Gas-phase adiponitrile hydrogenation over modified Ni-P/SiO2 amorphous catalysts

Article abstract of DOI:10.1246/cl.2000.1048

Gas-phase hydrogenation of adiponitrile was carried out in a fixed-bed reactor at 1 atm pressure, and in the absence of ammonia. In comparison with other Ni-based catalysts, the Ni-P/SiO₂ amorphous catalyst exhibited higher activity and/or better selectivity to 1,6-hexanediamine, which could be further improved by W or MgO-additives.

Full text of DOI:10.1246/cl.2000.1048

1
048
Chemistry Letters 2000
Gas-Phase Adiponitrile Hydrogenation over Modified Ni–P/SiO Amorphous Catalysts
2
Hexing Li,* Minghui Wang, and Yeping Xu
Department of Chemistry, Shanghai Normal University, Shanghai 200234, P. R. China
(Received May 29, 2000; CL-000517)
The XRD patterns9 revealed that the as-prepared
Gas-phase hydrogenation of adiponitrile was carried out in a
fixed-bed reactor at 1 atm pressure, and in the absence of ammo-
nia. In comparison with other Ni-based catalysts, the Ni–P/SiO2
amorphous catalyst exhibited higher activity and/or better selec-
tivity to 1,6-hexanediamine, which could be further improved by
W or MgO-additives.
Ni–P/SiO , Ni–W–P/SiO , and Ni–P/SiO –MgO samples were
2
2
2
present in the typical amorphous structure as found in the Ni–P
1
0
amorphous alloy obtained by the rapid quenching technique.
When these samples were treated at 873 K in N flow, the crys-
2
tallization occurred gradually since the amorphous structure is
metastable thermodynamically. The radial distribution function
(
RDF) curves, obtained by the fast Fourier transformation based
3
1
,6-Hexanediamine (HMA) is the most important intermedi-
on the χ(k)k Ni edge EXAFS data, further confirmed the amor-
phous structure of the as-prepared Ni–P based samples since
only one weak FT peak around R = 2.1 Å was observed for the
fresh samples while very strong FT peak around R = 2.1 Å and
two additional small peaks at longer distance (ca. 3.7–4.6 Å )
1
ate in the manufacturing of Nylon-6,6. Large scale HMA is usu-
ally produced via adiponitrile (ADN) liquid phase hydrogenation
and Raney Ni and Raney Co are probably the most frequently
used catalysts.2 Because these kinds of catalysts are subject to
severe leaching by the aqueous reaction medium, substantial
nickel losses and product cleanup are inevitable. The develop-
ment of a fixed-bed catalyst system for ADN hydrogenation with
a supported metal catalyst that was resistant for such leaching
was believed to be feasible. However, the selectivity to HMA
was very low (< 40%), where the catalyst was proved to be the
key.4 Recently, the amorphous alloy catalysts have caused
much attention for their higher activity, better selectivity, and
stronger sulfur resistance in hydrogenation reactions.7 In this
,3
9
were observed after being treated at 873 K, showing that the
fresh samples had no long-range but only short-range ordering
structure, which may transfer to the well-ordered crystalline
structure at high temperature.
The XPS spectra of the Ni–P/SiO , Ni–W–P/SiO and
2
2
Ni–P/SiO –MgO amorphous catalysts revealed that the nickel
2
–6
in all the as-prepared catalysts was mainly present in the metal-
lic state with the binding energy (BE) of 853.1 eV (Ni2P3/2
level) and the phosphorus was in the alloying state with Ni cor-
responding to 130.0 eV (P level). Most of the W species was
present in the W state with BE of 32.8 eV in W4f level. In
–9
paper, we report a Ni–P/SiO amorphous catalyst which seems
2
2P
4+
more powerful in the gas-phase ADN hydrogenation to HMA
due to its excellent activity and selectivity at 1 atm pressure and
even in the absence of ammonia. Promoting effects of W- and
MgO-additives were also observed.
The Ni–P/SiO amorphous catalyst was prepared by the elec-
troless plating described as follows: 1.0 g SiO (40–60 meshes, 182
1
1
comparison with the BE values of pure metallic Ni and red P,
no significant BE shift of both the metallic Ni and the alloying
P was observed, indicating that electronic interaction between
Ni and P in the Ni–P alloy was not significant.
2
2
2
m /g) was impregnated with small amount of NiCl solution
2
overnight and dried at 423 K. After being calcined at 623 K for
2
.0 h, the precursor was reduced by NaH PO in alkaline solution
2 2
to create crystalline nuclei. Then, the sample was transferred into
50 mL solution containing 40 g/L of Na C H O ⋅2H O (sodium
2
3
6
5
7
2
citrate), CH COONa, NiCl ⋅6H O, NaH PO ⋅H O (pH 4–5) with
3
2
2
2
2
2
vigorous stirring for electroless plating at 363 K until the evolution
of a gas had ceased, usually within 2 h. The resulting black solid
−
was washed thoroughly with distilled H O until free from Cl ions,
2
then washed with absolute alcohol (EtOH), and finally, dried at
3
63 K in the N atmosphere. The Ni–P/SiO –MgO amorphous
The Ni loading and the composition of the as-prepared cata-
2
2
catalyst was obtained by using SiO –MgO (Si/Mg molar ratio = 7)
instead of SiO , which was prepared by impregnating the SiO
support with desired amount of Mg(NO ) solution followed by a
calcination at 773 K for 6 h. The W-doped Ni–P/SiO amorphous
catalyst (Ni–W–P/SiO ) was prepared in the same procedure by
adding desired amount of Na WO into the plating solution.
lysts were analyzed by ICP. The surface active area (S ) was
2
act
2
2
determined by hydrogen chemisorption assuming H/Ni(s) = 1.
As shown in Table 1, the presence of W-dopant caused an
increase in P content in the Ni–P alloy composition. The Sact of
3
2
2
2
Ni–W–P/SiO slightly decreased since partial surface Ni atoms
2
2
4
were covered by W oxides. The lower Sact of Ni–P/SiO –MgO
2
For comparison, the crystallized Ni–P/SiO catalyst was also
sample could be attributed to the decrease in the surface area of
the support, leading to the decrease in the dispersion degree of
2
prepared by treating the fresh Ni–P/SiO sample at 873 K in N
flow for 2 h. The Ni/SiO catalyst was prepared by H reduction
2
2
2
2
the Ni active sites. The crystallization of the Ni–P/SiO amor-
2
at 773 K for 2 h of the precursor obtained by a regular impregna-
tion, where Ni(NO ) was used instead of NiCl to avoid the poi-
soning effect from Cl ions.
phous catalyst caused a considerable decrease in Sact owing to the
gathering of Ni–P alloy particles at high temperature. Similar to
the crystallized Ni–P/SiO , the Ni/SiO catalyst also had low S
3
2
2
−
2
2
act
Copyright © 2000 The Chemical Society of Japan

Chemistry Letters 2000
1049
since the catalyst was prepared by H reduction at high tempera-
the Ni/SiO catalyst), which facilitated the surface addition of
2
2
ture (773 K).
dissociated H to C≡N group; (3) the stronger interaction between
Ni and Ni atoms owing to the shorter Ni–Ni bonding length in
the Ni–P amorphous alloy than that in the pure Ni catalyst and
The ADN hydrogenation was carried out in a tubular glass
fixed-bed reactor (0.8 cm i.d.) in a continuous process. Our pre-
liminary study revealed that the catalyst did not show diffusion
restrictions under present conditions. The ADN dissolved in
EtOH was vaporized in a bubbling-type evaporator at 523 K, and
then fed into the reactor with hydrogen mixture as a carrier gas
6
the more highly unsaturated Ni active sites, which were proved to
7
be favorable for the hydrogenation. Since no significant change
in the azacycloheptane yield was observed, the better selectivity
to HMA over the Ni–P/SiO amorphous catalyst was mainly
2
(
75 mL/min). The effluents were analyzed by means of an on-
attributed to its higher activity which promoted the further hydro-
genation of 6-aminohexanenitrile to HMA, as shown in Table 2.
2. Sueiras et al. reported that the higher acidity of the support
favored the formation of azacycloheptane. This may success-
fully account for the promoting effect of the MgO additive (an
alkaline oxide) on the selectivity to HMA, since the formation of
the azacycloheptane was effectively inhibited. 3. The role of the
line gas chromatograph (GC 1102) with a flame ionization detec-
tor equipped with a 25 m OV 101 capillary column, and a
3
The reaction conditions and the experimental results are summa-
rized in Table 2.
4,5
53–533 K oven temperature programmed at a ramp of 4 K/min.
W-dopant in the Ni–W–P/SiO amorphous catalyst is relatively
2
complex. The XPS characterization revealed that most of the W
4+
modifiers were present in the form of the low-valent state (W ).
4+
On one hand, these electropositive W species could strongly
adsorb the cyano group, leaving more Ni active sites for adsorb-
ing hydrogen. Our kinetic studies revealed that the ADN hydro-
genation was first order with respect to hydrogen and zero order
to ADN. Therefore, the activity was enhanced by W-dopant
since more hydrogen molecules could be adsorbed and dissociat-
ed on the Ni active sites. The higher activity of the
As shown in Table 2, besides the main product (HMA), two
additional products, 6-aminohexanenitrile, and azacycloheptane
were identified over each catalyst. The 6-aminohexanenitrile is
an intermediate resulted from the half-hydrogenation of ADN in
which only one cyano group (C≡N) was hydrogenated, while the
azacycloheptane is a real side-product (a secondary amine),
Ni–W–P/SiO amorphous catalyst could also account for the
2
lower yield of 6-aminohexanenitrile since it favored the further
hydrogenation of 6-aminohexanenitrile to HMA, as discussed
4+
above. However, the W species could act as Lewis adsorption
12
sites. the ADN molecule being adsorbed via the donation of a
lone electron pair from the nitrogen of cyano group. This bond-
ing polarized the C≡N bond which was favorable for a nucle-
ophilic attack on the carbon atom by the nitrogen atom with a
2
whose formation could be illustrated in the following scheme.
lone electron pair in NH group on the other side of the ADN
2
molecule to form azacycloheptane. Therefore, no much
improvement on the selectivity to HMA was observed due to the
presence of larger amount of azacycloheptane, although the 6-
aminohexanenitrile formation was effectively inhibited.
This work was supported by the National Natural Science
Foundation of China (29973025), the Committee of Shanghai
Education (98SG44), and the Natural Science Foundation of
Shanghai, China (98QMA11402).
Obviously, the high activity of the catalyst favors the further
hydrogenation of 6-aminohexanenitrile to produce HMA. The
polarization of the cyano group adsorbed on the catalyst seems to
be favorable for the nucleophilic attack on the carbon atom by
References
nitrogen atom in the NH group on the other side of the ADN
2
1
2
3
4
5
A. Baiker, J. De Pietro, M. Maciejewski, and B. Walz, Stud. Surf. Sci.
Catal., 67, 169 (1991).
1
2
molecule and thus for the formation of the azacycloheptane.
C. De Bellefon and P. Fouilloux, Catal. Rev.—Sci. Eng., 36, 459
Based on above assumption, the results in Table 2 could be
explained as follows.
(1994).
D. Gavroy, C. Joly-Vuillemin, G. Cordier, P. Fouilloux, and H.
Delmas, Catal. Today, 24, 103 (1995).
1
. The Ni–P/SiO amorphous catalyst exhibited higher
2
activity than either the crystallized Ni–P/SiO catalyst or the
F. Medina, P. Salagre, and J. E. Sueiras, J. Mol. Catal., 81, 363
2
(1993).
Ni/SiO catalyst. As no significant electronic interaction
2
F. Medina, P. Salagre, and J. E. Sueiras, J. Chem. Soc. Faraday
Trans., 90, 1455 (1994).
between Ni and P in the Ni–P amorphous alloy was determined
by XPS, the structural modification by the alloying P was consid-
ered as a main promoting factor. Such a structural modification
6 H. X. Li, Y. P. Xu, and J. F. Deng, New J. Chem., 23, 1059 (1999).
7 A. Baiker, Faraday Discuss. Chem. Soc., 87, 239 (1989).
8 A. Molnar, G. V. Smith, and M. Bartok, Adv. Catal., 36, 329 (1989).
9 J. F. Deng, H. X. Li, and W. J. Wang, Catal. Today, 51, 113 (1999).
10 S. Yoshida, H. Yamashita, T. Funabiki, and T. Yonezawa, J. Chem.
Soc., Chem. Commun., 1982, 964.
resulted in (1) the higher dispersion degree (S ), as shown in
act
Table 1. However, this could not explain the higher TOF value
–
1
of the Ni–P/SiO amorphous catalyst (0.012 s ) than that of the
2
–
1
11 H. Li, H. X. Li, W. L. Dai, W. J. Wang, Z. Fang, and J. F. Deng,
Appl. Surf. Sci., 152, 25 (1999).
12 P. Gallezot, P. J. Cerino, B. Blanc, G. Fleche, and P. Fuertes, J.
Catal., 146, 93 (1994).
Ni/SiO catalyst (0.0029 s ); (2) the relatively weaker Ni–H
2
bonging, as determined by TPD (the lower H desorption peak
2
on the Ni–P/SiO amorphous catalyst in comparison with that on
2