Catalysis Letters

, Volume 118, Issue 1, pp 50–54

Liquid phase hydrogenation of maleic anhydride over nickel catalyst supported on ZrO2–SiO2 composite aerogels

Authors

  • Chun-guang Gao
    • School of Chemistry and Chemical EngineeringShanxi University
    • Institute of Advanced ChemistryShanxi University
    • School of Chemistry and Chemical EngineeringShanxi University
    • Institute of Advanced ChemistryShanxi University
Article

DOI: 10.1007/s10562-007-9135-4

Cite this article as:
Gao, C., Zhao, Y. & Liu, D. Catal Lett (2007) 118: 50. doi:10.1007/s10562-007-9135-4

Abstract

The promotional effects of ZrO2 on Ni/ZrO2–SiO2 catalysts were investigated by the comparison of Ni/SiO2 and Ni/ZrO2–SiO2 activity in the hydrogenation of maleic anhydride (MA) to γ-butyrolactone (GBL), and by the measurements of X-ray diffraction (XRD), Fourier transform infrared (FT-IR), temperature-programmed reduction (TPR) and ammonia temperature-programmed desorption (NH3-TPD). The presence of ZrO2 led to an obvious increase of GBL yield. The promotion effect could be attributed to the possible presence of Zr4+ species on the catalyst surface owing to the higher ionicity of the Zr–O bonds, and to the proper interaction of Ni with the ZrO2-SiO2 support that is regulated by the presence of ZrO2.

Keywords

ZrO2–SiO2 composite aerogelsγ-butyrolactonehydrogenation of maleic anhydridenickel

Introduction

GBL is currently one of the most valuable alternatives to the environmentally harmful chlorinated solvents, which have been widely used in the polymer and paint industries. Hydrogenation of MA is the most direct way to produce GBL and it does not involve the use of hazardous materials. In published works, different catalysts were used in the MA hydrogenation in different ways, such as liquid-phase, gas phase and homogeneous reactions, and the activity performance is dependent on the catalysts active phase, support and promoter.

Uwe Herrmann and Gerhard Emig [1] investigated the liquid phase hydrogenation of MA in a packed bubble column reactor using different copper-based catalysts. A copper-zinc catalyst was found to be active in the formation of 1,4-butanediol (BDO), whereas on zinc-free copper catalysts, mainly succinic anhydride (SA) and GBL were formed. They had also investigated the liquid phase hydrogenation of MA and intermediates in a stirred tank slurry reactor using copper-based and noble metal catalysts [2]. The copper chromites and noble metal catalysts were found to be suitable for the hydrogenation of MA.

R.M. Deshpande et al. [3] investigated the liquid phase hydrogenation of SA to GBL and BDO using ruthenium–cobalt bimetallic catalysts in a semi-batch slurry reactor. It was found that addition of small amounts of ruthenium to cobalt catalyst increased the overall hydrogenation activity of the cobalt catalyst. Ruthenium doping mainly influenced the formation of THF from SA and further hydrogenation of butanediol to propanol. The hydrogenation of GBL to BDO and THF to butanol steps seemed to be independent of ruthenium.

Yoshinori Hara and Kazunari Takahashi [4] prepared a ruthenium catalyst system consisting of Ru salts, trialkylphosphine and p-toluene sulfonic acid, which was found very effective for the hydrogenation of SA affording GBL in liquid-phase, and exhibited excellent catalyst performance with exceeding 97% selectivity for GBL and high activity.

Seong Moon Jung et al. [5,6] investigated the promotional effects of Sn on Pd/SiO2 by the comparison of Pd/SiO2 and Pd–Sn/SiO2 activity in the hydrogenation of MA and SA to GBL. As a result, it is proposed that Pd–Sn/SiO2 catalyst can be an effective one-step catalyst for hydrogenation of MA to GBL. The promotional effect of tin on Pd–Sn/SiO2 catalyst can be explained by the change of electronic configuration of Pd due to the interaction with Sn. The effective interaction between Pd and Sn in the hydrogenation of MA to GBL is enhanced when the reduction temperature is performed at higher temperature.

Jian Xu et al. [7,8] prepared a Pd/TiO2 catalyst for the liquid phase hydrogenation of MA. Over Pd/TiO2 catalyst prepared by sol–gel and super critical fluid of ethanol drying, 93.5% yield to butyric acid and 100% conversion of MA were obtained at 240 °C, 3 MPa.

Heondo Jeong et al. [9] prepared a catalyst composed of palladium, molybdenum and nickel on a silica support of high surface area. This multi-component mixed metal oxide catalyst system in a batch-type reactor is found to be suitable for the hydrogenation of MA selectively to GBL. The hydrogenation of MA using this catalyst is not observed by-products and the selectivity of GBL is very high.

Tongjie Hu et al. [10] prepared Cu–Zn–Ti catalysts by coprecipitation method and applied the catalyst to gas phase hydrogenation of MA to GBL. The Cu–Zn–Ti catalysts with Cu:Zn:Ti ratios of 1:2:1 and 2:2:1.5 showed high catalytic activity in the gas phase hydrogenation of MA to GBL.

In this paper, we report a study on the hydrogenation of maleic anhydride on ZrO2–SiO2 composite aerogels and SiO2 aerogels supported nickel catalysts. The catalysts were characterized by means of XRD, FT-IR, TPR and NH3-TPD. The aim of this work is to realize an investigation of the effects of the presence of ZrO2 on supported nickel catalysts. Moreover, the effects of the reaction conditions have also been investigated to taking into account the potential industrial relevance of "single-step" hydrogenation of MA to GBL with the Ni/ZrO2–SiO2 catalysts.

Experiment procedure

Preparation of catalysts

The synthesis procedure for ZrO2–SiO2 composite aerogels and SiO2 aerogels adapted from that already reported by our group in [11] is described briefly below. A proper amount of zirconyl nitrate dihydrate (ZrO(NO3)2·2H2O) was dissolved in an ethanol solution, then acetic acid, water and tetraethoxysilicane (TEOS) were added into under constant stirring. The mixture was transferred into an autoclave and heated to the supercritical condition of ethanol (516 K, 6.36 MPa) and kept the conditions for about 0.5 h. Then unclosed the leak valve let the gas leaking slowly from the autoclave, and swept with N2 before room temperature was reached. The obtained composite aerogels monolith ZrO2–SiO2 was crushed into 40–60 mesh. The ZrO2 content 5 wt. % was selected in this work according to a content selection test (data not shown). The SiO2aerogels was prepared with the same above process without adding ZrO(NO3)2·2H2O.

A proper amount of Ni(NO3)2 was impregnated with the above ZrO2–SiO2 and SiO2 aerogels respectively to obtain Ni/ZrO2–SiO2 (catalyst code: 30Ni/ZrO2–SiO2) and Ni/SiO2 (catalyst code: 30Ni/SiO2) catalysts, and both were calcined at 673 K for 3 h and reduced with hydrogen at 673 K for 3 h. The content of Ni is selected to 30 wt. % according to the previous work of our group [12].

Characterization of catalysts

Catalysts were characterized with FT-IR (Shimadzu 8400S FT-IR, Japan), XRD (Rigaku D/max 2500, Cu Ka radiation, 50 kV, 60 mA), N2 adsorption-desorption isotherms (Micromeritics ASAP 2020), TPR, NH3-TPD and elemental analysis (AtomScan 16).

TPR measurements were carried out from 50 to 700 °C in a conventional lab-made apparatus equipped with a thermal conductivity detector (TCD) under the following conditions: heating rate, 10 °C min−1; flow rate of nitrogen gas containing 10 vol.% of hydrogen, 30 cm3 min−1; catalyst weight, 30 mg.

NH3-TPD experiments were also recorded on the above apparatus fitted with a TCD. The samples (0.1 g) were pretreated at 500 °C in a He stream for 2 h in order to remove surface contaminants, and ammonia was absorbed at 50 °C. The heating rate and sweep gas (helium) flow during the measurement of ammonia desorption were 10 °C and 60 cm3 min−1, respectively.

Catalytic test

The catalytic hydrogenation of MA (reaction conditions: 453 K, 3.0 MPa, 3 h) in liquid phase was performed on the prepared catalysts within an autoclave (FYX 0.1, Da-Lian No.4 Instrument Plant, China) using 4.9 g MA and 0.1 g catalysts, and THF was charged in the autoclave as solvent. The reactor was heated to the desired temperature while being filled with known amounts of hydrogen based on the set pressure.

The components MA, SA and GBL were analyzed with a gas chromatography (GC-930, Hai-Xin Company, Shanghai China) equipped with a capillary column (SE-52, Lan-zhou Institute of Chemical Physics, China), column temperature was 333 K, and quantitative analysis was performed using a CDP-4A integrator (Wu-Feng Scientific Instruments Ltd. Shanghai, China). Peak areas were converted into concentration using the internal standard normalization.

Results and discussion

Characterization of catalysts

XRD analysis

The features exhibited on the XRD patterns (figure 1.) of the catalysts 30Ni/SiO2 and 30Ni/ZrO2–SiO2 showed Ni and NiO were both existed in the two catalysts and the two catalysts had no obvious difference in crystal structure and particle size. That is, the addition of ZrO2had no influence on the catalysts crystal structure, and the characteristic peaks of ZrO2 were not detected meaning that ZrO2 should be present in amorphous phase. The previous published work of our group [11] has studied the crystal composition of ZrO2–SiO2 composite aerogels, and the result suggests when ZrO2 content is lower than 30 wt. %, the ZrO2–SiO2 composite aerogels are X-ray amorphous due to the incorporation of zirconium atoms in the framework of amorphous SiO2.
https://static-content.springer.com/image/art%3A10.1007%2Fs10562-007-9135-4/MediaObjects/10562_2007_9135_Fig1_HTML.gif
Figure 1

The XRD patterns of 30Ni/SiO2 and 30Ni/ZrO2–SiO2 samples

FT-IR analysis

The figure 2 is the FT-IR spectra of 30Ni/SiO2, 30Ni/ZrO2–SiO2 and pure SiO2. For the two catalysts, the peak 964 cm−1 assigned to Si-OH bond stretching of SiO2 [13] decrease indicating a presence of M–O–Si band (in this work M = Zr or Ni) [11,12]. And the 30Ni/ZrO2–SiO2 has the lowest 964 cm−1 peak, meaning more M–O–Si bonds are formed.
https://static-content.springer.com/image/art%3A10.1007%2Fs10562-007-9135-4/MediaObjects/10562_2007_9135_Fig2_HTML.gif
Figure 2

The FT-IR spectra of 30Ni/SiO2 and 30Ni/ZrO2–SiO2 samples

N2 adsorption–desorption isotherm analysis

The N2 adsorption–desorption isotherm analysis results of the two catalysts and the two supports are listed in table 1. Because the supports were not calcined before impregnation, the SBET of the supports are large more than those of the corresponded catalysts. But the addition of ZrO2 slightly decreases the SBET, and no influence on the isotherm type and pore distribution.
Table 1

The results of N2 adsorption–desorption isotherm analysis

Samples

SBET(m2/g)

Pore volume (cm3/g)

ZrO2–SiO2 aerogels support

898

2.14

SiO2 aerogels support

1015

2.13

30Ni/ZrO2–SiO2

279

0.38

30NiO2/SiO2

285

0.39

TPR analysis

The TPR profiles of 30Ni/SiO2 and 30Ni/ZrO2–SiO2 are shown on figure 3. The effect of ZrO2 on the reduction of NiO is clearly shown by the changes of the temperature of the maximum of the main peaks and an upheaval of broad bands at about 500–750 °C.
https://static-content.springer.com/image/art%3A10.1007%2Fs10562-007-9135-4/MediaObjects/10562_2007_9135_Fig3_HTML.gif
Figure 3

The TPR patterns of 30Ni/SiO2 and 30Ni/ZrO2–SiO2 samples

Comparing the TPR patterns of the sample 30Ni/SiO2 and 30Ni/ZrO2–SiO2, we could suppose that the ZrO2 has an effect of regularity on the interaction strength between the Ni and the ZrO2–SiO2 support, that is, decrease the interaction strength of the strong interaction sites and increase the interaction strength of the weak interaction sites, causing the TPR maximum temperature of the main peak slightly move to higher temperature, hunch broad bands at about 500–750 °C and increase hydrogen consumption over the tested temperature range. This regularity effect can be attributed to the higher ionicity of Zr–O bond than that of covalent Si–O band, and the molecular-level mixture of ZrO2 with SiO2 matrix.

NH3-TPD analysis

The NH3-TPD pattern of the 30Ni/SiO2 sample (figure 4. dotted line) is divided into three regions: region I-90–200 °C; region II-200–500 °C and region III-500–800 °C. The 30Ni/ZrO2–SiO2 sample shows a two regions NH3-TPD pattern (figure 4 solid line): regionI-90–550 °C and region II-550–800 °C. The NH3-TPD patterns of the supports SiO2 and ZrO2–SiO2 (shown in fig.4 with two transparent labeled lines) are also divided into two regions.
https://static-content.springer.com/image/art%3A10.1007%2Fs10562-007-9135-4/MediaObjects/10562_2007_9135_Fig4_HTML.gif
Figure 4

The NH3-TPD curves of 30Ni/SiO2, 30Ni/ZrO2–SiO2, SiO2 support and ZrO2–SiO2 support

For analysis the factors that affect the acid distribution features, we could roughly treat the 90∼500 °C range as the weak acid region and 500∼800 °C range as the strong acid region. In the weak acid region the acid density is fall in the order of ZrO2–SiO2-support, 30Ni/ZrO2–SiO2, 30Ni/SiO2 and SiO2-support. But in the strong acid region, the acid density is fall in the order of SiO2-support, ZrO2–SiO2-support, 30Ni/SiO2 and 30Ni/ZrO2–SiO2. Then it could be concluded that the presence of Ni and ZrO2 both increases the density of weak acid strength sites and decreases the density of strong acid strength sites; but in the strong acid region Ni is the main factor, and in the weak acid region ZrO2 is the main factor and which is related more with the catalytic selectivity.

The surface acidity of ZrO2–SiO2 mixed oxide has been studied intensively by several groups [1417] and it is widely accepted that the presence of ZrO2 increase both Lewis and Brønsted acid centers over ZrO2–SiO2 mixed oxide.

S. Damyanova and co-workers [14] have studied the acid promotion mechanism of ZrO2 with x-ray photoelectron spectroscopy (XPS) and FT-IR. They supposed the zirconium is responsible for Lewis acidity in ZrO2/SiO2 mixed oxide due to the higher ionicity of the Zr–O bond, and Brønsted acid sites are produced owing to the electron density of the OH bonds is reduced by the Zr–O higher ionicity bonds that are neighboring the more covalent Si–O units.

James A. Anderson and co-workers [15] have especially studied the surface acidity of silica-zirconia aeeogels with different Si/Zr ratio. They also proved the Lewis acid sites originate from zirconia and a roughly linear increase in Lewis acid density as a function of mol% Zr is expectable.

Applying above ideas about the surface acidity generation in ZrO2–SiO2 mixed oxides to the 30Ni/SiO2 sample, we could believe that the increase of the density of weak acid strength sites is owing to the presence of ZrO2, and also could suppose most of the increased acid sites are Lewis acid sites. Meanwhile, as the FT-IR and TPR analysis shown (figures 2 and 3), the presence of ZrO2 also increased the interaction between Ni species and ZrO2-SiO2 supports, and more Si–O–M bonds are formed on the surface of 30Ni/ZrO2–SiO2 samples than those formed on the 30Ni/SiO2 samples, which decrease the formation of acidic hydroxyls on the surface of the supports and resulting in a decrease of the density of Brønsted acid sites.

Elemental analysis

According to the results of elemental analysis, the Ni content for sample 30Ni/SiO2 is 30.15 wt. %; and for sample 30Ni/ZrO2–SiO2, the Ni content is 31.18 wt. %, and the ZrO2 content is 4.68 wt. %.

Catalytic test

Table 2 is the results of catalytic hydrogenation of MA. Over the two catalysts, MA was converted completely into SA and GBL, and no other relative hydrogenated products were detected. The great difference displayed in the selectivity to GBL. The presence of 5 wt. % ZrO2 component plays an important role in catalytic selectivity to GBL. On the optimized conditions, over 30Ni/ZrO2–SiO2 the selectivity to GBL is 84.8 %, while over 30Ni/SiO2 only small amount of GBL was obtained no matter the reaction conditions.
Table 2

Catalytic performance of MA hydrogenation on different conditions

Catalyst

Reaction temperature (K)

Reaction pressure (MPa)

Reaction time (h)

MA conversion (%)

GBL selectivity (%)

30Ni/ZrO2–SiO2

393

6

3

100.0

2.9

423

6

3

100.0

28.3

453

6

3

100.0

63.2

453

3

3

100.0

15.0

453

4

3

100.0

22.3

453

5

3

100.0

52.4

453

6

2

100.0

31.1

453

6

4

100.0

69.4

453

6

6

100.0

76.4

453

6

8

100.0

84.8

30NiO2/SiO2

453

3

3

100.0

0.7

453

6

8

100.0

0.6

As shown in Scheme 1, the hydrogenation of MA to GBL follows two-step reactions. Firstly, C=C bond in MA is reduced and converted to SA. And then, GBL is formed via the subsequent oxidative-hydrogenation of C=O bond in SA.
https://static-content.springer.com/image/art%3A10.1007%2Fs10562-007-9135-4/MediaObjects/10562_2007_9135_Sch1_HTML.gif
Scheme 1

The catalytic hydrogenation pathway of MA to GBL

The obvious enhancement of the selectivity to GBL over 30Ni/ZrO2–SiO2 can be explained by the modification of its surface electronic properties towards the improved adsorption of C=O bonds.

Just as the acid promotion mechanism of ZrO2, firstly, we can attribute the effect of activation of C=O bond to the possible presence of Zr4+ species on the catalyst surface owing to the higher ionicity of the Zr–O bonds. And another reason for the increase of the selectivity to GBL could attribute to the proper interaction of Ni with the ZrO2–SiO2 support that is regulated by the presence of ZrO2.

Comparing with other reported liquid-phase MA hydrogenation systems [110], the advantage of this 30Ni/ZrO2–SiO2 catalyst is that the GBL can be obtained with a high yield via a one-step continuous operation and no other by-products are formed besides the SA. And it could be proposed that nickel based catalyst supported on ZrO2–SiO2 composite aerogels can be an effective one-step catalyst for hydrogenation of MA to GBL.

Conclusions

The promotional effects of ZrO2 on nickel catalyst supported on ZrO2–SiO2 composite aerogels were studied systematically. The results indicates that the introduction of ZrO2 into SiO2 matrix greatly improved the selectivity to GBL than that on SiO2 supported Ni catalysts. And it could be proposed that nickel based catalyst supported on ZrO2–SiO2 composite aerogels can be an effective one-step catalyst for hydrogenation of MA to GBL.

In the catalysts, the presence of Ni and ZrO2 both increases the density of weak acid strength sites and decreases the density of strong acid strength sites; but in the strong acid region Ni is the main factor, and in the weak acid region ZrO2 is the main factor and which is related more with the catalytic selectivity.

The presence of ZrO2 in the catalyst also has an effect of regularity on the interaction strength between the Ni and the ZrO2–SiO2 support.

Acknowledgments

This work is supported by the Industry Development Fund of Shanxi Province (041122 and 031099).

Copyright information

© Springer Science+Business Media, LLC 2007