Catal Lett (2011) 141:1458–1463 DOI 10.1007/s10562-011-0642-y
Glycerol Hydrogenolysis over Co Catalysts Derived from a Layered Double Hydroxide Precursor Xiaohui Guo • Yong Li • Wei Song Wenjie Shen
•
Received: 17 February 2011 / Accepted: 31 May 2011 / Published online: 15 June 2011 Ó Springer Science+Business Media, LLC 2011
Abstract Co catalysts, obtained from a layered double Co–Zn–Al hydroxide, are highly active and stable towards the hydrogenolysis of glycerol to 1,2-propanediol (1,2PDO) in aqueous media. The Co-673 catalyst, containing a CoO species, provided a glycerol conversion of 67.7% and a 1,2-PDO selectivity of 50.5%. The Co-873 catalyst comprising 16 nm Co nanoparticles gave a glycerol conversion of 70.6% and a 1,2-PDO selectivity of 57.8%. It was revealed that the CoO species in the Co-673 catalyst was readily converted to 50 nm Co particles under the glycerol hydrogenolysis conditions. The Co catalysts maintained a stable size and phase in recycling tests. Keywords Glycerol Hydrogenolysis 1,2-Propanediol Layered double hydroxide Co catalyst
1 Introduction Biodiesel is regarded as a renewable and clean fuel, but its production via the transesterification of plant oils and animal fats generates significant levels of glycerol as a byproduct. To make the process more cost effective, the utilization of the crude glycerol to produce value-added fine chemicals and fuel has attracted significant attention in recent years. The hydrogenolysis of glycerol to propanediol is a promising option [1, 2]. Until now, supported Cu [3],
X. Guo Y. Li W. Song W. Shen (&) State Key Laboratory of Catalysis, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, China e-mail:
[email protected]
123
Co [4], Ni [5], Ru [6], Pt [7], Rh [8] and Ir [9] catalysts have been reported to be active towards glycerol hydrogenolysis. However, the sintering of the metal particles during the course of the reaction often resulted in catalyst deactivation [10–13]. For example, the sintering of Cu particles in the Cu/ZnO catalysts reduced the conversion of glycerol from 46 to 10% upon reuse [10]. A similar deactivation pattern was observed for Cu/Al2O3 catalysts [11]. Few catalysts have been demonstrated to be both active and stable towards glycerol hydrogenolysis in aqueous media [13–16]. Recently, Ga2O3-modified Cu/ ZnO catalysts have exhibited an enhanced stability, even after four consecutive glycerol hydrogenolysis reactions at 493 K, without apparent deactivation [16]. ReOx-modified Ir nanoparticles effectively catalyzed glycerol hydrogenolysis in aqueous media at 393 K and the catalyst could be reused at least three times without changes in reaction rate or product selectivity [9]. Layered double hydroxides (LDHs), known as hydrotalcite-like materials, are interesting precursors for preparing well dispersed metal oxides or metal nanoparticles [17]. For instance, Cu-Zn-Al oxides, derived from the corresponding hydrotalcite precursor, proved to be highly active and selective in glycerol hydrogenolysis at 473 K, giving a glycerol conversion of 48.0% and a 1,2- propanediol (1,2-PDO) selectivity of 93.9% [18]. We have recently reported that Co/MgO catalysts provided a glycerol conversion of 44.8% and a 1,2-PDO selectivity of 42.2% in glycerol hydrogenolysis at 473 K, but the hydration of MgO to Mg(OH)2, under the reaction conditions, caused severe aggregation of the cobalt particles [11]. In this work, we have used a LDH to synthesize stable Co catalysts for glycerol hydrogenolysis in aqueous media, and their structural properties and the catalytic activities have been investigated.
Glycerol Hydrogenolysis over Co Catalysts
1459
2 Experimental
2.3 Catalytic Test
2.1 Catalysts Preparation
A 40 g 10 wt% aqueous glycerol solution and 0.3 g Co-T catalyst were added to a stainless steel autoclave (100 mL). After purging with hydrogen three times, the reactor was pressured to 2.0 MPa and heated to 473 K. The reaction was then conducted at 473 K for 12 h. The products from both the gas and liquid phases were analyzed using gas chromatography (Agilent 7890 A). A Carbowax 20 M capillary column connected to a flame ionization detector was employed for the liquid product and a Hayesep D packed column connected to a thermal conductivity detector was used for the analysis of the gas product.
The Co content in the catalysts was 40.1 wt %, as determined by ICP analyses and the Co/Zn/Al mole ratio was 4/2.2/1, in accordance with the stoichiometric composition of the initial solution. Figure 1 shows the XRD patterns for the precursor, the oxide, and the Co-T samples. For the LDH precursor, the characteristic diffraction lines in the (003), (006), (012), (015), (110) and (113) planes reflected the hydrotalcite-like structure [19]. The very weak diffraction lines at 2h = 31.8° and 36.3° may be assigned to small ZnO phase (JCPDS 89-1397). The formation of a layered double hydroxide usually requires a ratio of divalent to trivalent cations in the range of 2 to 4 [19]. Though
*
ZnO (89-1397) Spinels CoO (75-393) fcc Co (89-4307)
*
*
10
20
*
30
40
(110) (113)
(018)
(100) (012) (101) (015)
(006)
The chemical compositions of the samples were measured using inductively coupled plasma-atomic emission spectroscopy (ICP, Plasma-Spec-II spectrometer). Powder X-ray diffraction (XRD) patterns were recorded on a Rigaku D/MAX 2500 diffractometer using Cu Ka radiation operated at 40 kV and 200 mA. The average crystallite size was calculated from the half width of the diffraction lines of CoO (111) or Co (111). Transmission electron microscopy (TEM) images were recorded on a Philip Tecnai G2 spirit microscope operated at 120 kV. The sample was prepared by ultrasonically dispersing the sample powder in ethanol, and droplets of the suspension were deposited on a carbon-coated copper grid and dried in air. Temperature-programmed reduction (H2-TPR) measurements were conducted with a U-type quartz reactor in conjunction with a mass spectrometer (Omnistar QMS 200). 30 mg sample was heated to 573 K under a He flow (50 mL/min) and maintained at that temperature for 0.5 h. After cooling to room temperature, a 20% H2/He mixture (30 mL/min) was introduced and the sample was heated to 1073 K at a rate of 10 K/min. X-ray photoelectron spectroscopy (XPS) measurements were performed on a VG ESCALAB MK2 spectrometer using Al Ka X-ray source. The sample was compressed into a thin disc and mounted on a sample rod in the analysis chamber. The spectra of C1s, O1s, Co2p, Al2p and Zn2p levels were recorded before and after hydrogen reduction at 673 K for 1 h. The charge effect was corrected by adjusting the binding energy of C1s to 284.6 eV.
3.1 Chemical and Physical Properties
(003)
2.2 Characterization
3 Results and Discussion
Intensity (a.u.)
The Co–Zn–Al precursor was prepared using a co-precipitation method. Appropriate amounts metal nitrate, with a Co/Zn/Al molar ratio of 4/2/1 (total ionic concentration of 0.5 M), were dissolved in 100 mL of water. A 100 mL of 1.0 M NaOH and 0.25 M Na2CO3 was used as the precipitating agent. The two solutions were simultaneously added to a 1000 mL flask containing 100 mL of water using peristaltic pumps at the rate of 5 mL/min, under vigorous stirring at 363 K. The mixture was aged at 363 K for 2 h. After filtration and washing with water, the obtained solid was dried at 353 K overnight and calcinated at 723 K in air for 4 h to yield the Co–Zn–Al oxide. The oxide samples were then reduced with H2 in a fixed-bed quartz reactor at 673 or 873 K for 1 h, to produce the Co-T catalysts, where T refers to the temperature of hydrogen reduction.
50
60
70
2 Theta (degree) Fig. 1 XRD patterns for a the LDH b the oxide c the Co-673 and d the Co-873 samples
123
1460
the (Co ? Zn)/Al molar ratio was 6/1 in the initial solution, the partial oxidation of Co2? to Co3? throughout the course of the preparation, facilitated the formation of LDH. This is commonly observed in the preparation of hydrotalcite-like materials that contain cobalt species [20]. After calcination at 723 K, reflections of typical spinel-like phases such as Co3O4 (JCPDS 80-1543), ZnCo2O4 (JCPDS 23-1390), ZnAl2O4 (JCPDS 82-1043) and CoAl2O4 (JCPDS 82-2251) appeared [21, 22]. However, the minor differences in lattice spacing make it difficult to differentiate the exact composition of each phase. After hydrogen reduction at 673–873 K, the spinel-like phases disappear entirely. Diffraction lines for CoO (JCPDS 75-393) appeared in the Co-673 sample, while diffraction lines for fcc Co (JCPDS 89-4307) and ZnO (JCPDS 89-1397) emerged in the Co-873 sample. The particle sizes for CoO and Co were calculated to be 9.2 and 14.5 nm, respectively. Figure 2 shows the TEM images for the LDH, the oxide, and the Co-T samples. The LDH possessed an irregular platelet shape with an average size of 80 nm. Upon calcination at 723 K, the platelets were transformed into a porous wormhole-like structure. This shape evolution is caused by the dehydration, dehydroxylation and decarbonation of the LDH precursor throughout the thermal treatment [23]. After hydrogen reduction, the Co-673 sample preserved the wormhole-like morphology, but the Co-873 catalyst comprised particles with an average size of 16 nm. Figure 3 illustrates the H2-TPR profile for the oxide, which shows the typical sequential reduction of Co3O4. The hydrogen consumption peak at 473–673 K represents the reduction of Co3O4 to CoO while the broad reduction peak at 673–1073 K corresponds to the subsequent reduction of CoO to Co. The levels of hydrogen consumed were 25.6 and 74.4%, respectively, reaffirming the sequential reduction pattern of Co3O4 [24]. This is further verified by the XPS spectra for Co2p in the samples. As shown in Fig. 4, binding energies of 779.7 eV (2p1/3) and 794.6 eV (2p3/2) were measured, with a spin–orbit separation of 14.9 eV, which is characteristic of Co3O4 [25]. The two shake-up satellites with spin–orbit separations of 6.0 eV (T1) and 10 eV (T2) were assigned to Co2? and Co3?, respectively [26]. The weaker intensity for the T1 peak suggests that the cobalt species are primarily in a trivalent state. After hydrogen reduction at 673 K, the binding energies shifted to higher values (781.0 eV for Co 2p3/2 and 796.9 eV for Co 2p1/2) and the shake-up satellite T1 intensified considerably. This reflects the Co3? species being partially reduced to Co2? [27]. This result is in accordance with the XRD and H2-TPR measurements where cobalt species were present as CoO in the Co-673 sample and Co in the Co-873 catalyst.
123
X. Guo et al.
3.2 Glycerol Hydrogenolysis Table 1 summarizes the reaction results of glycerol hydrogenolysis over the Co-T catalysts. For the Co-673 catalyst, the conversion of glycerol was 67.7% and the selectivity of 1,2-PDO approached 50.5%, giving a 1,2PDO yield of 34.2%. The Co-873 catalyst, containing Co nanoparticles of about 16 nm, provided a glycerol conversion of 70.6% and a 1,2-PDO selectivity of 57.8%, corresponding to a 40.8% yield. These results are comparable to those obtained for the typical Cu catalysts under similar conditions. Over a copper-chromite catalyst, the 1,2-PDO yield reached 21.7%, using a 20% glycerol aqueous solution at 473 K for 24 h [28]. On a Cu–ZnO catalyst, the 1,2-PDO yield approached 31.6% using a 10% glycerol aqueous solution at 473 K for 16 h [29]. The Co-T catalysts also exhibited promising stabilities in the recycling tests. For instance, in the second run for the Co-673 catalyst, the conversion of glycerol and the selectivity of 1,2-PDO maintained at 65.8 and 51.2%, respectively. For the Co-873 catalyst, the conversion decreased slightly from 70.6 to 63.0% upon reuse, but the selectivity of 1,2-PDO was maintained at 53.3%. It is noteworthy that the Co-673 catalyst containing CoO performed similarly to the Co-873 catalyst, which is composed of Co nanoparticles (about 16 nm). This may imply that the CoO species are reduced under the reaction conditions, since only Co nanoparticles are effective for glycerol hydrogenolysis. As shown in Figs. 2 and 5, the Co particle sizes in the Co-873 catalysts increased slightly, to about 20 and 25 nm, after the 1st and the 2nd runs, respectively. The situation for the Co-673 catalyst was more complicated. Diffraction lines for ZnO and fcc Co appeared in the used sample and the diffraction lines for CoO vanished completely. Co particles with an average size of 50 nm were dispersed over large ZnO aggregates (Fig. 2e), which maintained their size (48 nm) with less uniform distribution, after the recycle test (Fig. 2f). This can perhaps explain the highly reproducible catalytic performance in the 2nd run. However, the particle sizes for cobalt, as calculated from the XRD profiles (Fig. 5), were only 20 nm, which is much smaller than that measured from the TEM images. This discrepancy in Co particle size may arise from the calculation from XRD pattern reflecting the average crystallite size. In contrast, the TEM images give the size of individual particles that are usually the agglomeration of crystallites. To elucidate the structural variation of the Co-673 catalyst, over the course of the reaction, controlled tests were undertaken. As shown in Fig. 6, diffraction lines for CoO disappeared, while diffraction lines for metallic cobalt emerged, in the sample used for glycerol hydrogenolysis at 473 K over 0.5 h, and the mean particle size was 18.2 nm.
Glycerol Hydrogenolysis over Co Catalysts
(a)
(b)
(c)
(d)
60 daverage= 16 nm
Counts
40
20
0 5
10
15
20
25
30
Particles size (nm)
20
(e)
80
(f)
d average= 50 nm
daverage= 48 nm
Counts
Counts
60
40
10
20
0
0 0
20
40
60
80
100
20
120
60
(g)
60
80
100
40
(h)
daverage= 20 nm
50
40
Particles size (nm)
Particles size (nm)
daverage = 25 nm
30
Counts
40
Counts
Fig. 2 TEM images for a the LDH, b the oxide, c the Co-673 and, d the Co-873 catalysts, e–f the used Co-673 samples after the 1st and 2nd runs, and g–h the used Co-873 samples after the 1st and 2nd runs. The insets are the size distributions for the particles
1461
30 20
20
10 10 0
0 10
15
20
25
30
Particles size (nm)
35
10
20
30
40
Particles size (nm)
123
1462
X. Guo et al.
ZnO (89-1397) fcc Co (89-4307)
Intensity (a.u.)
Intensity (a.u.)
(d)
(c) (b)
(a) 273
473
673
873
1073
Temperature (K) 20
30
40
Fig. 3 H2-TPR profile for the oxide
50
60
70
2 Theta (degree) Fig. 5 XRD patterns for the used Co-673 catalyst after the 1st (a) and the 2nd (b) runs and the Co-873 catalyst after the 1st (c) and the 2nd (d) runs
T2
Intensity (a.u.)
(b)
*
T1
*
2p3/2 Intensity (a.u.)
2p1/2
(a)
810
800
790
ZnO (89-1397) Spinels fcc Co (89-4307) CoO (75-393)
*
(c)
*
* *
(b)
780
Binding Energy (eV) Fig. 4 The XPS core-level spectra for Co 2p of (a) the oxide and (b) the Co-673 catalyst
(a) 20
30
40
50
60
70
80
2 Theta (degree) Table 1 Glycerol hydrogenolysis over the Co-T catalysts Catalyst
Run
Conv. (%)
Selectivity (Cmol%) Ethanol
Co-673 Co-873
1,2-PDO
EG
Others
1st
67.7
4.6
50.5
17.7
23.8
2nd
65.8
5.3
51.3
22.3
13.0
1st
70.6
4.5
57.8
21.0
11.0
2nd
63.0
3.9
53.3
22.1
12.9
Reaction conditions: 40 g 10 wt% glycerol solution, 0.3 g catalyst, 473 K, 2.0 MPa H2, 12 h
123
Fig. 6 The XRD profiles for the samples obtained by treating the asprepared Co-673 catalyst at 473 K, in a 40 g 10 wt % glycerol solution, 2.0 MPa H2, 0.5 h, b 40 g pure water, 2.0 MPa H2, 0.5 h, c 40 g 10 wt % glycerol solution, 2.0 MPa N2, 12 h
Distinct diffraction lines for ZnO were also observed. When the Co-673 sample was treated with H2 in the absence of glycerol (in pure water) for 0.5 h, the particle size was 13.3 nm and diffraction lines for Co appeared with relatively weakened intensities. When the Co-673
Glycerol Hydrogenolysis over Co Catalysts
sample was treated in glycerol aqueous solution, under N2 atmosphere instead of hydrogen, the CoO species remained unchanged, even after 12 h. The particle size was only 9.4 nm compared to the reduced sample in Fig 1c. This confirms that the CoO species in the as-prepared Co-673 catalyst was readily reduced to Co particles by H2 under glycerol hydrogenolysis conditions, and that glycerol significantly promoted the reduction process. This could be related to the facilitated dissolution of H2 in glycerol solution rather than in pure water as reported by Gandarias et al. [30].
4 Conclusion Layered double Co–Zn–Al hydroxide is efficient precursor in preparing active and stable Co catalysts for the hydrogenolysis of glycerol in aqueous media. Upon reduction with hydrogen at 673 K, the catalyst mainly contains CoO species which are further reduced to Co nanoparticles, of about 50 nm, under the reaction conditions. When reduced with hydrogen at 873 K, the resulting catalyst, consisting of Co nanoparticles of 16 nm, allowed a glycerol conversion of 70.6% and a 1,2-PDO selectivity of 57.8%. Moreover, the Co particles remained stable upon recycling, with no obvious variations in conversion and selectivity.
References 1. Brandner A, Lehnert K, Bienholz A, Lucas M, Claus P (2009) Top Catal 52:278 2. Zhou CHC, Beltramini JN, Fan YX, Lu GQM (2008) Chem Soc Rev 37:527 3. Wang S, Liu HC (2007) Catal Lett 117:62 4. Liu QY, Guo XH, Chen JL, Shen WJ (2008) Nano Technol 19:365608 5. Yu WQ, Zhao J, Ma H, Miao H, Song Q, Xu J (2010) Appl Catal A 383:73
1463 6. Miyazawa T, Kusunoki Y, Kunimori K, Tomishige K (2006) J Catal 240:213 7. Maris EP, Davis RJ (2007) J Catal 249:328 8. Shinmi Y, Koso S, Kubota T, Nakagawa Y, Tomishige K (2010) Appl Catal B 94:318 9. Nakagawa Y, Shinmi Y, Koso S, Tomishige K (2010) J Catal 272:191 10. Bienholz A, Schwab F, Claus P (2010) Green Chem 12:290 11. Guo LY, Zhou JX, Mao JB, Guo XW, Zhang SG (2009) Appl Catal A 367:93 12. Guo XH, Li Y, Shi RJ, Liu QY, Zhan ES, Shen WJ (2009) Appl Catal A 371:108 13. Mane RB, Hengne AM, Ghalwadkar AA, Vijayanand S, Mohite PH, Potdar HS, Rode CV (2010) Catal Lett 135:141 14. Roy D, Subramaniam B, Chaudhari RV (2010) Catal Today 156:31 15. Balaraju M, Rekha V, Devi BLAP, Prasad RBN, Prasad PSS, Lingaiah N (2010) Appl Catal A 384:107 16. Bienholz A, Blume R, Knop-Gericke A, Girgsdies F, Behrens M, Claus P (2011) J Phys Chem C 115:999 17. Zhang FZ, Xiang X, Li F, Duan X (2008) Catal Surv Asia 12:253 18. Meher LC, Gopinath R, Naik SN, Dalai AK (2009) Ind Eng Chem Res 48:1840 19. Debecker DP, Gaigneaux EM, Busca G (2009) Chem Eur J 15:3920 20. Benito P, Herrero M, Labajos FM, Rives V, Royo C, Latorre N, Monzon A (2009) Chem Eng J 149:455 21. Jing FL, Zhang YY, Luo SZ, Chu W, Qian WZ (2010) Appl Clay Sci 48:203 22. Velu S, Suzuki K, Hashimoto S, Satoh N, Ohashi F, Tomura S (2001) J Mater Chem 11:2049 23. Herrero M, Benito P, Labajos FM, Rives V (2007) Catal Today 128:129 24. Ribet S, Tichit D, Coq B, Ducourant B, Morato F (1999) J Solid State Chem 142:382 25. Oku M, Hirokawa K (1976) J Electron Spectrosc Rel Phenom 8:475 26. Wei WF, Chen WX, Ivey DG (2008) Chem Mater 20:1941 27. Karamat S, Ke C, Tan TL, Zhou W, Lee P, Rawat RS (2009) Appl Surf Sci 255:4814 28. Dasari MA, Kiatsimkul PP, Sutterlin WR, Suppes GJ (2005) Appl Catal A 281:225 29. Balaraju M, Rekha V, Sai Prasad PS, Prasad RBN, Lingaiah N (2008) Catal Lett 126:119 30. Gandarias I, Arias PL, Requies J, Gemez MB, Fierro JLG (2010) Appl Catal B 97:248
123