was also confirmed
2
3
that this tendency was in good agreement with XPS results and reactivity of the reductive
amination. Therefore, it appeared that the Ni-0.3Fe/Al O catalyst developed herein could be
2
3
applied in practical applications for the formation of MIPA, due to its good activity and
stability. In addition, it is considered that it can be effectively applied to various fields such as
hydrogenation and dehydrogenation processes requiring the metallic functionalities.
4
. Conclusion
Ni-Fe/γ-Al O catalysts containing different Fe/Ni molar ratios (i.e., 0–0.7) were
2
3
successfully synthesized via a co-impregnation method. From the X-ray diffraction results,
we could conclude that impregnated metal precursors existed on the surface of the Al O
2
3
support in the form of NiO and Fe O following calcination at 500 °C for 2 h under a flow of
2
3
air. Following subsequent reduction at 600 °C for 3 h, an Ni-Fe alloy was formed. In addition,
H -temperature-programmed reduction measurements indicated that compared to a Ni/Al O
2
2
3
catalyst, the areas of reduced Ni related peaks increased, while their reduction temperatures
decreased in Fe-promoted Ni-xFe/Al O3 catalysts. Furthermore, analysis by X-ray
2
0
photoelectron spectroscopy indicated that the surface Ni content increased in the presence of
an Fe promoter after reduction at 600 °C for 3 h. These results clearly demonstrated that the

Fe species can improve the reducibility of the Ni component. Moreover, the reducibility and
0
surface Ni content varied in the presence of different Fe/Ni molar ratios, with a Ni-
0
0
.3Fe/Al O catalyst exhibiting the highest reducibility and surface Ni content. We also
2
3
found that in the reductive amination of iso-propanol (IPA), the IPA conversion well
0
correlated with the surface Ni content. We could therefore conclude that the Ni-0.3Fe/Al O
2
3
catalyst showed the highest catalytic performance amongst the all catalysts examined herein
for the reductive amination reaction, and it maintained constant conversion and selectivity
over 100 h on stream in the presence of hydrogen.
Acknowledgement
This research was supported by Basic Science Research Program through the National
Research Foundation of Korea (NRF) funded by the Ministry of Education (2017 xxxxx grant
number).

References
[
[
[
1] K. Hayes, Appl. Catal. A: Gen. 221 (2001) 187-195.
2] H.-Y. Jeon, C.-H. Shin, H.J. Jung, S.B. Hong, Appl. Catal. A: Gen. 305 (2006) 70-78.
3] J.H. Cho, J.H. Park, T.-S. Chang, G. Seo, C.-H. Shin, Appl. Catal. A: Gen. 417 (2012)
3
13-319.
[
[
[
[
4] J.H. Cho, S.H. An, T.-S. Chang, C.-H. Shin, Catal. Lett. 146 (2016) 811-819.
5] R.P. Tripathi, S.S. Verma, J. Pandey, V.K. Tiwari, Curr. Org. Chem. 12 (2008) 1093-1115.
6] S. Gomez, J.A. Peters, T. Maschmeyer, Adv. Syn. Catal. 344 (2002) 1037-1058.
7] M.E. Domine, M.C. Hernández-Soto, M.T. Navarro, Y. Pérez, Catal. today 172 (2011) 13-
2
[
1
[
1
[
1
0.
8] J.H. Cho, J.-H. Park, T.-S. Chang, J.-E. Kim, C.-H. Shin, Catal. lett. 143 (2013) 1319-
327.
9] Y. Zhang, Y. Zhang, C. Feng, C. Qiu, Y. Wen, J. Zhao, Catal. Commun. 10 (2009) 1454-
458.
10] O. Solcova, D.-C. Uecker, U. Steinike, K. Jiratova, Appl. Catal. A: Gen. 94 (1993) 153-
60.
[
[
[
11] G. Sewell, C. O'Connor, E. Van Steen, Appl. Catal. A: Gen. 125 (1995) 99-112.
12] B.H. Davis, Catalysis Today 84 (2003) 83-98.
13] S.-H. Kang, J.-H. Ryu, J.-H. Kim, S.-J. Seo, Y.-D. Yoo, P.S.S. Prasad, H.-J. Lim, C.-D.
Byun, Korean J. Chem. Eng. 28 (2011) 2282-2286.
[
[
[
14] D. Huang, M. Ke, X. Bao, H. Liu, Ind. Eng. Chem. Res. 55 (2016) 1192-1201.
15] F. Meshkani, M. Rezaei, Korean J. Chem. Eng. 32 (2015) 1278-1288.
16] L. Wang, D. Li, M. Koike, S. Koso, Y. Nakagawa, Y. Xu, K. Tomishige, Appl. Catal. A:
Gen. 392 (2011) 248-255.
17] Q. Chen, S. Hu, J. Xiang, S. Su, L. Sun, Y. Wang, L. Zhang, H. Chi, Fuel Process.
Technol. 146 (2016) 56-61.
[
[
[
[
(
[
18] S. Abelló, E. Bolshak, D. Montané, Appl. Catal. A: Gen. 450 (2013) 261-274.
19] F. Meng, P. Zhong, Z. Li, X. Cui, H. Zheng, J. Chem. 2014 (2014).
20] F. Liu, H. He, C. Zhang, Z. Feng, L. Zheng, Y. Xie, T. Hu, Appl. Catal. B: Environ. 96
2010) 408-420.
21] J.W. Bae, S.-M. Kim, S.-H. Kang, K.V. Chary, Y.-J. Lee, H.-J. Kim, K.-W. Jun, J. Mol.
Catal. A: Chem. 311 (2009) 7-16.
22] V. Zamlynny, L. Kubelková, E. Babůrek, K. Jirátová, J. Nováková, Appl. Catal. A: Gen.
69 (1998) 119-125.
23] M.-Y. Wang, T. Shen, M. Wang, D.-E. Zhang, Z.-w. Tong, J. Chen, Sensor. Actuat. B:
Chem. 190 (2014) 645-650.
24] C. Yang, G. Wang, Z. Lu, J. Sun, J. Zhuang, W. Yang, J. Mater. Chem. 15 (2005) 4252-
257.
[
1
[
[
4
[
[
25] A. Baiker, W. Caprez, W.L. Holstein, Ind. Eng. Chem. Prod. RD. 22 (1983) 217-225.
26] A. Baiker, D. Monti, Y.S. Fan, J. Catal. 88 (1984) 81-88.
1
9

Tables and figure captions
Table 1. Specific surface areas (SBET), total pore volumes (V ), and average pore diameters
P
(
D ) of the γ-Al O support and the Ni-xFe/Al O catalysts calcined at 500 °C for 2 h
P 2 3 2 3
a
x in
Ni-xFe/Al O
S
V
DP
BET
P
2
3
(m /g)
(cm /g)
0.699
0.472
0.401
0.384
0.380
0.367
(nm)
14.4
10.8
8.5
2
3
γ-Al O
2
3
1
94
support
0
146
144
134
151
154
0
0
0
0
.1
.3
.5
.7
8.8
8.4
8.2
a
Calculated from the desorption branch of the N isotherm by the BJH method
2
2
0

Table 2. Evolved peak temperatures and relative areas obtained from the H -TPR profiles of
2
Ni-xFe/Al O catalysts. The samples were calcined at 500 °C for 2 h prior to H -TPR
2
3
2
measurements
Peak temperature (°C)
x in
(Relative peak area, a.u.)
Ni-xFe/Al O
2
3
δ-1
δ-2
α
β
γ
4
30.0
604.6
(100.8)
701.8
(1.8)
0
-
-
(19.2)
3
61.7
399.1
(1.0)
455.2
(75.1)
581.2
(107.4)
667.9
(29.0)
0
0
0
0
.1
.3
.5
.7
(1.0)
3
54.6
392.8
(22.9)
431.0
(188.3)
576.3
(76.3)
-
-
-
(
(
(
33.5)
3
66.5
406.9
(51.7)
457.0
(145.2)
585.9
(54.6)
41.1)
3
70.6
409.0
(47.9)
460.0
(134.7)
590.0
(27.0)
40.0)
2
1

Table 3. Peak positions and relative atomic compositions of the Ni-xFe/Al O catalysts
2
3
reduced at 600 °C for 3 h under H flow, following calcination at 500 °C for 2 h. These
2
results are derived from the Ni 2p3/2 and Fe 2p XPS spectra
3/2
Ni 2p3/2
Fe 2p3/2
x in Ni-
Binding energy (eV)
(Atomic composition, %)
Binding energy (eV)
(Atomic composition, %)
0
xFe/Al O
2
3
Ni /
Ni
2
+
2
+
2+
0
Ni in
NiO
Ni in
NiAl O4
0
2+
3+
Ni
Fe
Fe
Fe
2
8
52.5
855.0
(48.5)
857.3
(26.0)
0
0.34
0.48
0.61
0.58
0.55
-
-
-
(25.5)
8
53.3
855.6
(59.7)
857.2
(7.8)
707.0
(13.6)
709.6
(52.1)
711.8
(34.3)
0
0
0
0
.1
.3
.5
.7
(32.5)
8
53.5
855.7
(62.1)
707.0
(13.3)
709.6
(50.1)
711.8
(36.6)
-
-
-
(37.9)
8
53.5
855.7
(63.4)
707.0
(13.5)
709.6
(45.9)
711.8
(40.6)
(36.6)
8
53.1
855.4
(64.5)
707.0
(12.0)
709.6
(55.4)
711.8
(32.6)
(35.5)
2
2

Table 4. Variations in the conversion and selectivities towards MIPA, DIPA, and acetone in
the reductive amination reaction over the Ni-0.3Fe/Al O catalyst under a range of
2
3
experimental conditions
Experimental Condition
b
Selectivity (%)
a
b
MIPA
Feed Composition
Conv.
%)
b
Variable
Yield
%)
SV
(
Temp. (°C)
-1 -1
(
(
mL∙min ∙g )
IPA
NH3
H2
MIPA
DIPA
Acetone
4
6
8
93.3
90.7
86.9
84.5
83.3
88.5
91.1
93.3
95.6
93.3
89.7
85.6
82.8
76.8
90.2
93.7
95.6
96.5
96.5
83.6
83.8
83.9
84.5
55.0
71.1
78.6
83.6
83.0
83.6
82.6
81.2
79.9
85.8
85.5
84.2
83.0
80.6
78.6
8.8
8.8
7.1
6.9
78.0
76.0
72.9
71.4
45.8
62.9
71.6
78.0
79.3
78.0
74.1
69.5
66.2
65.9
77.1
78.9
79.3
77.8
75.8
H Partial
Pressure
2
1
8
150
500
8.3
7.2
1
0
8.0
7.1
2
4
6
8
24.4
15.4
11.1
8.8
19.9
12.9
9.8
NH3
Partial
Pressure
1
1
4
150
150
500
7.1
3
5
00
00
11.7
8.8
5.2
7.1
SV
8
4
700
00
100
6.6
10.1
12.8
15.4
10.8
7.2
9
5.2
1
4.5
1
1
1
1
1
1
30
40
45
50
55
60
3.3
6.9
9.1
6.1
Temp.
1
8
4
300
11.7
13.9
14.8
5.2
5.4
6.1
a
IPA/NH /H molar ratio with a constant IPA partial pressure of 3 kPa
3
2
b
Conversion and selectivities were obtained at 30 min on stream
2
3

Fig. 1. (a) N adsorption-desorption isotherms, and (b) pore size distributions of the γ-Al O
2
2
3
support and the Ni-xFe/Al O catalysts calcined at 500 °C for 2 h.
2
3
Fig. 2. XRD patterns of the Ni-xFe/Al O catalysts: (a) calcined at 500 °C for 2 h, and (b)
2
3
then consecutively reduced at 600 °C for 3 h.
Fig. 3. (a) H -TPR profiles of the Ni-xFe/Al O catalysts calcined at 500 °C for 2 h, and (b)
2
2
3
the areas of the peaks corresponding to NiO, and their sums. α represents NiO that weakly
interacts with the support, and β represents NiO that strongly interacts with the support.
Fig. 4. XPS spectra of (a) Ni 2p, and (b) Fe 2p for the Ni-xFe/Al O catalysts calcined at
2
3
5
00 °C for 2 h and reduced at 600 °C for 3 h under a flow of H (50 mL/min).
2
Fig. 5. Influence of the Fe content in the Ni-xFe/Al O catalysts on the reductive amination
2
3
-1
-1
of IPA. Reaction conditions: T = 150 °C, SV = 500 mL∙min ∙g , and IPA/NH /H feed
3
2
composition (molar ratio) = 1/8/6 with a constant IPA partial pressure of 3 kPa. (a)
Conversion versus time on stream, (b) conversion and selectivities as a function of the Fe/Ni
0
molar ratio, and (c) correlations of the conversions and metallic Ni composition with the
Fe/Ni molar ratio.
Fig. 6. H on/off experiments for the reductive amination of IPA over Ni-0.3Fe/Al O catalyst.
2
2
3
The reductive amination reaction was initially conducted in the presence of hydrogen (T =
-1
-1
1
50 °C, SV = 300 mL∙min ∙g , and IPA/NH /H molar feed composition = 1/8/4). The total
3 2
flow of the N + H mixture was changed to pure N after 4 h and maintained for 6 h prior to
2
2
2
returning to the initial feed composition.
2
4

Fig. 7. Long-term stability of the Ni-0.3Fe/Al O catalyst in the reductive amination of IPA.
2
3
-1
-1
Reaction conditions: T = 150 °C, SV = 300 mL∙min ∙g , and IPA/NH /H molar feed
3
2
composition = 1/8/4 with a constant IPA partial pressure of 3 kPa.
2
5