GModel
APCATA-13879; No. of Pages7
ARTICLE IN PRESS
S.A. Regenhardt et al. / Applied Catalysis A: General xxx (2012) xxx–xxx
7
In this case, not only the yield in CH4 was the lowest, but GBL
production was much more important than PA one. This means
that the main reaction pathway with Ni/SiO –Al O was the selec-
an important influence over the final structure of the metallic nickel
surface, as it was showed by temperature programmed decom-
2+
position of NH , a structure sensitive reaction. The Ni –support
2
2
3
3
tive hydrogenolysis of SA into GBL. Thus, selectivity to GBL was
about 88% (Table 4) and it kept almost constant during the whole
run. Finally, the trends with Ni/H-BEA showed a very important
decay for GBL production rate with time on stream, while amaz-
ingly PA production rate kept almost constant (Figs. 7B and 8B).
This is almost the opposite behavior to that observed at 170 C
with the same catalyst (Figs. 7A and 8A). Thus, with Ni/H-BEA, the
selectivity to GBL diminished from 88 to 79% (Table 4) during the
interaction can be regulated by the concentration and strength of
the acid sites on the support surface. Thus, the pattern found for
2+
Ni –support interaction is: Ni/H-BEA > Ni/SiO –Al O > Ni/SiO .
2
2
3
2
The surface structure of the active metal nickel phase is strongly
depending on this interaction degree. In this work, it was shown
that the surface structure of metal nickel and, as a consequence,
the performance of Ni-based catalyst on the selective hydrogenol-
ysis of succinic anhydride into ␥-butyrolactone, that is also a
structure sensitive reaction, can be optimized by modulating the
◦
3
hour run. These results are indicating that the active sites for SA
2+
hydrogenolysis on Ni/H-BEA are suffering an important selective
deactivation during the reaction, i.e. especially the sites that are
active for SA hydrogenolysis into GBL. Then, this selective deactiva-
tion becomes more important with temperature, which is opposite
to that observed for Ni/SiO2 catalyst.
Ni –support interaction. The best stability and GBL production is
2+
reached when a medium Ni –support interaction is obtained, as
in the case of Ni/SiO –Al O catalyst. Thus, a selectivity as high
2
2
3
as 88%, at succinic anhydride conversions of 60% or higher, are
◦
2+
obtained at 220 C. If the Ni –support interaction is too low, as
◦
From the results obtained at 220 C, the following initial activ-
in the case of Ni/SiO , the metal nickel surface is more selective for
2
ity pattern is suggested: Ni/H-BEA > Ni/SiO > Ni/SiO –Al O .
the hydrogenolysis of succinic anhydride into propionic acid than
2
2
2
3
However,
after
3 hours,
this
pattern
changed
to:
in the case of Ni/SiO –Al O . If the interaction is too high, as in the
2 2 3
Ni/SiO > Ni/SiO –Al O > Ni/H-BEA. These results indicate that
case of Ni/H-BEA, a loss of stability is observed during the run due to
selective blockage of hydrogenolytic sites located in the micropores
of the zeolite, probably by strong adsorption of carbon compounds
2
2
2
3
Ni/SiO –Al O and Ni/SiO2 catalysts were more stable than
2
2
3
Ni/H-BEA, especially at the highest temperature used in this
work. The activity loss of Ni/H-BEA is mainly due to deactivation
of hydrogenolytic sites, probably due to micropore blockage
by compounds formed during reaction. The pattern for ini-
◦
formed during reaction, especially at 220 C.
References
◦
tial selectivity to GBL at 220 C can be resumed as follows:
∼
[1] M. Messori, A. Vaccari, J. Catal. 150 (1994) 177–185.
Ni/H-BEA = Ni/SiO –Al O > Ni/SiO . However, after 3 hours on
2
2
3
2
[
2] R. Zhang, H. Yin, D. Zhang, L. Qi, H. Lu, Y. Shen, T. Jiang, Chem. Eng. J. 140 (2008)
88–496.
[3] G. Budroni, A. Corma, J. Catal. 257 (2008) 403–408.
◦
stream at 220 C, the pattern changed to: Ni/SiO –Al O > Ni/H-
2
2
3
4
BEA > Ni/SiO . This is because the deactivation of Ni/H-BEA is
2
[
4] G.L. Castiglioni, A. Vaccari, G. Fierro, M. Inversi, M. Lo Jacono, G. Minelli, I. Pettiti,
P. Porta, M. Gazzano, Appl. Catal. A: Gen. 123 (1995) 132–144.
5] Y.L. Zhu, J. Yang, G.Q. Dong, H.Y. Zheng, H.H. Zhang, H.W. Xiang, Y.W. Li, Appl.
Catal. B: Environ. 57 (2005) 183–190.
affecting selectively the conversion of SA into GBL during the
catalytic test. Ni/SiO2 was more active than Ni/SiO –Al O . How-
2
2
3
[
ever, the large metal particle formed on Ni/SiO2 is favoring the
formation of a metal nickel surface with sites that promote planar
adsorption of SA and the subsequent preferential hydrogenol-
ysis into PA. Instead, small metal nickel particles formed on
[6] H. Jeong, T.H. Kim, K.I. Kim, S.H. Cho, Fuel Process. Technol. 87 (2006) 497–503.
[
[
7] S.H. Vaidya, C.V. Rode, R.V. Chaudhari, Catal. Commun. 8 (2007) 340–344.
8] S.M. Jung, E. Godard, S.Y. Jung, K.C. Park, J.U. Choi, J. Mol. Catal. A: Chem. 198
(
2003) 297–302.
2+
Ni/SiO –Al O , from reduction of small NiO particles and Ni
[9] W. Lu, G. Lu, Y. Guo, Y. Guo, Y. Wang, Catal. Commun. 4 (2003) 177–181.
10] C. Ohlinger, B. Kraushaar-Czarnetzki, Chem. Eng. Sci. 58 (2003) 1453–1461.
11] A.J. Marchi, J.L.G. Fierro, J. Santamaría, A. Monzon, Appl. Catal. A: Gen. 142
2
2
3
[
[
interacting strongly with support surface, would be more selective
for SA hydrogenolysis into GBL.
In summary, selective hydrogenolysis of SA into GBL with Ni-
based catalysts is strongly depending on the nature of support
and reaction conditions. The highest stability, production rate and
(
1996) 375–386.
[12] C.I. Meyer, A.J. Marchi, A. Monzón, T.F. Garetto, Appl. Catal. A: Gen. 367 (2009)
22–129.
1
[
13] S. Qi, B.A. Cheney, R. Zheng, W.W. Lonergan, W. Yu, J.G. Chen, Appl. Catal. A:
Gen. 393 (2011) 44–49.
◦
selectivity to GBL were obtained at 220 C with Ni/SiO –Al O . This
[14] A. Saadi, R. Merabti, Z. Rassoult, M.M. Bettahar, J. Mol. Catal. A: Chem. 253 (2006)
79–85.
2
2
3
2+
catalyst has an intermediate interaction Ni –support for the series
used in this work, indicating that there is an optimum interaction
that favors the formation of a metallic nickel surface that is active,
stable and selective for the hydrogenolysis of SA into GBL. If the
[
15] C.I. Meyer, S.A. Regenhardt, A.J. Marchi, T.F. Garetto, Appl. Catal. A: Gen.
17–418 (2012) 59–65.
4
[16] A. Simon-Masseron, J.P. Marques, J.M. Lopes, F. Ramoa Ribeiro, I. Gener, M.
Guisnet, Appl. Catal. A: Gen. 316 (2007) 75–82.
[17] R. Fréty, L. Tournayan, M. Primet, G. Bergeret, M. Guenin, J.B. Baumgartner, A.
Borgna, J. Chem. Soc. Faraday Trans. 89 (17) (1993) 3313–3318.
[18] P. Casta n˜ o, B. Pawelec, J.L.G. Fierro, J.M. Arandes, J. Bilbao, Fuel 86 (2007)
2262–2274.
2
+
interaction Ni –support is very low, as in the case of Ni/SiO , an
2
active and stable metal nickel phase is formed, but with lower selec-
tivity to GBL. Finally, if the Ni2+–support interaction is too high, a
very active and selective metal nickel phase is obtained, but with
poor stability, especially at high temperature.
[
[
[
19] S.R. Kirumakki, B.G. Shpeizer, G.V. Sagar, K. Chary, A. Clearfield, J. Catal. 242
2006) 319–331.
20] E.G.M. Kuijpers, M.W.C.M.A. Nieuwesteeg, G.J. Wermer, J.W. Geus, J. Catal. 112
1988) 107–115.
21] N.M. Bertero, A.F. Trasarti, C.R. Apesteguía, A.J. Marchi, Appl. Catal. A: Gen. 394
(2011) 228–238.
(
(
4
. Conclusions
[
[
22] H. van’t Blik, R. Prins, J. Catal. 97 (1986) 188–199.
23] E. van Steen, G.S. Sewell, R.A. Makhote, C. Micklethwaite, H. Manstein, M. de
Lange, C.T. O’Connor, J. Catal. 162 (1996) 220–229.
It was shown that the support nature affects the activity,
selectivity and stability of Ni-based catalysts in the gas phase
hydrogenation of maleic anhydride. This effect is clearly influenc-
ing the step of hydrogenolysis of succinic anhydride over nickel,
in which ␥-butyrolactone and propionic acid are obtained as main
[24] C.M.N. Yoshioka, M.H. Jordao, D. Zanchet, T.F. Garetto, D. Cardoso, Appl. Catal.
A: Gen. 355 (2009) 20–26.
[
25] A. Fúnez, A. de Lucas, P. Sánchez, M.J. Ramos, J.L. Valverde, Chem. Eng. J. 136
2008) 267–275.
(
[26] C. Plana, S. Amenise, A. Monzón, E. García-Bordejé, J. Catal. 275 (2010) 228–235.
[27] X.-K. Li, W.-J. Ji, J. Zhao, S.-J. Wang, C.-T. Au, J. Catal. 236 (2005) 181–189.
2
+
products. The Ni –support interaction in the oxide precursor has