Nickel and nickel-magnesia catalysts active in the hydrogenation of 1,4-butanedinitrile

Article abstract of DOI:10.1006/jcat.2000.3084

Several NiO-MgO systems were synthesized to be studied as nickel catalysts for the hydrogenation of 1,4-butanedinitrile in the gas phase and compared with a bulk NiO of controlled morphology. All samples were characterized by XRD, BET, TPR, TPD, SEM, and H₂ chemisorption techniques. The Ni-MgO systems had higher activities than the Ni bulk catalyst. The most active catalyst at all reaction temperatures was type R4C_B which had homogeneous particles of about 1000 A, the highest metal surface area, and the highest coverage with weakly bound hydrogen. The presence of basic magnesia suppresses the condensation reactions and consequently favors the elimination of amines, and prevents catalyst deactivation. The selectivity toward the different products not only depends on the catalytic properties but can also be modified by controlling the hydrogen/dinitrile ratio. The highest selectivity to 4-aminobutanenitrile was achieved by catalyst R4C_B, with 85% at 100% conversion and working at a space velocity of 13,000 h^-1 and 343 K. This selectivity could be increased by lowering the hydrogen/butanedinitrile ratio.

Full text of DOI:10.1006/jcat.2000.3084

Journal of Catalysis 197, 210–219 (2001)
doi:10.1006/jcat.2000.3084, available online at http://www.idealibrary.com on
Nickel and Nickel–Magnesia Catalysts Active in the Hydrogenation
of 1,4-Butanedinitrile
Marc Serra, Pilar Salagre, ,1 Yolanda Cesteros, Francisco Medina,† and Jesu´s E. Sueiras†
Facultat de Qu´ımica, Universitat Rovira i Virgili, Pl. Imperial Tarraco, 1, 43005, Tarragona, Spain; and †Escola Te`cnica Superior d’Enginyeria
Qu´ımica, Universitat Rovira i Virgili, Pl. Imperial Tarraco, 1, 43005, Tarragona, Spain
Received July 28, 2000; revised September 28, 2000; accepted September 28, 2000
in the post Prins (6) has recently proposed the catalytic
hydrogenation of 2-methylglutaronitrile as an alternative
process to obtain these picoline compounds. The 2-methyl-
glutaronitrile is a by-product in the DuPont process for
manufacturing adiponitrile from butadiene (6, 7) and,
therefore, such a catalytic process means that by-products
can be used as raw materials.
Several NiO–MgO systems were synthesized to be studied as
nickel catalysts for the hydrogenation of 1,4-butanedinitrile in the
gas phase and compared with a bulk NiO of controlled morphol-
ogy. All samples were characterized by XRD, BET, TPR, TPD,
SEM, and H2 chemisorption techniques. The Ni–MgO systems had
higher activities than the Ni bulk catalyst. The most active catalyst
at all reaction temperatures was type R4CB which had homoge-
Nitrile hydrogenation products are composed mainly
of a mixture of primary, secondary, and tertiary amines
(8, 9). The condensation reactions between a highly reac-
tive intermediate imine and the primary amine always lead
to the formation of such products as secondary and ter-
tiary amines together with the primary amines (10). The
selectivity of the nitrile hydrogenation process is of great
importance and depends on the properties of the starting
nitrile, the catalyst used, and the reaction conditions. How-
ever, the catalyst is the most important factor (2, 11). Cata-
lysts based on Co, Ni, and Ru are mostly used to produce
primary amines. Cu and Rh catalysts are mostly used to
obtain secondary amines while tertiary amines can be pre-
pared by using Pt and Pd catalysts (12, 13).
˚
neous particles of about 1000 A, the highest metal surface area,
and the highest coverage with weakly bound hydrogen. The pres-
ence of basic magnesia suppresses the condensation reactions and
consequently favors the elimination of amines, and prevents cata-
lyst deactivation. The selectivity toward the different products not
only depends on the catalytic properties but can also be modified by
controlling the hydrogen/dinitrile ratio. The highest selectivity to
4-aminobutanenitrile was achieved by catalyst R4CB, with 85% at
1
100% conversion and working at a space velocity of 13,000 h
and 343 K. This selectivity could be increased by lowering the
c
hydrogen/butanedinitrile ratio.
2001 Academic Press
Key Words: nickel–magnesia catalysts; magnesia; succinonitrile;
dinitrile; 1,4-butanedinitrile hydrogenation.
Amines are usually prepared industrially in the liquid
phase at high hydrogen pressures. Nickel catalysts, how-
ever, have been used at lower pressures, in the gas phase,
although an excess of ammonia was essential to suppress
secondary and tertiary amine formation (10, 14). Also, the
addition of potassium in dopant amounts enhances the
selectivity of nickel catalysts toward primary amines in
adiponitrile hydrogenation (15–18). Recently, precursors
such as hydrotalcites have made it possible to vary the
MgO/Al2O3 ratio and thus control the acidity of the cat-
alysts (19, 20). When this ratio increases, the selectivity
to primary amines in the acetonitrile hydrogenation also
increases.
Double selective catalysts are required if 6-amino-
hexanenitrile is to be obtained selectively in the adiponi-
trile hydrogenation. Such catalysts lead to primary amines
with only one hydrogenated nitrile group. This selectivity
could be related to the crystallinity of the active phase. In
this respect, Ni/ -Al2O3 catalysts with nickel particles of
INTRODUCTION
Hydrogenation of nitriles is one of the most widely used
methods to obtain amines commercially (1, 2), and pro-
duction of aliphatic amines has been the most important
application. One interesting industrial process is the hydro-
genation of 1,6-hexanedinitrile (adiponitrile) to obtain 1,6-
hexanediamine which is used as a precursor in the prepa-
ration of Nylon-6,6 (3–5).
Recent studies have shown that the hydrogenation–
cyclization of dinitriles is a new interesting research field in
organic synthesis to obtain N-heterocyclic compounds. To-
day, pyridine derivatives such as - and -picolines are used
to obtain precursors of many chemical products which are
important in medicine, farming, and industry. These com-
pounds have usually been prepared by aldol condensation
1 To whom correspondence should be addressed. Fax: +34 977559563.
E-mail: salagre@quimica.urv.es.
210
0021-9517/01 $35.00
c
Copyright
2001 by Academic Press
All rights of reproduction in any form reserved.

Ni CATALYSTS FOR 1,4-BUTANEDINITRILE HYDROGENATION
211
octahedral morphology (21) and crystalline Raney nickel by adhesive tape for X-ray diffraction (XRD) monitoring.
catalysts (22) have shown very high selectivity to the A glove box was used for mounting. The catalytic activity
monoamine compound. It has also been observed that was measured in situ, in the same reactor after reduction,
the morphology and size of nickel particles are related to where gas purges, positive gas pressures, and Schlenk tech-
the method used to prepare NiO–MgO systems (23).
In this study, we propose that Ni–MgO systems with
nickel metallic phases of different crystallinity can be used
to control the selectivity to monoamine or to cyclic deriva-
tives. Different magnesia sources will suppress the conden-
sation reactions and consequently the lifetime of the cata-
lysts will increase. The catalytic tests were performed in
the gas phase at 1 atm pressure for the hydrogenation of
1,4-butanedinitrile.
niques were used when necessary.
BET Surface Areas
Surface areas were calculated from the nitrogen adsorp-
tion isotherms at 77 K using a Micromeritics ASAP 2000
surface analyzer and a value of 0.164 nm2 for the cross sec-
tion of the nitrogen molecule.
X-ray Diffraction (XRD)
Powder X-ray diffraction patterns of the different sam-
ples were obtained with a Siemens D5000 diffractometer
using nickel-filtered Cu K radiation. Samples were dusted
on double-sided adhesive tape and mounted on glass mi-
croscope slides. The patterns were recorded over a range
of 2 angles from 10 to 90 and crystalline phases were
identified using the Joint Committee on Powder Diffrac-
tion Standards (JCPDS) files.
The XRD patterns of MgO are very similar to those
of NiO. The peak positions corresponding to 2 angles
(with the relative intensities in parentheses), taken from
the JCPDS files, are 36.95 (10), 42.91 (100), 62.31 (52),
74.68 (4), and 78.61 (12) for the MgO phase and 37.28 (91),
43.30 (100), 62.92 (57), 75.44 (16) and 79.39 (13) for the NiO
phase. Both 2 peaks are assigned to the faces (1 1 1), (2 0
0), (2 2 0), (3 1 1), and (2 2 2), respectively. The 2 angles
and the relative intensities (in parentheses) are 44.51 (100),
51.85 (42), and 76.36 (21) for the Ni phase.
EXPERIMENTAL
Catalysts Preparation
NiO of octahedral morphology was prepared according
to the method of Estelle´ (24). It involved thermally decom-
posing Ni(NO3)2 6H2O at 373 K for 14 days in order to
obtain Ni3(NO3)2(OH)4 as a single phase which was then
calcined at 623 K for 40 min to obtain the NiO sample, re-
ferred to as NiO. The corresponding catalyst was obtained
by reduction under H2 at 673 K for 4 h. This sample was
named Ni-Oh.
Three NiO–MgO samples were prepared by modifying
the MgO source and/or the preparative procedure accord-
ing to the method previously described (23) which is briefly
summarized as follows. The NiO/MgO weight ratio was 4 : 1
for all samples (referred to as 4). Two MgO sources were
used: commercial MgOC (BET surface area 223 m2 g 1) and
a magnesia obtained by thermal decomposition of magne-
sium nitrate hexahydrate, MgON (BET surface area 2.5 m2
This technique was also used to determine the reduc-
tion degree ( ) of the catalysts by means of the Rietveld
method (25). This method implies quantitative phase anal-
ysis of multicomponent mixtures from the X-ray powder
diffraction data. It is only necessary to know the crystal
structure of each phase of interest.
g
1), referred to as C and N, respectively.
Two preparative procedures were performed: path A and
path B (referred to as subscripts A and B, respectively).
Path A involved a direct calcination of a Ni(NO3)2 6H2O
and MgO mixture at 673 K (samples 4CA and 4NA). In path
B, a controlled thermolysis of a nickel nitrate and MgOC
mixture was carried out to obtain Ni3(NO3)2(OH)4. This
sample was later calcined at a lower temperature (523 K)
than in case A (sample 4CB).
Temperature-Programmed Reduction (TPR)
Temperature-programmed reduction (TPR) studies
were carried out in a Labsys/Setaram TG DTA/DSC ther-
mobalance equipped with a 273–1273 K programmable
Finally, all samples obtained by paths A and B were re-
duced under pure H2 at 673 K for 6 h (catalysts R4CA,
R4NA, and R4CB).
temperature furnace. The accuracy was
1 g.
Each sample (90 mg) was first heated at 423 K in an
Ar flow (80 cm3/min) until no change of weight was de-
tected. The sample was then heated in a 5 vol% H2/Ar flow
(80 cm3/min) from 423 K to 1173 K at a rate of 5 K min
1
Air-free Sampling
and maintained at 1173 K until the reduction process was
finished.
After the reduction step, the catalysts were always han-
dled under air-free conditions. The samples were trans-
ferred in degassed isooctane under a hydrogen atmosphere
at room temperature. The isooctane surface-impregnated
samples were further isolated from the air either by gold
Scanning Electron Microscopy (SEM)
Scanning electron micrographs were obtained with a
film for the scanning electron microscopy (SEM) study or JEOL JSM6400 scanning microscope operating at an

212
SERRA ET AL.
accelerating voltage in the range 30–35 kV, a work distance
(w.d.) between 7–9 mm and at a magnification factor of
40,000 to 50,000.
RESULTS AND DISCUSSION
X-Ray Diffraction (XRD)
Sample NiO shows peaks which can be attributed to crys-
talline NiO whereas the powder diffraction patterns of the
NiO–MgO systems (23) show peaks which correspond to a
crystalline solid solution of the components.
Temperature-Programmed Desorption (TPD)
TPD spectra were obtained with a Fisons QTMD 150 gas
desorption unit equipped with a 273–1273 K programmable
temperature furnace and a mass spectrometer detector.
Samples were reduced in a pure H2 flow (80 cm3/min)
from room temperature at 1 K min 1 up to 673 K. They were
isothermally maintained at this last temperature for 4 h to
obtain the catalyst Ni-Oh and for 6 h to obtain catalysts
R4CA, R4CB, and R4NA. The samples were then cooled to
room temperature in the hydrogen flow, evacuated for 1 h
at low pressures (<1 Pa), and heated at a rate of 10 K/min
up to 1123 K. The hydrogen desorption was detected with
the mass spectrometer.
Table 1 shows the crystalline phases obtained by XRD for
the catalysts. The Ni-Oh catalyst contains crystalline nickel
as a single phase. However, the reduced catalysts Ni–MgO
(samples R4CA, R4NA, and R4CB) have powder diffraction
patterns corresponding to two phases: a crystalline solid so-
lution phase and a less crystalline Ni phase (see Table 1).
The detection by XRD of a solid solution phase for the
catalysts indicates that the reduction of NiO is not com-
plete. Therefore, the presence of magnesia hinders the re-
duction process, in agreement with what has been reported
by Arena et al. (26–28).
The degrees of reduction ( ) were determined from
the corresponding diffraction patterns using the Rietveld
method (25) for all catalysts (Table 1). The Ni-Oh catalyst
had the highest degree of reduction, as expected.
When path A was used to prepare the Ni–MgO catalysts,
the degrees of reduction were lower than those when path
B was used (0.12 for R4CA and 0.29 for R4CB) for the same
magnesia source. The reduction degree of catalyst R4CA
is lower because the diffusion between the NiO and MgO
phases (which hinders the reduction of NiO) is better at the
higher calcination temperature used in path A (673 K) than
in path B (523 K).
Hydrogen Chemisorption
Hydrogen chemisorption was measured with a Mi-
cromeritics ASAP 2010C instrument equipped with a tur-
bomolecular pump. Samples had been reduced previously
under the same conditions under which the catalysts had
been prepared. After reduction, the chemisorbed hydrogen
was removed in a stream of 30 ml/min of He for 30 min at
683 K. The sample was subsequently cooled to 303 K in the
same He stream. The chemisorbed hydrogen was analyzed
at 303 K. The nickel surface area was calculated assuming
a stoichiometry of one hydrogen molecule per two surface
nickel atoms and an atomic cross-sectional area of 6.49
10 20 m2/Ni atom.
The reduction degree for catalyst R4NA is higher (
=
0.56) than that for catalyst R4CA ( = 0.12). Thisisin agree-
ment with the diffusion problems observed (23) for the
Determination of the Catalytic Activity
˚
MgON which had larger particles (crystallite sizes >1200 A)
˚
than the MgOC (crystallite sizes of 47 A). TPR results, dis-
cussed below, confirm this behavior.
The gas-phase hydrogenation of 1,4-butanedinitrile was
studied in a tubular fixed-bed flow reactor heated by an
oven equipped with a temperature control system. The re-
actor was filled with catalyst (500 mg) and the catalytic re-
action for 1,4-butanedinitrile vapor at 353 K was measured
at 1 atm pressure, using space velocities between 6500–
13000 h 1 and reaction temperatures between 343–453 K.
No diffusional retardation was observed as the residence
times lie on the linear portion of the curve obtained when
plotting conversion against residence time for different
catalyst volumes. Reaction products were analyzed by an
on-line HP 5890 gas chromatograph equipped with a 30-m
“commercial Rtx-5 amine” capillary column and a flame
ionization detector.
BET Surface Areas
The surface area of the NiO sample is 11 m2 g 1 and those
for the NiO–MgO systems range between 24 and 36 m2 g
.
1
After reduction of the catalytic precursors, a small decrease
in surface areas was observed due to the formation of a new
TABLE 1
Characterization of Catalysts
Ni-Oh
Ni
R4CA
Ni
R4NA
Ni
R4CB
Ni
Conversion and selectivity were defined by the following
equations: conversion (% ) = [moles of 1,4-butanedinitrile
consumed] 100/[moles of 1,4-butanedinitrile charged];
Crystalline phases (XRD)
NiO–MgO NiO–MgO NiO–MgO
Reduction degree (XRD)a
1
0.2
0.12
1.6
0.56
2.0
0.29
4.1
1
Metallic areab (m2 g
)
selectivity (% )
tion] 100/[moles of 1,4-butanedinitrile consumed]. The
carbon mass balance of the process was always maintained.
=
[moles of one product of reac-
a Reduction degree obtained by using the Rietveld method (25).
b Metal surface area, calculated from H2 chemisorption.

Ni CATALYSTS FOR 1,4-BUTANEDINITRILE HYDROGENATION
213
FIG. 1. Scanning electron micrograph of precursor NiO.
crystalline and more compact Ni phase. The low surface Scanning Electron Microscopy (SEM)
areas of the catalysts R4CA and R4CB (and their cata-
Scanning electron microscopy was used to monitor the
morphology and particle size of the different precursors
and their reduced forms. Figures 1, 2, and 3 show the mi-
crographs of the NiO, 4CA, and 4CB samples, respectively.
The NiO sample and its reduced form, Ni-Oh, ap-
pear to have the same homogeneous octahedral particles
lyst precursors) contrast with the high area of their MgOC
source (223 m2 g 1). This is in agreement with the results
reported by Anderson and Hohrlock (29), who observed
considerable sintering of high-surface-area magnesia sam-
ples in the presence of low amounts of water vapor at the
calcination temperatures used.
˚
with dimensions around 2500 A. Therefore, the use of
Ni3(NO3)2(OH)4 as a single precursor of NiO phase be-
comes a good method to obtain homogeneous octahedral
particles in a reproducible way. This is in agreement with
Temperature-Programmed Reduction (TPR)
The reducibility of the NiO sample was estimated by TPR what has been reported by Estelle´ (24).
considering its initial reduction temperature in order to al-
For the NiO–MgO systems, different pathsofpreparation
low comparison with the initial reduction temperature ob- lead to samples with different morphologies and particle
tained for the NiO–MgO systems previously (23). Our goal size distributions. The sample prepared via path B is highly
˚
in performing these reducibility studies is to confirm the dif- homogeneous and has small particles (around 1000 A) with
ferences in the degrees of reduction ( ) observed by XRD non-well-defined octahedral forms. In contrast, the sample
for the four catalysts, as mentioned above. The initial reduc- prepared via path A has spherical particles between 1000
˚
tion temperature values (TR) are 583, 598, 604, and 613 K and 10000 A. Ni–MgO catalysts do not show significant
for the samples NiO, 4NA, 4CB, and 4CA, respectively. The differences with respect to their catalytic precursors.
initial reduction temperature of the NiO sample (583 K)
is lower than that of the systems prepared with magnesia
Temperature-Programmed Desorption (TPD)
whereas its reduction degree ( = 1) is higher (see Table 1).
This confirms that the interaction between NiO and MgO
To obtain information about their active sites, tem-
leads to the formation of a NiO–MgO solid solution for perature-programmed desorption of hydrogen was per-
these systems (27, 28). This hinders the reduction of NiO. formed for all the catalysts. The mechanisms of adsorption/
Comparing the three Ni–MgO catalysts, the catalyst which desorption of hydrogen are extremely complex (30–51), es-
has the highest reduction degree (R4NA) is the one whose pecially when dealing with supported catalysts. This is be-
precursor showed the lowest initial reduction temperature. cause phenomena related to the interaction between the

214
SERRA ET AL.
FIG. 2. Scanning electron micrograph of precursor 4CA.
active phase and the support can interfere although highly experimental conditions of the measurement such as the
complicated TPD spectra have also been obtained for pure weight of the sample examined, the flow rate of the carrier
metal catalysts (43). The number and population distribu- gas, the use of ultrahigh vacuum (UHV), or the shape of
tion of adsorbed species depend on many factors: how the the reactor system which affect conditions for removal of
catalyst was prepared, the kind of support used, and the desorbed hydrogen.
FIG. 3. Scanning electron micrograph of precursor 4CB.

Ni CATALYSTS FOR 1,4-BUTANEDINITRILE HYDROGENATION
215
TABLE 2
of hydrogen (due to a difference in morphology and size of
the nickel particles) that we also observed for other nickel
catalysts (50, 51).
Top Temperatures in the H2 TPD Spectra for Catalysts Ni-Oh,
R4CA, R4NA, and R4CB
With all catalysts hydrogen mainly desorbed at low tem-
peratures (<700 K). An important point to emphasize
is that H2 populations of the low-temperature peaks are
higher for the systems containing magnesia than for the
pure nickel catalyst (Ni-Oh). This may be related to the
higher metallic area obtained for the systems with magnesia
(see Table 1) and especially to the presence of specific active
sites. Therefore, there will be a greater amount of hydro-
gen available at the reaction temperatures for the Ni–MgO
catalysts.
In the reaction mechanisms proposed in the litera-
ture for the hydrogenation of dinitrile compounds (8),
different reaction products can be obtained if different
amounts of hydrogen are consumed. The production of
monoamine demands the lowest hydrogen consumption.
This fact together with the possibility to correlate the
binding strength of the chemisorbed hydrogen with the
activity for a specific reaction, as has been reported by
some authors (41, 50–52), lead to our interest in the
use of TPD results in the interpretation of the catalytic
activity.
Position of peak maxima
Catalysts
TD < 700 K
TD > 700 K
Ni-Oh
R4CA
R4NA
R4CB
476(w)
396(m)
435(m)
450(s)
545(w)
600(w)
687(s)
821(w)
—
811(w)
825(w)
634(m)
Note: s = strong; m = medium; w = weak.
Table 2 summarizes the hydrogen desorption temper-
ature of the peak maxima for catalysts Ni-Oh, R4CA,
R4NA, and R4CB. TPD plots show two peaks in the low-
temperature region (with maxima at 400–475 K and 545–
685 K, respectively) for all catalysts and one peak in the
high-temperature region (with maxima at 810–825 K) for
catalysts Ni-Oh, R4NA, and R4CB. The low-temperature re-
gion has been associated with different adsorption states of
the hydrogen for other nickel catalysts (41, 50, 51). Several
authors attribute the occurrence of the high-temperature
peaks to hydrogen spillover taking place during high-
temperature (above 773 K) treatments of the catalysts (30,
42). Their arguments cannot be applied to the Ni-Oh cata-
lyst which also shows a peak at 821 K. Therefore, a more
probable explanation, for all catalysts, is that the high-
temperature peak is associated with other adsorption states
Catalytic Activity
Table
3 shows conversions and selectivities for
hydrogenation of 1,4-butanedinitrile on nickel and
TABLE 3
Catalytic Behavior of Four Nickel Catalysts As Influenced by the Space Velocity and the Reaction Temperature
Selectivity (% )
Space
Reaction
temp
(K)
velocity
Conversion
(% )
Cracking Condensation
1
Catalyst
Ni-Oh
(h
)
4-Aminobutanenitrile 1,4-Butanediamine Pyrrolidine 1-Pyrroline products
products
6500
6500
6500
9750
13000
473
453
393
403
403
100
100
100
78
0
9
49
61
77
1
1
0
0
0
16
13
0
0
0
55
52
46
35
20
0
0
0
0
0
28
25
5
4
3
45
R4CA
R4NA
R4CB
6500
13000
13000
13000
13000
453
453
403
383
353
100
100
100
99
0
0
4
16
56
0
0
16
23
7
71
76
30
36
0
3
18
46
23
32
23
5
0
0
0
3
1
4
2
5
84
6500
6500
13000
13000
13000
453
433
403
383
363
100
100
100
100
93
0
0
23
60
68
0
0
17
18
0
36
59
7
0
0
0
20
53
20
28
64
21
0
0
0
0
0
0
0
0
6500
6500
6500
13000
13000
453
423
363
403
343
100
100
100
100
100
0
0
19
41
85
0
0
15
0
0
72
0
0
0
0
0
66
56
12
100
28
0
0
0
0
0
0
3
3
0

216
SERRA ET AL.
nickel-magnesia catalysts (Ni-Oh, R4CA, R4NA, R4CB) un- (453–343 K). This catalyst is the most active (100% conver-
der the conditions given in the Experimental section.
The activity of the two pure magnesia samples was also
tested and they were shown to be catalytically inactive un-
der the reaction conditions studied.
As Table 3 shows, the reaction products were 4-amino-
butanenitrile, 1,4-butanediamine, pyrrolidine, 1-pyrroline,
and cracking and condensation species. The selectivities to-
ward these reaction products depend on the catalyst used.
This dependence has been studied by optimizing the reac-
tion conditions (reaction temperature, space velocity, and
H2/dinitrile ratio).
High-molecular-weight products are obtained in higher
amounts when the catalyst is used without magnesia
(Ni-Oh) at a low space velocity (6500 h 1) and at high tem-
peratures (473–453 K). On the other hand, nickel–magnesia
catalysts hardly produce any condensation products and
consequently they do not undergo catalyst deactivation.
This is probably a consequence of the basic character of
magnesia which favors the elimination of amines and pre-
vents secondary reactions. These condensation products
have been described by Prins (6) as dimers obtained as a
result of an intermolecular amine–imine condensation re-
action. However, if these basic catalysts are used at low
space velocity (6500 h 1) and high temperatures (453–423
K), cracking products are obtained. The catalyst which pro-
vides most cracking products is the one with the highest
metallic area, R4CB (see Table 1), and the highest amount
of hydrogen desorbed at the reaction temperatures studied
sion) at all temperatures and space velocities used.
Formation of 4-aminobutanenitrile and 1,4-butane-
diamine. All the catalysts showed low selectivities toward
1,4-butanediamine. The Ni-Oh catalyst hardlyproducesany
of this diamine while the selectivity to the monoamine, 4-
aminobutanenitrile (77% ), is highest at the lowest conver-
1
sion value (45% ), working at a space velocity of 13,000 h
and a reaction temperature of 403 K.
However, the basic Ni–MgO catalysts produced low
amounts of the diamine (selectivity <24% ) and conversions
and selectivities toward 4-aminobutanenitrile were high.
The most active (100% ) and selective catalyst to 4-
aminobutanenitrile (85% ) was type R4CB. This result was
obtained when working at the same space velocity as with
the NiOh catalyst but at a lower reaction temperature
(343 K).
The main difference between the two types of systems is
their activity:the Ni–MgO catalystsare more active than the
Ni-Oh. This means that the nickel–magnesia systems can be
used at lower reaction temperatures in which cyclization
is not favored. This kind of catalysts makes it possible to
increase the selectivity to monoamine without a significant
loss of conversion.
The fact that the highest activity was found for catalyst
R4CB at 343 K is probably related to the presence of one
intense hydrogen TPD peak with a maximum at 450 K (see
Table 2). This provides enough desorbed hydrogen at this
FIG. 4. Product selectivities for the hydrogenation of 1,4-butanedinitrile on R4CB catalyst as a function of the hydrogen/1,4-butanedinitrile ratio.
(The modification of these ratios has been realized by diluting hydrogen with argon.)

Ni CATALYSTS FOR 1,4-BUTANEDINITRILE HYDROGENATION
217
reaction temperature for the monoamine to be selectively hydrogen available probably decreases, and consequently
obtained with high conversion values (100% ). the formation ofthe lesshydrogenated species, monoamine,
The Ni-Oh catalyst probably has a lower activity be- increases.
cause there is less hydrogen available at lower reaction
Pyrrolidine and 1-pyrroline are amine and imine cyclic
temperatures (catalyst Ni-Oh has a weak desorption peak compounds, respectively. The formation of the imine prob-
with a maximum at 476 K) and additionally it is partially ably needs lower amounts of hydrogen if both products
deactivated by the condensation products generated, as are obtained from the same intermediate compound (4-
mentioned above.
aminobutaneimine). In order to investigate how hydrogen
In a previous study (21), the selective hydrogenation of influences the production of pyrrolidine and 1-pyrroline,
one nitrile group was related to the presence of a consid- several experiments were performed with catalyst R4CB,
erable number of octahedral crystal sites. Our experiments varyingthe H2/1,4-butanedinitrile ratio and maintainingthe
have confirmed this since no diamine was observed when same total flow (by dilution with argon) at a temperature
the nickel catalyst (Ni-Oh) was used. In contrast, the Ni– of 423 K and a space velocity of 6500 h 1. The results are
MgO catalysts may provide diamine as a minor product shown in Fig. 4.
because the nickel particles do not show well-defined faces,
as observed by SEM and as confirmed by XRD.
When 100% H2 was used, the selectivity to pyrrolidine
was high (72% ); when hydrogen was diluted with 50% Ar,
the selectivity to 1-pyrroline clearly increased (69% ). Nev-
ertheless, a further decrease in the amount of hydrogen in
the totalflow(25% H2, 75% Ar) did not significantlychange
the selectivity to 1-pyrroline but there was an increase
in the selectivity to 4-aminobutanenitrile (18% ).
These results enabled us to propose Scheme 1 which
could explain the relative formation of 1-pyrroline, pyrro-
lidine, and 4-aminobutanenitrile as a function of the avail-
ability of hydrogen under the reaction conditions. Once the
aminoimine intermediate has been obtained, there are two
possible side reactions: a ring closure of the intermediate
with loss of ammonia and uptake of one H2 and a ring clo-
sure with the loss of ammonia only. The second side re-
action is probably only possible if the proportion of hy-
drogen in the total flow is decreased. It is of considerable
importance that 4-aminobutanenitrile appeared at the low-
est H2/Ar ratio amount (see Fig. 4). Hence, when less hy-
drogen is available, a higher selectivity to 1-pyrroline will
be found. Finally, under more diluted conditions more 4-
aminobutanenitrile will be formed, even at high reaction
temperatures.
Formation of pyrrolidine and 1-pyrroline. The Ni-Oh
catalyst shows low selectivities to pyrrolidine. The highest
value (16% ) was obtained by working at 473 K and a space
velocity of 6500 h 1 at 100% conversion. All Ni–MgO sys-
tems provide higher selectivities to pyrrolidine than the Ni-
Oh catalyst. Catalyst R4CA has a 76% selectivity at 453 K
and 13,000 h 1 and catalyst R4CB has a 72% selectivity at
423 K and 6500 h 1 for a 100% of conversion for both cata-
lysts.
The Ni-Oh catalyst shows the highest selectivity to 1-
pyrroline (55% ) at the highest reaction temperature tested
1
(473 K) and at a space velocity of 6500 h for a 100%
conversion. However, catalysts containing magnesia show
the highest 1-pyrroline selectivities between 46–66% at a
conversion of 100% at lower reaction temperatures (363
and 403 K). The most active catalyst, R4CB, gave the high-
est selectivity to 1-pyrroline, 66% at 363 K, whereas R4CA
and R4NA, which are less active, gave the highest val-
ues to 1-pyrroline (46% and 53% , respectively) at 403 K.
For these catalysts, at lower temperature the amount of
Scheme 1. Pyrrolidine and 1-pyrroline obtained by catalytic hydrogenation of 1,4-butanedinitrile (succinonitrile).

218
SERRA ET AL.
CONCLUSIONS
selectively. If the reaction temperature is decreased and/or
the hydrogen/adiponitrile ratio is decreased to a large ex-
tent, the monoamine selectivity increases.
Several nickel–magnesia catalysts have been prepared by
varying the magnesia source and the preparative method.
Their performance has been compared with the behavior
of a nickel catalyst of octahedral morphology in the hy-
drogenation of 1,4-butanedinitrile. All samples were struc-
turally characterized by the application of BET, XRD, TPR,
SEM, H2 chemisorption, and TPD techniques. XRD was
used to identify the solid solutions obtained for all the
NiO–MgO catalyst precursors and the catalysts. The Ni-Oh
catalyst contained nickel-only crystalline phase and all the
Ni–MgO catalysts consisted of Ni and a NiO–MgO solid so-
lution. XRD combined with TPR enabled usto establish the
sequence of reducibility of the catalyst precursors: NiO >
4NA > 4CB > 4CA.
SEM showed homogeneous octahedral crystallites for
the Ni-Oh catalyst, whereas the micrographs obtained for
the systems containing magnesia showed nickel particles
with non-well-defined faces, the homogeneity and sizes of
which depended on the preparative path used. The NiO–
MgO prepared by path B gave rise to a homogeneous sys-
REFERENCES
1. Weissermel, K., and Arpe, H. J., “Industrial Organic Chemistry.”
Verlag Chemie, Berlin, 1978.
2. Volf, J., and Pasek, J., in “Catalytic Hydrogenation, Sudies in
Surface Science and Catalysis” (L. Cerveny´, Ed.), Vol. 27. Elsevier,
Amsterdam, 1986.
3. Lazaris, A. Y., Zilberman, E. N., Lunicheva, E. V., and Vedin, A. M.,
Zh. Prikl. Khim. 38, 1097 (1965).
4. Bolle, J., French Patent 2 063378, 1971.
5. Medina, F., Salagre, P., Sueiras, J. E., and Fierro, J. L. G., J. Mol. Catal.
68, L17 (1991).
6. Prins, R., Catal. Today 37, 103 (1997).
7. Harrison, C. R., in “Catalysis and Chemical Processes” (R. Pearce and
W. R. Patterson, Eds.), Chapter 7. 1981.
8. Mares, F., Galle, J. E., Diamond, S. E., and Regina, F. J., J. Catal. 112,
145 (1988).
9. Huang, Y., and Sachtler, W. M. H., Appl. Catal. A: Gen. 182, 365 (1999).
10. Augustine, R. L., Catal. Rev. 13, 285 (1976).
11. Frank, G., and Neubauer, G., German Patent 3402734A1, 1984(BASF
A.G.).
˚
tem with particle sizes of around 1000 A.
12. Greenfield, H., Ind. Eng. Chem. Prod. Res. Dev. 6, 142 (1967).
The Ni–MgO catalysts showed higher activities than the 13. Pasek, J., Kostova, N., and Dvorak, B., Collect. Czech. Chem. Commun.
46, 1011 (1981).
Ni-Oh catalyst. This may be due to their higher metal
surface areas and hence the greater amount of available
chemisorbed hydrogen (H2 TPD results). R4CB was the
14. Schowoegler, E. J., and Adkins, H., J. Am. Chem. Soc. 61, 3449 (1939).
15. Medina, F., Salagre, P., Sueiras, J. E., and Fierro, J. L. G., Appl. Catal.
A: Gen. 92, 131 (1992).
most active catalyst under all the reaction conditions stud-
ied. This is because its metal surface area was the highest
(4.1 m2 g 1) as was its population of weakly bound H2.
In the nickel–magnesia catalysts, the presence of MgO
suppresses the condensation reactions and consequently
there is no catalyst deactivation, probably due to the ba-
sic character of magnesia which favors the elimination of
amines and prevents secondary reactions.
16. Medina, F., Salagre, P., Sueiras, J. E., and Fierro, J. L. G., Solid State
Ionics 59, 205 (1993).
17. Medina, F., Salagre, P., Sueiras, J. E., and Fierro, J. L. G., Appl. Catal.
A: Gen. 99, 115 (1993).
18. Medina, F., Salagre, P., Sueiras, J. E., and Fierro, J. L. G., J. Chem. Soc.,
Faraday Trans. 90(10), 1455 (1994).
19. Medina, F., Dutartre, R., Tichit, D., Coq, B., Dung, N. T., Salagre, P.,
and Sueiras, J. E., J. Mol. Catal. A: Chem. 119, 201 (1997).
20. Dung, N. T., Tichit, D., Chiche, B. H., and Coq, B., Appl. Catal. A: Gen.
169, 179 (1998).
The selectivity toward cyclic products (pyrrolidine and
1-pyrroline) can be modified by controlling the space veloc-
ity, the reaction temperature, and the hydrogen/adiponitrile
21. Medina, F., Salagre, P., Sueiras, J. E., and Fierro, J. L. G., J. Chem. Soc.,
Faraday Trans. 89(21), 3981 (1993).
22. Yu, X., Li, H., and Deng, J. F., Appl. Catal. A: Gen. 199, 191 (2000).
ratio. High temperatures and low space velocities were re- 23. Serra, M., Salagre, P., Cesteros, Y., Medina, F., and Sueiras, J. E., Solid
State Ionics 134, 229 (2000).
24. Estelle´, J., Ph.D. thesis, Rovira i Virgili University, Tarragona, Spain,
quired in order to obtain pyrrolidine. A decrease in the
hydrogen/1,4-butanedinitrile ratio involves an increase in
the selectivity to 1-pyrroline.
1998.
25. Bish, D. L., and Howard, S. A., J. Appl. Crystallogr. 21, 86 (1988).
Diamine was absent in the product stream when the Ni-
Oh catalyst was used but it was detected as a minor product
when the Ni–MgO catalysts were used. This may be due
to the higher crystallinity of the nickel phase in the Ni-Oh
catalyst.
26. Arena, F., Frusteri, F., Parmaliana, A., Plyasova, L., and Shmakov,
A. N., J. Chem. Soc., Faraday Trans. 92, 469 (1996).
27. Parmaliana, A., Arena, F., Frusteri, F., and Giordano, N., J. Chem. Soc.,
Faraday Trans. 86(14), 2663 (1990).
28. Arena, F., Frusteri, F., and Parmaliana, A., Appl. Catal. A: Gen. 187,
127 (1999).
Catalyst R4CB shows the highest selectivity to 4-
aminobutanenitrile (85% ). This has to be ascribed to the
characteristic surface properties of this catalyst, leading to
a suficient coverage with weakly bound hydrogen which
permits the choice of lower reaction temperatures. At
these lower temperatures, cyclization is not favored and the
29. Anderson, P. J., and Hohrlock, R. F., Trans. Faraday Soc. 58, 1993
(1962).
30. Pa´al, Z., and Menon, P. G., Catal. Rev.-Sci. Eng. 25(2), 229 (1983).
31. “Hydrogen Effects in Catalysis” Z. Paa´l and P. G. Menon, Eds.).
Marcel Dekker, New York, 1988.
32. Falconer, J. L., and Schwarz, J. A., Catal. Rev.-Sci. Eng. 25, 141 (1983).
33. Raupp, G. B., and Dumesic, J. A., J. Catal. 95, 587 (1985).
amount of hydrogen available could produce monoamine 34. Raupp, G. B., and Dumesic, J. A., J. Catal. 97, 85 (1986).

Ni CATALYSTS FOR 1,4-BUTANEDINITRILE HYDROGENATION
219
35. Arai, M., Nishiyama, Y., Masuda, T., and Hashimoto, K., Appl. Surf.
Sci. 89, 11 (1995).
36. Ferreira-Aparicio, P., Guerrero-Ruiz, A., and Rodriguez-Ramos, I.,
J. Chem. Soc., Faraday Trans. 93(19), 3563 (1997).
37. Popova, N. M., Babenkova, L. V., and Sokol’skii, D. V., Kinet. Katal.
10(5), 1177 (1969).
38. Bartholomew, C. H., Catal. Lett. 7, 27 (1990).
39. Lee, P. I., and Schwartz, J. A., J. Catal. 73, 272 (1982).
40. Zielinski, J., Polish J. Chem. 69, 1187 (1995).
41. Smeds, S., Salmi, T., Lindfors, L. P., and Krause, O., Appl. Catal. A:
Gen. 144, 177 (1996).
42. Kramer, P., and Andre, M., J. Catal. 58, 287 (1979).
43. Konvalinka, J. A., Van Oeffelt, P. H., and Scholten, J. J. F., Appl. Catal.
1, 141 (1981).
44. Weatherbee, G. D., and Bartholomew, C. H., J. Catal. 87, 55 (1984).
45. Spinicci, R., and Tofanari, A., React. Kinet. Catal. Lett. 27, 65 (1985).
46. Ikushima, Y., Arai, M., and Nishiyama, Y., Appl. Catal. 11, 305 (1984).
47. Glugla, P. G., Bailey, K. M., and Falconer, J. L., J. Catal. 115, 24
(1989).
48. Rankin, J. L., and Bartholomew, C. H., J. Catal. 100, 533 (1986).
49. Stockwell, D. M., Bertucco, A., Coulston, G. W., and Bennet, C. O.,
J. Catal. 113, 317 (1988).
50. Cesteros, Y., Salagre, P., Medina, F., and Sueiras, J. E., Appl. Catal. B:
Environ. 25, 213 (2000).
51. Cesteros, Y., Salagre, P., Medina, F., and Sueiras, J. E., Appl. Catal. B:
Environ. 22, 135 (1999).
52. Mare´cot, P., Paraiso, E., Dumas, J. M., and Barbier, J., Appl. Catal. 74,
261 (1991).