Polyamide synthesis from 6-aminocapronitrile, part 2: Heterogeneously catalyzed nitrile hydrolysis with consecutive amine amidation

Article abstract of DOI:10.1002/chem.200601898

To test the potential of heterogeneous catalysts for the nylon-6 synthesis from 6-aminocapronitrile, a number of zeolites, aluminum silicate, and metal oxides were tested as catalysts for the model reaction of pentanenitrile with water and hexylamine to N-hexylpentanamide. All zeolitic and aluminum silicate systems showed an insufficient performance, while the metal oxides (TiO ₂, ZrO₂, Nb₂O₅) showed very promising results. The kinetic behavior of the metal oxides was further investigated. First the nitrile was catalytically hydrolyzed to the terminal amide and subsequently the amidation of the hexylamine occurred. To polymerize 6-aminocapronitrile into nylon-6, more than 99% nitrile conversion was required to obtain a high-molecular-weight polymer. Pentanenitrile conversions larger than 99% can be obtained within six hours, at 230°C, by using ZrO ₂ as the catalyst. A kinetic study (by using IR spectroscopy) on the behavior of the metal oxides demonstrated that the adsorbed nitrile was catalytically hydrolyzed at the surface, but remained tightly bound to the surface. Zirconia-catalyzed polymerizations of 6-amino-capronitrile demonstrated that high-molecular-weight nylon-6 is feasible by using this route.

Full text of DOI:10.1002/chem.200601898

FULL PAPER
DOI: 10.1002/chem.200601898
Polyamide Synthesis from 6-Aminocapronitrile, Part 2: Heterogeneously
Catalyzed Nitrile Hydrolysis with Consecutive Amine Amidation
Adrianus J. M. van Dijk,[a, b] Robbert Duchateau,*[a, b] Emiel J. M. Hensen,[c]
Jan Meuldijk,[b, d] and Cor E. Koning[a, b]
Abstract: To test the potential of het-
erogeneous catalysts for the nylon-6
further investigated. First the nitrile
was catalytically hydrolyzed to the ter-
minal amide and subsequently the ami-
dation of the hexylamine occurred. To
polymerize 6-aminocapronitrile into
nylon-6, more than 99% nitrile conver-
sion was required to obtain a high-mo-
lecular-weight polymer. Pentanenitrile
conversions larger than 99% can be
obtained within six hours, at 2308C, by
using ZrO2 as the catalyst. A kinetic
study (by using IR spectroscopy) on
the behavior of the metal oxides dem-
onstrated that the adsorbed nitrile was
catalytically hydrolyzed at the surface,
but remained tightly bound to the sur-
ACHTREUNG
synthesis from 6-aminocapronitrile, a
number of zeolites, aluminum silicate,
and metal oxides were tested as cata-
lysts for the model reaction of pentane-
nitrile with water and hexylamine to
N-hexylpentanamide. All zeolitic and
aluminum silicate systems showed an
insufficient performance, while the
metal oxides (TiO2, ZrO2, Nb2O5)
showed very promising results. The ki-
netic behavior of the metal oxides was
face.
Zirconia-catalyzed
polymeri-
zations of 6-amino-capronitrile demon-
strated that high-molecular-weight
nylon-6 is feasible by using this route.
Keywords: amides · heterogeneous
catalysis · hydrolysis · metal oxides ·
nitriles · zirconium
Introduction
nylon-6 is the requirement of complete conversion of the ni-
trile functionality into an amide functionality. Unreacted ni-
triles or, for example, imine side products will function as
chain stoppers that will dramatically affect the polymer mo-
lecular weight. Furthermore, side reactions such as dispro-
portionation of the primary amine are also very undesirable
as they will lead to branching of the polymer. In an attempt
to reach the requirement of complete nitrile conversion, we
recently reported a comparative study of two methods to
improve the performance of the nitrile hydrolysis.[2] One
From an economic point of view, 6-aminocapronitrile
(ACN) is an attractive alternative for e-caprolactam as a
monomer for nylon-6 formation (Scheme 1).[1] ACN can be
prepared starting from butadiene by consecutive hydrocya-
nation and hydrogenation. The bottleneck for the polymeri-
zation of 6-aminocapronitrile into high-molecular-weight
[a] Dr. A. J. M. van Dijk, Dr. R. Duchateau, Prof. Dr. C. E. Koning
Laboratory of Polymer Chemistry
Eindhoven University of Technology
PO Box 513, 5600 MB Eindhoven (The Netherlands)
Fax : (+31)40-246-3966
E-mail: r.duchateau@tue.nl
[b] Dr. A. J. M. van Dijk, Dr. R. Duchateau, Dr. J. Meuldijk,
Prof. Dr. C. E. Koning
Dutch Polymer Institute
PO Box 902, 5600 AX Eindhoven (The Netherlands)
[c] Dr. E. J. M. Hensen
Laboratory of Inorganic Chemistry and Catalysis
Eindhoven University of Technology
PO Box 513, 5600 MB Eindhoven (The Netherlands)
[d] Dr. J. Meuldijk
Process Development Group
Eindhoven University of Technology
PO Box 513, 5600 MB Eindhoven (The Netherlands)
Scheme 1. Essential chemistry of the nylon-6 synthesis from 6-aminocap-
ronitrile.
Chem. Eur. J. 2007, 13, 7673 – 7681
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7673

method comprised the use of a mild hydrothermal process
and the other method was based on the use of a homogene-
ous ruthenium catalyst. Although high conversions were ob-
tained, neither of the systems resulted in complete nitrile
hydrolysis. Nevertheless, the study yielded insight into the
mechanism of the nitrile hydrolysis reaction, which is benefi-
cial for studying nitrile hydrolysis in the presence of hetero-
geneous catalysts.
Changing from homogeneous to heterogeneous catalysis
would be very advantageous. It dramatically simplifies the
process operation and allows a high catalyst to reactant
ratio that might reduce the reaction time of the hydrolysis
step.
Except for patent literature on the nylon-6 synthesis out
of ACN, there has not been much attention for the hetero-
geneously catalyzed formation of N-substituted amides from
nitrile and amine functionalities (Scheme 1). The only rele-
vant study available in the open literature deals with the
heterogeneously catalyzed nitrile hydrolysis and subsequent
esterification by using benzonitrile, water, and ethanol as
substrates and several zeolites as catalysts.[3a] Contrary to
the heterogeneously catalyzed formation of N-substituted
amides by using nitrile hydrolysis, the use of heterogeneous
catalysis for nitrile conversion into its corresponding termi-
nal amide is receiving more and more attention.[4] The best
heterogeneous systems for this reaction described so far are
ruthenium hydroxide on alumina[4a] and magnesia-supported
copper systems.[4b] Although these catalysts also seem very
attractive for the polymerization of 6-aminocapronitrile, the
presence of the amine functionality in this system dramati-
cally affects the behavior of the acidic catalysts. For exam-
ple, the abovementioned ruthenium on alumina fails com-
pletely due to amine-induced leaching of the ruthenium
from the acidic Al2O3 surface.[5] During the course of our
study, a BASF patent appeared describing the polymeri-
zation of 6-aminocapronitrile by using TiO2 as the heteroge-
neous catalyst.[1e,h] Unfortunately, nothing was disclosed
about the mechanisms involved and information about the
characteristics of the catalyst and catalyst activity was very
limited.
use of a model system also facilitates the mechanistic stud-
ies. Clearly, it is desirable to remove NH3 from the reaction
mixture as its presence is expected to limit the conversion
by chemical equilibrium.[2] However, for practical reasons
the reactions were performed in a closed system. The results
from the model reaction, obtained after 2 h at 2308C, are
Table 1. Performance of the different catalyst systems for the reaction
shown in Scheme 2.[a]
[b]
[c]
[d]
[e]
Entry
Catalyst
XPN
XHA
SPN
SHA
[%]
[%]
[%]
[%]
1
2
3
4
5
6
7
8
9
–
0
3
2
11
44
53
32
86
93
84
0
9
–
0
44
13
90
80
80
67
100
100
100
SiO2/Al2O3
zeolite HY
zeolite NaY
zeolite Hb
CuI–zeolite b
CuII–zeolite b
TiO2
>99
>99
99
99
99
15
10
54
65
46
79
84
78
99
>99
>99
>99
ZrO2
Nb2O5
10
[a] 2 h, 2308C, in a closed system, 0.29 g PN, 0.35 g HA, 0.13 g H2O and
0.2 g catalyst. [b] XPN =The conversion of pentanenitrile. [c] XHA =The
conversions of hexylamine. [d] SPN =The selectivity for the conversion of
pentanenitrile into pentanamide and N-hexylpentanamide. [e] SHA =The
selectivity for the conversion of hexylamine into N-hexylpentanamide.
collected in Table 1. The selectivity for the nitrile hydrolysis
was determined by the combined yield of the terminal
amide and N-hexylpentanamide that account for nearly all
the nitrile conversion.[2] The small deviation from complete
nitrile selectivity was caused by the formation of traces of
pentanoic acid. Nitrile hydrolysis was not observed for a
noncatalyzed reference experiment (entry 1, Table 1) and is
therefore completely attributed to the catalytic activity of
the oxide surfaces.
As zeolites show good performance for nitrile hydroly-
A
densation of alcohols,[3a] we started by testing some zeolitic
systems. Since the target application was the combined
hydrolysis and polymerization of 6-aminocapronitrile and
N
Herein we describe a study aimed at identifying heteroge-
neous catalysts that catalyze the consecutive nitrile hydroly-
sis and amine amidation. Furthermore, IR spectroscopy has
been used to investigate the three steps of the catalytic
cycle, being nitrile adsorption, nitrile hydrolysis on the oxide
surface, and desorption of the hydrolysis products from the
catalyst surface. The catalyst with the best performance in
the model reaction was also tested in preliminary polycon-
densation reactions.
partial oligomerization during the hydrolysis step could not
be avoided, the zeolites used were restricted to the large-
pore-size forms faujasite and beta zeolite. In addition to
zeolites, amorphous aluminum silicate was screened. Un-
fortunately, all silica-based systems (entries 2–5, Table 1) dis-
played a low activity towards nitrile hydrolysis, mostly in
combination with a poor selectivity for the amine amidation.
The low selectivity was mainly the result of dihexylamine
formation.[6] The disproportionation of hexylamine was a
posteriori not a big surprise for zeolitic systems. Note that,
among others, zeolites are industrially used to catalyze the
disproportionation of methylamine into dimethylamine and
Results and Discussion
trimethyl
amine.[7] Despite the fact that amorphous alumi-
A
To simplify the catalyst screening for the hydrolysis and sub-
sequent polymerization of 6-aminocapronitrile, a model
system of monofunctional reactants (n-pentanenitrile and n-
hexylamine) was investigated (Scheme 1). Furthermore, the
num silicate and sodium faujasite show much better selectiv-
ity with respect to the amine conversion into N-alkyl amide,
their activities are very low. The Hb zeolite (entry 5,
Table 1) has a distinctly higher activity than the faujasite
7674
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 7673 – 7681

Polyamide Synthesis
FULL PAPER
zeolite (entries 2 and 4, Table 1), but the observed conver-
sion is still far from sufficient. In an attempt to improve the
catalytic activity Hb zeolite was impregnated with copper
(entries 6 and 7, Table 1). Copper is known to catalyze the
nitrile hydrolysis.[3d,4d] Although both CuI- and CuII-modified
zeolites catalyze the model reaction, both systems showed
similar, or even worse nitrile conversion and selectivity than
Hb. Besides the zeolites, three metal oxides, ZrO2, TiO2,
and Nb2O5 (entries 8–10, Table 1), were tested. All three
metal oxides showed a high conversion of the nitrile and
showed excellent selectivity for the conversion of hexyl-
amine into N-hexylpentanamide. Owing to their promising
activity and selectivity, we decided to study these metal
oxide systems in more detail.
ZrO2 clearly showed the highest activity of the three
metal oxides (Table 1). After 2 h, at 2308C, already 93% of
nitrile had been converted against 86 and 84% for TiO2 and
Nb2O5, respectively. The conversion-time history and the
composition development of the metal oxide systems mea-
sured at a lower temperature (1808C) indicate that ZrO2 is
by far the most active catalyst with more than 50% nitrile
conversion within 2 h (Figure 1 and Figure 2). For all three
*
Figure 2. Selectivity development of pentanamide ( ) and N-hexyl
A
^
G
G
system with: a) ZrO2 and; b) Nb2O5, at 1808C.
of pentanamide and less than 1% of carboxylic acid were
formed as the only side products beside 95+ % of N-
hexylpentanamide. On the other hand, in the Nb2O5-cata-
G
lyzed reaction, pentanamide was initially the major product
(Figure 2b). Extrapolation of the pentanamide selectivity to
zero conversion according to the Delplot methodology iden-
tified the terminal amide as the initial product.[8] N-Hexyl-
pentanamide was the major species with increasing conver-
sion, comprising more than 90% of the converted nitrile
after 18 h. This implies that pentanenitrile was first hydro-
lyzed into pentanamide, which subsequently reacted with n-
hexylamine to give N-hexylpentanamide (Scheme 2).
Figure 1. Conversion–time history for the reaction of n-pentanenitrile
~
^
*
and n-hexylamine at 1808C with ZrO2 ( ), TiO2 ( ), and Nb2O5 ( ) in a
closed system.
Despite the initial absence of pentanamide in the ZrO2-
and TiO2-catalyzed systems, their reaction mixture develop-
ment can also be explained by Scheme 2. For the Nb2O5-cat-
catalyst systems pentanamide, N-hexylpentanamide, and
traces of pentanoic acid are the sole products.
The Group 4 metal oxides
(ZrO2 and TiO2) showed a very
similar composition develop-
ment, which surprisingly enough
is completely different with re-
spect to that obtained for
Nb2O5. For ZrO2 and TiO2 N-
hexylpentanamide was the main
product from the start of the re-
action. Only for nitrile conver-
sions higher than 50%, a small
amount of the terminal amide
was observed (Figure 2a). For a
nitrile conversion of 85%, 4%
Scheme 2. Reaction sequence for the formation of N-hexylpentanamide from n-pentanenitrile and n-hexyl-
amine. All reactions are catalyzed by metal oxides.
Chem. Eur. J. 2007, 13, 7673 – 7681
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7675

R. Duchateau et al.
nia-catalyzed system, a significant amount (up to 19%) of
pentanoic acid is present at lower conversions than 90%,
which is gradually consumed with increasing conversion. In
the niobium-oxide-catalyzed amidation reaction only a small
amount of pentanoic acid was present and for the uncata-
lyzed amidation, no pentanoic acid was observed at all. This
difference in pentanoic acid concentration is a strong indica-
tion that the difference between the two catalysts is that zir-
conia effectively catalyzes the hydrolysis of pentanamide to
pentanoic acid whereas niobia only effectively catalyzes the
nitrile hydrolysis to the amide.[10] Indeed, the zirconia-cata-
lyzed hydrolysis of pentanamide in the absence of hexyl-
amine afforded 30% of pentanoic acid after 15 h at 1408C,
while pentanamide was hardly hydrolyzed without ZrO2.[2]
Although the subsequent amine acylation with a carboxylic
acid was significantly faster than direct amine amidation and
proceeded satisfactorily without the addition of a catalyst,
also this reaction was found to be catalyzed by hydrated zir-
conia.[11] Therefore route IIbin Scheme 2 is clearly more im-
portant than route IIa and both steps of route IIb are cata-
lyzed by zirconia. Niobia is significantly less efficient in hy-
drolyzing the pentanamide into pentanoic acid compared to
U
zirconia and, at this point, it is not clear whether or not
niobia catalyzed the subsequent amine acylation with penta-
noic acid or whether the N-hexylpentanamide was formed
mainly through the direct amine amidation route IIa.
Figure 3. Conversion–time history (a) for the reaction of pentanamide
~
^
with hexylamine in the presence of water and ZrO2
( ) or Nb2O5 ( ) in
So far, ZrO2 turned out to be the most active and highly
selective catalyst for the combined nitrile hydrolysis and
amine amidation. Regardless of how promising this might
seem, the 93% nitrile conversion obtained within two hours,
at 2308C (see Table 1) is not sufficient for the target reac-
tion, being the conversion of 6-aminocapronitrile into a
high-molar-mass nylon-6. To find out whether the required
99+ % nitrile conversion can be obtained at all, the conver-
sion–time history of the model reaction using ZrO2 was fol-
lowed at 1808C, 2108C, and 2308C without NH3 removal
(Figure 4). At 2308C, thermodynamic equilibrium was
reached with a 99% nitrile conversion and with complete
selectivity of the amine conversion into N-hexylpentan-
a closed system. The uncatalyzed acylation reaction at 1408C (a) has
been included as a reference. The selectivity development with the con-
*
*
version (b) is shown for pentanoic acid ( : ZrO2 and : Nb2O5) and N-
^
^
hexylpentanamide ( : ZrO2 and : Nb2O5).
alyzed reactions, the nitrile hydrolysis and the amidation
proceed with time constants of the same order of magnitude.
So, no rate determining step could be assigned. For Group 4
metal oxides, however, the nitrile hydrolysis was found to be
rate limiting. These observations suggest that Group 4 metal
oxides catalyze the amine amidation better than Nb2O5. To
prove this hypothesis, the amidations of hexylamine with
A
pentanoic amide in an aqueous environment with zirconia
and niobia as catalysts were compared (route II, Scheme 2).
In the hope of tracing possible reaction intermediates, the
reaction was performed at 1408C instead of 1808C. As
shown in Figure 3, both Nb2O5 and ZrO2 increase the amide
conversion compared to the reaction without an oxide.
Taking into account that the reaction is performed in a
closed system, which unavoidably leads to an equilibrium
situation,[9] it is clear that ZrO2 catalyzes the hexylamine
amidation with pentanamide far more effectively than
Nb2O5. Whereas with ZrO2, most of the amide is already
converted within 1.5 hour at 1408C, the pentanamide con-
sumption is then only 17% for the Nb2O5 system and shows
a gradual increase over more than 6 h. For both catalyst sys-
tems, after about 10 h a thermodynamic equilibrium (Keq
ꢀ30, conversion ꢀ90%) is approximately reached with a
product mixture consisting of 91% N-hexylpentanamide,
7% pentanamide, and 2% of pentanoic acid. For the zirco-
C
7676
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 7673 – 7681

Polyamide Synthesis
FULL PAPER
ble, even in a closed system. Clearly, these excellent results
satisfy the requirement for the polymerization of 6-amino-
capronitrile (see below).
Aboulayt and co-workers for a dehydrated oxide surface.[12]
The molecular adsorption of CD3CN on a zirconia surface
gives rise to four adsorption bands in the IR spectrum, as
shown in Figure 5b. The bands at 2299 and 2262 cmÀ1 origi-
nate from the stretch vibration of the acetonitrile nitrile
group, n(CN), bound to Lewis acidic zirconia sites[12,13] and
physisorbed acetonitrile,[14] respectively. The symmetrical
At all temperatures, a high initial nitrile hydrolysis rate
was observed that decreased strongly at higher conversion.
The origin of the strongly decreasing nitrile hydrolysis rate
at higher conversions might be due to product inhibition, as
was also found in the homogeneous ruthenium catalyst
system.[2] An IR spectroscopic study was performed to get
more insight into the mechanism of the nitrile hydrolysis re-
action. Acetonitrile, often used as a probe molecule for the
surface site characterization of heterogeneous catalysts, was
used as the nitrile.
The first step of the catalytic cycle for nitrile hydrolysis is
believed to be the reversible adsorption of the nitrile onto
the zirconia surface (Figure 5a). The difference between the
hydroxide bands in vacuum (n1/n2) and under acetonitrile
pressure (p) indicate that the adsorption of acetonitrile is re-
versible and that the hydroxy groups on the surface are not
directly consumed upon nitrile adsorption on a hydrated sur-
face. This is in contrast to the observations reported by
stretch vibration, ns
A
bration, nas(CD3), of the methyl group are observed at 2113
H
and 2250 cmÀ1, respectively.[15] Both these methyl bands are
characteristic for physisorbed acetonitrile.
While still highly hydroxylated after calcination at 308C
for 120 min, the amount of zirconia-coordinated acetonitrile
increases more than 65% in comparison to the amount of
coordinated acetonitrile on a zirconia surface only activated
by evacuation for 10 min at 3008C, as shown in Figure 5b
R
trile is strongly dependent on the degree of hydration of the
surface, as an increased amount of physisorbed water is
competing with physisorbed acetonitrile for Lewis acid coor-
dination. Hence, a less hydrated surface results in a distinct-
ly higher amount of chemisorbed acetonitrile and conse-
quently affords a higher nitrile hydrolysis rate compared to
a more hydrated surface, in agreement with Figure 5b. Fur-
thermore, an increase of the acetonitrile pressure results ac-
cordingly in an increase of the absorption bands of both
chemi- and physisorbed acetonitrile (Figure 6). The increase
^
Figure 6. Adsorption isotherm for chemisorbed ( ) and physisorbed ace-
~
tonitrile ( ).
of the amount of coordinated acetonitrile can be rational-
ized by ligand displacement of zirconia-bound water as pro-
posed by Morterra and co-workers.[13]
The zirconia surface used for the catalytic nitrile hydroly-
sis will always be highly hydrated, as at least stoichiometric
amounts of water are required to completely convert the ni-
trile. As discussed above, this limits the amount of absorbed
acetonitrile activated for hydrolysis, and consequently de-
creases the hydrolysis rate. The surface reaction of adsorbed
acetonitrile at 308C is evidenced by a decrease of the inten-
Figure 5. a) IR spectra (4000–3000 cmÀ1) for the successive application of
vacuum (n1) and acetonitrile pressure (p), and vacuum (n2). b) IR spectra
(2350–2050 cmÀ1) after introduction of CD3CN on a zirconia surface cal-
cined at 308C for 10 min and during 120 min.
sity of the n(C
band at 2299 cmÀ1 during nitrile hydroly-
ꢁ
N)ads
Chem. Eur. J. 2007, 13, 7673 – 7681
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7677

R. Duchateau et al.
Figure 7. Effect of time (0.5 h increments) on the intensity of the nitrile
À1
signals at 2299 cmÀ1
(
~
) and 2262 cm
^
( ). Zirconia surface
n(C
ꢁ
N)ads
evacuated for 120 minutes. The solid line represents a first-order decay
of the n(C absorption band.
Figure 8. a) Effect of time (0.5 h time increments) on the intensity of the
absorption bands of the nitrile hydrolysis products and. b) After gradual
heating to 50 (d), 150 (b), and 1508C (c).
ꢁ
N)ads
sis (Figure 7a). Remarkably, the amount of physisorbed ni-
trile remains constant, while both the physisorbed and
chemisorbed acetonitrile are in equilibrium with gaseous
acetonitrile, according to the adsorption isotherm shown in
Figure 6. The change in chemisorbed acetonitrile coverage is
therefore ascribed to the consumption of surface sites due
to nitrile hydrolysis. Proton abstraction from acetonitrile,
observed for highly dehydroxylated metal oxides,[12,13,16] is
unlikely to occur at a highly hydroxylated surface. Indeed,
signals corresponding to [CD2CNÀ] and the formation of a
broad OD band at 2170 cmÀ1 are not observed.[12,16] Up to
about 60% conversion, the nitrile hydrolysis obeys first
order kinetics. At higher nitrile conversion the hydrolysis
rate starts to deviate from first-order kinetics (Figure 7b).
The bands resulting from the surface reaction of bound
acetonitrile are observed in the mid-IR range (1800–
1000 cmÀ1) as shown in Figure 8a.[12,17] An important obser-
vation is that although the adsorbed acetonitrile is hydro-
lyzed on a zirconia surface, most of the products are only
converted into acetamide anions. Only the signal at
1651 cmÀ1 can be assigned to molecularly adsorbed acet-
amide, which shows a good resemblance to the bands of
acetamide absorbed on TiO2.[17a] The other bands show a
strong similarity with the bands in spectra reported by
Aboulayt and co-workers. These authors assigned the bands
a and b to two kinds of anion acetamide species, resulting
from the nitrile hydrolysis on a zirconia surface.[12] Even
though an excess of molecular water is present, acetamide
ions are not fully converted into molecular amide, which ob-
structs the completion of the nitrile hydrolysis. This might
suggest that hydrolysis and desorption of the surface-bound
acetamide species are both governing the rate of the reac-
tion. So product inhibition indeed plays a crucial role in the
catalytic nitrile hydrolysis process.
Figure 8bshows the evolution of the infrared bands of the
hydrolysis products upon stepwise heating. The gradual
change in surface species is a result of the hydrolysis of the
acetamide surface ions into surface acetate species.[12,17c] De-
spite the fact that there are still surface hydroxyls present to
complete the catalytic cycle, the ionic products remain
strongly bound to the zirconia surface. The ionic products
are not readily transformed into molecular acetamide or
acetic acid and will simply not desorbupon heating. The
presence of the primary amine might facilitate the comple-
tion of the catalytic cycle of the nitrile hydrolysis.
7678
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 7673 – 7681

Polyamide Synthesis
FULL PAPER
Polymerization of 6-aminocapronitrile (ACN): To study
whether the results obtained for the model system are also
valid for making high-molecular-weight nylon-6, we per-
formed the zirconium-catalyzed polymerization of 6-amino-
capronitrile. Before synthesizing nylon-6 out of ACN by
using ZrO2, the uncatalyzed synthesis was performed as a
reference first. Both time and amount of water were varied,
as shown in Table 2.
actions demonstrated that when the nitrile hydrolysis is per-
formed in a closed system, distinct amounts of amide func-
tionalized oligomers can be expected. These are expected to
have a lower amidation rate with the amine functionality
than a carboxylic acid end-group, and could therefore be re-
sponsible for the limited molar mass built up. However, by
performing the polycondensation of adipamide with 1,6-hex-
anediamine it was found that amide-functionalized mono-
mers lower the polycondensation rate but do not affect the
final molecular weight. The molecular weight is probably
limited due to the presence of water required during the
ACN hydrolysis. A successive further amidation under
water-free conditions is therefore required to achieve the
maximum molecular weight. Since the amount of polymer
product obtained from the experiment (Table 3, entry 1) was
too low to perform a controlled-melt polymerization, the
semicrystalline product has been submitted to a solid-state
post-condensation (SSP) step (Table 3, entry 2). Whereas
the post-condensation only results in a minor increase of the
number average molecular weight (Mn), the polydispersity is
distinctively higher, and a weight average molar mass (Mw)
of approximately 12000 gmolÀ1 is reached. This value is al-
ready close to that of commercial low-Mw-grade nylons. The
amidation during solid-state post-condensation was per-
formed at 1858C, which is well below the melting tempera-
ture of the nylon-6 prepolymer. At this temperature, the
molecules present in the crystals do not participate in the
post-condensation reaction. Only the chain ends present in
the amorphous phase will undergo chain extension. This ex-
plains the broad PDI obtained after the SSP. Although not
satisfactory yet, this build-up of molecular weight during the
SSP step at 1858C clearly shows the potential of increasing
the molecular weight by post-condensation and it is very
likely that a melt-polymerization step will result in the re-
quired molecular weight.
Table 2. The molar mass distribution of the nylon-6 synthesized from 6-
aminocapronitrile without a catalyst, at 2308C.
Entry
ACN/H2O
t [h]
Mn [gmolÀ1
]
PDI
1
2
3
4
1:2
1:2
1:10
1:10
8.5
22.5
8.5
1200
2100
1100
1300
2.5
1.6
1.6
1.4
22.5
Although some oligomeric material was obtained, no
high-molecular-weight material was formed. The molecular
weight was largely independent of the amount of water used
and of the reaction time, although a smaller amount of
water initially resulted in a higher polydispersity (entry 1,
Table 2). After more than 22 hours with two equivalents of
water (entry 2, Table 2) the polydispersity index (PDI) was
similar to that obtained in the polymerizations with ten
equivalents of water, but the number average molecular
weight (Mn) is in this case about a factor of two higher. As
the amount of water used seems to affect the polymerization
of ACN, polymerization using the zirconia catalyst was also
performed with varying amounts of water (Table 3). A dis-
Table 3. The molar mass distribution of the nylon-6 synthesized from 6-
aminocapronitrile, by using ZrO2 as the catalyst.[a]
Entry
ACN/H2O
Mn [gmolÀ1
]
PDI
1
1:2
1:2
1:4
1:8
1:10
4100
4700
4000
4000
3900
1.4
2.6
1.4
1.5
1.4
2[b]
3
4
5
Conclusion
[a] 2308C, 8 h. [b] Nylon of entry 1 after 24 h post-condensation, at
1858C, in the solid state.
In an attempt to develop a catalytic system for the com-
bined hydrolysis and subsequent polymerization of 6-amino-
capronitrile to nylon-6, some zeolites, amorphous aluminum
silicate, and metal oxides were tested as catalysts for the
combined nitrile hydrolysis and amine amidation of n-pent-
tinct increase of the molecular weight was observed for all
experiments compared to the polymerization without zirco-
nia, demonstrating that zirconia indeed catalyzes the nitrile
hydrolysis of ACN, as was earlier observed for the studied
model compounds (see above). In contrast to the uncata-
lyzed ACN polymerization, the polymerization catalyzed by
ZrO2 does not depend on the amount of water used, at least
not within experimental error.
Although the use of the zirconia catalyst increases the
molar mass, the molar mass is still low with respect to the
required molar mass of about 1.5”104 gmolÀ1. The most
trivial origin for the formation of the low-molecular-weight
nylon is that the product obtained after the ACN hydrolysis
still contains some residual nitrile end-groups. The absence
of nitrile absorption bands in the IR spectrum of the prod-
uct excludes the presence of residual nitrile end-groups as
the cause of the moderate molecular weight. The model re-
A
as they gave low nitrile conversion. Furthermore, the ob-
served secondary amine formation is detrimental to the tar-
geted polymerization process, since secondary amines lead
to branched products and loss of stoichiometry. The metal
oxides gave high nitrile conversions and excellent selectivity.
Among the oxides zirconia was found to be by far the best
catalyst system. With a nitrile conversion above 99%, ZrO2
is a truly interesting catalyst for the combined hydrolysis
and polymerization of 6-aminocapronitrile. The composition
development of the reaction mixture shows that the nitrile is
first hydrolyzed into the amide that either directly, or after
further hydrolysis to the corresponding carboxylic acid, acy-
Chem. Eur. J. 2007, 13, 7673 – 7681
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7679

R. Duchateau et al.
SEM images indicated an average particle size of 5–10 mm. The BET sur-
face area was 288 m2 gÀ1 and the pore diameter was 4.9 nm. Niobium
oxide (HY 340) was acquired from CBMM Ltd. (Araxµ, Brazil) and was
used after calcination at 2008C in an air stream for 6 h. The BET surface
area was 193 m2 gÀ1 and the pore diameter was 4.3 nm. The specific sur-
face area (BET) and average pore size of the oxides were determined on
a Micromeretics ASAP 2020 instrument. To desorbcontaminating mole-
cules (mainly water) from the catalyst surface, the catalyst was first pre-
treated at 2008C, under vacuum, for over 2 h. For the determination of
the BET surface area, values of p/p0 in the range 0<p/p0 ꢂ0.3 was used.
For the pore size measurements, the value of p/p0 was further increased
to 1 and subsequently reduced to 0.14. Pentanenitrile (99.5%, Aldrich),
hexylamine (99%, Merck), pentanamide (99%, Acros), and p-xylene
(99+ %, Aldrich) were used as received. 6-Aminocapronitrile was kindly
donated by BASF AG The reaction products were identified by using
1H NMR spectroscopy (Varian Mercury 400). The product mixtures were
analyzed by gas chromatography by using a CP 9000 gas chromatograph
(GC) equipped with a 30 m”0.32 mm id, capillary CP volamine column,
and a FID detector. p-Xylene was used as an internal standard. Samples
were injected at 1208C and after a stabilization time of 1 min the temper-
ature was raised with 158CminÀ1 to 2908C. All single-point experiments
were performed in duplicate.
lates the amine to the corresponding N-hexylpentanamide.
Group 4 metal oxides not only catalyze the nitrile hydrolysis
but also prove to be highly active catalysts for the amine
acylation. Consequently, whereas the Nb2O5 catalyzed nitrile
hydrolysis and amine acylation proceeds with time constants
of similar order of magnitude, for zirconia and titania, the
nitrile hydrolysis is clearly rate determining.
IR spectroscopy clearly demonstrates that the catalytic
cycle starts with reversible nitrile adsorption, followed by
hydrolysis. The so formed surface-bound amide species are
further hydrolyzed to surface-bound carboxylates instead of
being protonated to form molecular amides. In fact both
amides and acetates are strongly bonded to the metal oxide
surface. The strong binding of the amides and carboxylates
to the metal oxide predicts product inhibition and by that a
drop in catalytic activity of the metal oxides at high nitrile
conversions. However, the rate of N-hexylpentanamide pro-
duction is strongly increased by the metal oxides, implying
that the completion of the catalytic cycle is strongly im-
proved by the presence of an amine. Tuning the amount of
water present in the system is also crucial, as a low water
concentration is beneficial for nitrile adsorption but results
in low rates of hydrolysis and product desorption. Converse-
ly, a high water concentration increases the rate of nitrile
hydrolysis and desorption of the hydrolysis products but
limits the nitrile adsorption.
ZrO2 is a very promising catalyst for the complete and se-
lective hydrolysis of 6-aminocapronitrile into nylon-6. In
comparison with the uncatalyzed hydrolysis of ACN, the use
of the ZrO2 catalyst significantly improves the molar mass
build-up of the product nylon-6. However, presence of
water required for the hydrolysis prevents build-up of the
molecular weight. Preliminary results by using an SSP step
clearly demonstrate that water removal leads to the desired
build-up of molecular weight.
Nitrile hydrolysis with subsequent amine amidation: The hydrolysis ex-
periments were performed in closed, stainless-steel 5 mL microreactors
by using a solvent-free system consisting of equimolar amounts of pent-
anenitrile and hexylamine in aqueous environment.[2] Reaction mixtures
consisting of n-pentanenitrile (8.6 mmol, 0.71 g), n-hexylamine
(8.6 mmol, 0.87 g), water (17.2 mmol, 0.31 g), and 0.2 g of the heterogene-
ous catalyst were prepared in an inert atmosphere. The internal standard
was p-xylene (3.56 mmol, 0.36 g). The rest of the experimental procedure
was identical to the procedure previously applied.[2] Reaction mixtures
were sufficiently stirred by using magnetic stirrer bars to avoid external
mass or heat transport limitations for the reactions on the catalyst sur-
face. For the analysis of the reaction mixture with GC, the product mix-
ture diluted with ethanol was filtered to remove the heterogeneous cata-
lysts. For the catalyst screening all the catalyst tests were performed in
duplicate. The results presented are the mean values of the conversion
and selectivity results.
IR spectroscopy: The zirconia used was obtained by calcination of a com-
mercially available precursor (MEL XZ01247/01, Mel Chemicals) at
308C, for 6 h, in an air stream. The zirconium oxide obtained was mainly
amorphous and exhibited a specific surface area of 288 m2 gÀ1. Infrared
spectra were recorded by using a Bruker IFS 113v spectrophotometer
equipped with a DTGS detector. The controlled-atmosphere transmission
cell was connected to a stainless-steel vacuum system, pumped by a tur-
bomolecular pump with a base pressure lower than 1”10À6 mbar. All IR
spectra were recorded at 308C with a resolution of 1 cmÀ1. The zirconia
material was pressed into self-supporting wafers (ꢀ9.5 mgcmÀ1). Before
the start of the surface studies, the zirconia was treated by dynamic
vacuum. Due to the largely irreversible nature of some of the surface re-
actions, virgin samples of (virtually) constant thickness were used for
each IR experiment. Deuterated acetonitrile was used as a reactant for
infrared studies. Due to the strong complication of the CN region for
acetonitrile by Fermi resonances between the n(CN) and the combination
Experimental Section
General: Zeolite NaY was obtained from Akzo Nobel and had a silica-
to-alumina ratio of 2.5 and a surface area of 500 m2 gÀ1. Zeolite HY was
synthesized through ion exchange of NaY with NH4NO3 followed by cal-
cination to deammonate the formed NH4Y into HY. Zeolite Hb was ob-
tained from Zeolist Int. and has a silica-to-alumina ratio of 37.5 and a
surface area of 680 m2 gÀ1. To avoid the formation of clustered copper
species during the synthesis of CuIb and CuIIb, the use of solid-state ion
exchange and a high Si/Al ratio for the zeolite was chosen.[18] The com-
plete ion-exchange procedure was performed in a nitrogen atmosphere.
Before ion exchange the zeolite was first calcined in situ. Subsequently,
the zeolite was well mixed with an equivalent amount to 2 wt% of CuICl
or CuIICl2 to obtain CuI/b and CuII/b loaded with 2 wt% of metal, respec-
tively. The obtained mixture was then heated for 2 h at 5008C in a nitro-
gen flow. In this way the hydrogen was replaced by copper and the HCl
formed was removed with the gas flow. Titanium oxide was prepared
from a commercially available precursor (FinnTi S150, acquired from
Kemira) by calcination at 3008C in an air stream for 6 h. SEM images in-
dicated an average particle size of 4–10 mm. The BET surface area was
201 m2 gÀ1 and the pore diameter was 5.8 nm. Zirconia was prepared
from a commercially available precursor (MEL XZ01247/01, acquired
from Mel Chemicals) by calcination at 3008C in an air stream for 6 h.
ds(CH3)+n(CC) frequencies, deuterated acetonitrile (CD3CN) was more
A
appropriate for infrared studies than CH3CN.[19] The CD3CN (liquid,
99.9%, Aldrich) was purified by several freeze-pump-thaw cycles before
use.
Nylon-6 synthesis from 6-aminocapronitrile: The polymerization of ACN
was performed by using the same system as applied in the catalytic nitrile
hydrolysis experiments.[2] A typical polymerization experiment was per-
formed by using ACN (0.90 g, 1 mL, 8.0 mmol, 1 equiv), water (0.29 g,
16 mmol, 2 equiv), and ZrO2 (0.40 g, Melchem, MEL XZ01247/01), at a
reaction temperature of 2308C. As the reactor set-up does not allow the
polymer–catalyst separation in the melt, the separation was achieved by
first dissolving the polymer product in formic acid. After removal of the
zirconia catalyst by filtration, the polymer was precipitated by addition of
water and separated from the liquid phase by ultracentrifugation for 1 h
7680
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 7673 – 7681

Polyamide Synthesis
FULL PAPER
at 40000 rpm. It should be noted that this method results in the loss of
the low-molecular-weight fraction of the product. The solid-state post-
condensation was performed by heating the finely ground polymer to
1858C, for 24 h, under a nitrogen flow, by using the method described in
literature.[20]
(BASF A.-G.), DE Patent 19753301, 1997; g) E. L. Martin (DuPont),
US Patent 2357484, 1944. The operating temperature is typically
near or above 3008C. However, under these conditions the uncata-
lyzed nitrile hydrolysis already proceeds with a decent rate, making
it difficult to measure the effect of the oxide. h) A. Kramer, S. Mit-
telstadt, H. Vogel, Chem. Eng. Technol. 1999, 22, 494–500. Further-
more, the process requires a highly dilute system to obtain the
lactam, which corresponds to a large excess of water in the process.
[4] a) K. Yamaguchi, M. Matsushita, N. Mizuno, Angew. Chem. 2004,
116, 1602–1606; Angew. Chem. Int. Ed. 2004, 43, 1576–1580; b) H.
Hayashi, H. Nishi, Y. Watanabe, T. Okazaki, J. Catal. 1981, 69, 44–
50; c) Y. Izumi, Catal. Today 1997, 33, 371–409; d) K. Mori, K. Ya-
maguchi, T. Mizugaki, K. Ebitani, K. Kaneda, Chem. Commun.
2001, 461–462; e) D. R. Millic, D. M. Opsenica, B. Adnadevic, B. A.
Solaja, Molecules 2000, 5, 118–126; f) T. Ushikubo, Y. Hara, K.
Wada, Catal. Today 1993, 16, 525–535; g) F. Nozaki, T. Sodesawa, T.
Yamamoto, J. Catal. 1983, 84, 267–269; h) K. K. Dzhumakaev, Y.
Isakov, G. T. Fedolyak, K. Minachev, T. A. Isakova, A. D. Kagarlit-
skii, Kinet. Katal. 1987, 28, 856–860.
Polymer analysis: The molecular weight distribution of the synthesized
nylon-6 was determined by using size-exclusion chromatography (SEC)
with hexafluoroisopropanol (HFIP) as eluent. The SEC set-up consecu-
tively consists of an eluent degasser (Alltech Elite), a gradient pump
(Shimadzu, LC-10 AD), an injector (Spark Holland, Midas), a two-
column set (PSS, PFG Linear xl 7 mm 8”300 mm) in series, and a differ-
ential refractive-index detector (DRI) (Waters, 2414). After injection of a
50-mL sample, the separation was established with
a flow rate of
0.8 mLminÀ1, at a constant temperature of 408C. The HFIP (Biosolve,
AR-S grade) was recycled and distilled when necessary. For evaluation of
the molecular weight distributions the DRI detector was used as a con-
centration detector. The calculated molecular weights were based on a
calibration curve for poly(methyl methacrylate) standards (molar mass
range 650–1.5”104 gmolÀ1) of narrow polydispersity (Polymer Laborato-
ries) in HFIP. The presence of nitrile functionalities in the polyamide
product was assessed with transmission IR spectroscopy (Biorad,
FTS 6000) on a thin polymer film, obtained by compression of a polymer
particle with a diamond press.
[5] Although ruthenium on Al2O3 is highly active for the selective ni-
trile hydrolysis (ref. [4a], with our system a 30% loss of amine se-
lectivity in the conversion of hexylamine into N-hexylpentanamide
is observed after 2 h, at 1608C. Furthermore, the presence of nitrile,
amine, and water results in leaching of the metal because a clear-
colored reaction mixture was observed, due to ruthenium in solu-
tion.
[6] S. Gomez, J. A. Peters, T. Maschmeyer, Adv. Synth. Catal. 2002, 344,
1037–1057.
[7] D. R. Corbin, S. Schwarz, G. C. Sonnichsen, Catal. Today 1997, 37,
71–102.
[8] N. A. Bhore, M. T. Klein, K. B. Bischoff, Ind. Eng. Chem. Res. 1990,
29, 313–316.
Acknowledgements
The authors thank the Dutch Polymer Institute (DPI) for their financial
support. This research forms part of the research program of the Dutch
Polymer Institute (DPI), project #297.
[9] a) B. Izzo, M. T. Klein, C. LaMarca, N. C. Scrivner, Ind. Eng. Chem.
Res. 1999, 38, 1183–1191; b) A. J. Belsky, T. B. Brill, J. Phys. Chem.
A 1999, 103, 3006–3012; c) B. Izzo, C. L. Harrell, M. T. Klein,
AIChE J. 1997, 43, 2048–2058.
[10] For the uncatalyzed amidation the only observed product is N-hex-
ylpentanamide.
[11] K. Takahashi, M. Shibagaki, H. Kuno, H. Kawakami, H. Matsushita,
Bull. Chem. Soc. Jpn. 1989, 62, 1333–1334.
[12] A. Aboulayt, C. Binet, J. C. Lavalley, J. Chem. Soc. Faraday Trans.
1995, 91, 2913–2920.
[13] a) C. Morterra, G. Cerrato, E. Novarino, M. P. Mentruit, Langmuir
2003, 19, 5708–5721; b) C. Morterra, M. Penarroya Mentruit, G.
Cerrato, Phys. Chem. Chem. Phys. 2002, 4, 676–687.
[14] R. E. Sempels, P. G. Rouxhet, J. ColloidInterface Sci. 1976, 55, 263–
273.
[15] a) A. G. Pelmenschikov, R. A. van Santen, J. Janchen, E. Meijer, J.
Phys. Chem. 1993, 97, 11071–11074; b) Vibrational Spectra of Poly-
atomic Molecules (Eds.: L. M. Sverdlov, M. A. Kovner, E. P. Krai-
nov), Jerusalem, Israel program for Scientific Translations, 1974.
[16] J. C. Lavalley, Catal. Today 1996, 27, 377–401.
[17] a) C. C. Chuang, W. C. Wu, M. X. Lee, J. L. Lin, Phys. Chem. Chem.
Phys. 2000, 2, 3877–3882; b) H. Knoezinger, H. Krietenbrink, H. D.
Mueller, W. Schulz, Proc. Int. Congr. Catal. 1977, 1, 183–194;
c) L. F. Liao, C. F. Lien, D. L. Shieh, F. C. Chen, J. L. Lin, Phys.
Chem. Chem. Phys. 2002, 4, 4584–4589.
[1] a) B. S. Curatolo, R. C. Sentman, G. P. Coffey (Standard Oil Co.),
US Patent 4629776, 1986; b) R. Mohrschladt, V. Hildebrandt (BASF
A.-G.), WO Patent 9943734, 1998; c) D. N. Marks (DuPont), EP
Patent 479306, 1991; d) E. K. Marchildon, A. T. Mutel (DuPont),
WO Patent 2000024808, 1998; e) R. Mohrschladt, H. Winterling, D.
Krauss (BASF A.-G.), DE Patent 19962573, 1995; f) R. Mohrs-
chladt, V. Hildebrandt (BASF A.-G.), WO Patent 9938905, 1998;
g) R. Mohrschladt, D. Krauss, V. Hildebrandt (BASF A.-G.), WO
Patent 9938906, 1998; h) R. Mohrschladt, V. Hildebrandt, E. Fuchs
(BASF A.-G.), WO Patent 9938908, 1998; i) J. D. Cohen, S. B. Fer-
gusson, E. K. Marchildon, D. N. Marks, A. Mutel (DuPont), US
Patent 6437089, 2001; j) C. H. Greenewalt (DuPont), US Patent
2245129, 1938; k) R. Weiss, D. Krauss, D. Keller, G. Pipper, W.
Harder, A. Ludwig, R. Mohrschladt (BASF A.-G.), WO Patent
9808889, 1997; l) H. Liehr (Zimmer A.-G.), US Patent 5627257,
1996; m) D. N. Marks (DuPont), US Patent 5185427, 1991; n) B. S.
Curatolo (Standard Oil Co.), US Patent 4640976, 1985; o) G. P.
Coffey (Standard Oil Co.), US Patent 4603192, 1985; p) B. S. Cura-
tolo (Standard Oil Co.), US Patent 4568736, 1984; q) B. S. Curatolo
(Standard Oil Co.), US Patent 4542205, 1984; r) R. Mohrschladt, H.
Winterling, D. Krauss (BASF A.-G.), DE Patent 19962573, 1996.
[2] Part 1 in this series: A. J. M. Van Dijk, T. Heyligen R. Duchateau J.
Meuldijk, C. E. Koning, Chem. Eur. J. 2007, 13, 10.1002/
chem.200601897).
[18] G. Spoto, A. Zecchina, S. Bordiga, G. Ricchiardi, G. Martra, G. Leo-
fanti, G. Petrini, Appl. Catal. B 1994, 3, 151–172.
[19] H. Knoezinger, H. Krietenbrink, J. Chem. Soc. Faraday Trans. 1975,
71, 2421–2430.
[20] M. A. G. Jansen, PhD Thesis, Eindhoven University of Technology
(The Netherlands), 2005.
[3] a) M. Lasperas, P. Graffin, P. Geneste, J. Catal. 1993, 139, 362–370.
A number of patents consider the gas-phase conversion of an ali-
phatic amino-nitrile into the corresponding N-substituted amide by
using metal oxides, mainly aimed at the synthesis of the correspond-
ing lactams. b) J.-P. Brunelle, C. Nedez (Rhodia), WO Patent
2000004994, 1999; c) H.-J. Buysch, O. Immel, R. Langer (Bayer
A.-G.), DE Patent 19632006, 1996; d) F. Mares (Allied Co.), EP
Patent 150295, 1984; e) M.-C. Cotting, L. Gilbert, N. Laurain, C.
Nedez (RhonÞ Poulenc), WO Patent 9622974, 1996; f) M. Eiermann
Received: December 30, 2006
Revised: March 6, 2007
Published online: June 26, 2007
Chem. Eur. J. 2007, 13, 7673 – 7681
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7681