- Email: [email protected]
Effect of synthesis route of mesoporous zirconia based Ni catalysts on coke minimization in conversion of biogas to synthesis gas
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 3 2 1 7 e3 2 2 8
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/he
Effect of synthesis route of mesoporous zirconia based Ni catalysts on coke minimization in conversion of biogas to synthesis gas Hande Mustu a, Sena Yasyerli a, Nail Yasyerli a, Gulsen Dogu a,*, Timur Dogu b, Petar Djinovic c, Albin Pintar c a
Gazi University, Department of Chemical Engineering, Maltepe, 06570 Ankara, Turkey Middle East Technical University, Department of Chemical Engineering, Cankaya, 06800 Ankara, Turkey c Laboratory for Environmental Sciences and Engineering, National Institute of Chemistry, Hajdrihova 19, SI-1001 Ljubljana, Slovenia b
article info
abstract
Article history:
Mesoporous zirconia and nickel incorporated zirconia catalysts were prepared following
Received 10 November 2014
different routes. Synthesis of mesoporous zirconia and Ni incorporated zirconia with very
Received in revised form
narrow pore size distributions and high surface area was achieved. Ni incorporated mes-
4 January 2015
oporous zirconia materials showed high activity in carbon dioxide reforming of methane,
Accepted 7 January 2015
performed at 600 C. Coke formation during dry reforming was eliminated over the Ni
Available online 28 January 2015
incorporated zirconia catalyst prepared by the one-pot procedure, using Pluronic P123 as the surfactant. It was shown that Ni was very well distributed within this material with
Keywords:
cluster sizes smaller than the detection limit of XRD. This catalyst also showed highly
Biogas
stable catalytic performance. However, the catalysts prepared by the impregnation method
Syngas
showed higher activity but much higher coke formation than the catalyst prepared by the
Dry reforming
one-pot route.
Mesoporous zirconia
Copyright © 2015, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
Coke
Introduction Fast depletion of fossil resources and the related environmental concerns initiated significant research for the production of alternative fuels and petrochemicals from biowaste. Biogas contains large quantities of two of the most abundant greenhouse gases, namely methane and carbon dioxide. Catalytic reforming of methane with CO2 (Eq. (1)) is a promising method to convert biogas to synthesis gas [1e4].
CH4 þCO2 42CO þ 2H2
1
DHo298 ¼ 247 kJ mol
(1)
Synthesis gas, containing CO and H2 may then be converted to valuable chemicals through FischereTropsch synthesis, or may be used to produce methanol and/or dimethyl ether. Product distribution of CO2 reforming of methane and hence the ratio of CO to H2 in the product stream, are strongly influenced by the occurrence of number of side reactions, namely reverse water gas shift (RWGS) reaction (Eq. (2)),
* Corresponding author. Tel.: þ90 312 5823559. E-mail address: [email protected] (G. Dogu). http://dx.doi.org/10.1016/j.ijhydene.2015.01.023 0360-3199/Copyright © 2015, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
3218
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 3 2 1 7 e3 2 2 8
methanation (Eq. (3)) and steam reforming of methane (Eq. (4)). CO2 þH2 4CO þ H2 O
1
DHo298 ¼ 41:2 kJ mol
1
CO2 þ4H2 4CH4 þ 2H2 O DHo298 ¼ 164:9 kJ mol CH4 þH2 O4CO þ 3H2
DHo298
1
¼ 206 kJ mol
(2)
Synthesis of catalysts
(3)
Synthesis of mesoporous zirconia by different routes
(4)
Noble metals were reported to show high activity for dry reforming of methane [5e8]. Much cheaper Ni and Co based catalysts were also reported to show good activity for this reaction [9e16]. However, coke formation is the major problem of using Ni based catalysts in dry reforming of methane [14,15]. Coke formation through methane cracking (Eq. (5)) and/or Boudouard reaction (Eq. (6)) may cause catalyst deactivation, as well as clogging of the reactor. Coke formation at lower temperatures is mainly due to Boudouard reaction. However, endothermic decomposition reactions become more significant at higher temperatures. 1
CH4 /C þ 2H2
DHo298 ¼ 75:2 kJ mol
2CO/C þ CO2
DHo298 ¼ 173:0 kJ mol
(5) 1
Experimental
(6)
Coke formation was reported to be decreased by the modification of Ni based catalysts through incorporation of noble metals. As it was shown in our recent studies, performance of MCM-41 incorporated Ni based catalysts could be improved by the addition of Ru or Rh to the catalyst structure [3,11]. Another very important factor effecting coke formation is the type of support material used in the synthesis of Ni based catalytic materials [16,17]. Conventionally, g-alumina or zeolites [6,18] were used as the catalyst support for this reaction. It was reported in the literature that mesoporous catalyst supports, like MCM-41, were less susceptible to catalyst deactivation due to coke formation than the conventional microporous materials [19]. In a more recent publication of ours, it was reported that mesoporous alumina supported Ni based catalysts showed good catalytic performance and coke formation was significantly decreased by the modification of these catalysts by tungsten [15]. Zirconia is another attractive support material for Ni based catalysts. There are few studies for catalytic performances of zirconia modified catalysts in the recent literature [20e26]. Zirconia (ZrO2) presents three crystalline phases, which are monoclinic, tetragonal and cubic. It has excellent properties, such as high mechanical strength, high thermal stability, high fracture toughness and hardness. ZrO2 is very stable at high temperatures and is quite suitable to be used as a catalyst support for endothermic catalytic reactions, which are generally performed at high temperatures and require catalysts with high thermal stability [27,28]. In the present study, mesoporous zirconia supported Ni based catalysts were synthesized following different routes using different templates, and catalytic performances of these materials were tested in dry reforming of methane. Ni based catalysts were prepared by both impregnation and one-pot procedures and the best catalyst synthesis route for coke elimination was illustrated.
Zirconia can be synthesized by various methods, such as precipitation, hydrothermal synthesis, surfactant-assisted, co-precipitation and solegel [29e32]. Considering the diffusion resistance minimization advantages of mesoporous materials, zirconia catalyst supports were synthesized in the present study using different templates in order to obtain ordered mesoporous pore structures [33]. These mesoporous zirconia materials were synthesized following three different procedures. In the first two procedures, zirconia supports were synthesized using two different surfactants as the templates, namely, cetyl-trimethyl-ammonium bromide CTMABr (denoted as ZrO2eC) and Pluronic P123 (denoted as ZrO2eP). In the third procedure, mesoporous zirconia was prepared following an ammonia solution route. This material was denoted as ZrO2. For the synthesis of ZrO2eC, a hydrothermal route similar to the procedure described in a recent publication of ours, for the synthesis of MCM-41, was used [34]. In this procedure, cetyl-trimethyl-ammonium-bromide (CTMABr, C16H33(CH3)3NBr, Merck) was used as the surfactant. CTMABr was first dissolved in deionized water by continuous stirring at 40 C. Zirconium (IV)-oxynitrate-hydrate (ZrO(NO3)2$xH2O, Aldrich) was then added to this solution. This is a basic synthesis route and in order to have a solution pH of 11.5, NaOH was added drop-wise to the solution during mixing. Hydrothermal synthesis of mesoporous zirconia was then achieved in a Teflon lined stainless steel autoclave at 60 C for four days. The product was then filtered, washed with deionized water, dried and calcined at 600 C in a tubular furnace in a flow of dry air. In the case of synthesis of ZrO2eP, the procedure described by Rezai et al. [20] was used. In this method, Pluronic P123 ((EO)20(PO)70(EO)20, SigmaeAldrich) was used as the surfactant and 28e30% aqueous ammonia solution (Merck) was used as the precipitation agent. A solution containing zirconium (IV)-oxynitrate-hydrate (ZrO(NO3)2$xH2O, Aldrich) in deionized water was prepared and then P123 was added (P123:ZrO2 molar ratio ¼ 0.03), by continuous stirring. A precipitate was obtained by drop-wise addition of ammonia. The slurry was then refluxed at 88 C for 24 h with continuous stirring. The resultant mixture was then filtered, washed with deionized water and calcined at 600 C in a flow of dry air. For the synthesis of zirconia by the hydrothermal ammonia route (ZrO2), the procedure described by Djinovic et al. for the ceria/zirconia support was modified and used [21]. In this procedure, zirconium (IV)-oxynitrate-hydrate (ZrO(NO3)2$xH2O, Aldrich) was dissolved in deionized water. This solution was added into the 30% aqueous ammonia solution (Merck) by continuous stirring. The final mixture was placed into a Teflon lined stainless steel autoclave and aged at 120 C for 6 h. The final product was filtered, washed with deionized water and ethanol and then calcined.
3219
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 3 2 1 7 e3 2 2 8
140
(a) ZrO2 -C 5Ni@ZrO2 -C
100
(b) ZrO2-P 5Ni@ZrO2-P 5Ni-ZrO2-P
120
Volume Adsorbed, cc/g (STP)
Volume Adsorbed, cc/g (STP)
125
100
75
50
25
80 60 40 20
0
0
0.0
0.2 0.4 0.6 0.8 Relative Pressure [P/P0 ] 275
0.2 0.4 0.6 0.8 Relative Pressure [P/P0 ]
1.0
(c)
250 Volume Adsorbed, cc/g (STP)
0.0
1.0
ZrO2 5Ni@ZrO 2
225 200 175 150 125 100 75 50 25 0 0.0
0.2 0.4 0.6 0.8 Relative Pressure [P/P0 ]
1.0
Fig. 1 e N2 adsorption-desorption isotherms of (a) ZrO2eC and 5Ni@ZrO2eC, (b) ZrO2eP, 5Ni@ZrO2eP and 5NieZrO2eP, (c) ZrO2 and 5Ni@ZrO2.
Synthesis of Ni containing mesoporous zirconia catalysts Zirconia supported catalysts containing 5 wt% Ni were prepared by impregnation of Ni into mesoporous zirconia materials, which were synthesized by the procedures described in the previous section. In addition to the catalysts prepared by the impregnation procedure, another mesoporous zirconia based Ni incorporated catalyst was also synthesized by a onepot route. For the impregnation of Ni into mesoporous zirconia, a solution containing 0.1 M Ni was prepared by dissolving nickel (II) nitrate hexahydrate (Ni(NO3)2$6H2O) in deionized water. Ethanol was added to this solution to have an ethanol/water ratio of two. The support material (mesoporous zirconia) was then added to this solution. Mixing was continued for 48 h. Then, this suspension was evaporated to dryness at 80 C. Solid product was then calcined at 800 C in a flow of dry air. The catalysts synthesized by using ZrO2eC, ZrO2eP or ZrO2 as the support materials were denoted as 5Ni@ZrO2eC, 5Ni@ZrO2eP and 5Ni@ZrO2, respectively. Ni contents of all of the catalysts prepared in this work were adjusted as 5 wt%.
One-pot synthesis of Ni incorporated zirconia catalyst was achieved following a hydrothermal route using Pluronic P123 as the template. Synthesis procedure is quite similar to the procedure described in the previous section for the synthesis of ZrO2eP. However, in this case a solution of Ni salt (Ni(NO3)2$6H2O) was directly added to the solution containing zirconium(IV)-oxynitrate-hydrate and P123, before the hydrothermal synthesis step. Amount of Ni added to the solution was determined to achieve 5 wt% Ni in the final product. Catalyst obtained by this one-pot procedure is denoted as 5NieZrO2eP [33].
Characterization of catalysts Pore size distribution and surface area values of catalysts were analyzed by using N2 adsorption-desorption technique with a Quanta Chrome-Autosorb-1C Sorptometer. X-ray diffraction (XRD) patterns of fresh and used catalysts were measured using a Rigaku Ultima-IV instrument, with a CuKa radiation source (l ¼ 0.15406 nm). Characterization of the surface
3220
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 3 2 1 7 e3 2 2 8
1.8 1.6 1.4
dV/d(logd)
the presence of dry air, using Perkin Elmer Pyris and Setaram Labsys instruments, respectively. In order to determine the nature and morphology of carbon formed over catalyst surface, scanning electron microscope (SEM) and transmission electron microscopy (TEM) analysis of the used catalyst were also performed, using QUANTA 400F Field Emission and Jeol Jem-2100F 200 kV HRTEM instruments, respectively. Chemical compositions of the synthesized materials were evaluated by the ICP-MS analysis, using a Perkin Elmer DRCII instrument, as well as energy dispersive spectroscopy (EDS) analysis using a QUANTA 400F Field Emission instrument. Besides these characterization tests, which were performed in the Central Laboratory of Middle East Technical University, temperature programmed oxidation (TPO) analysis of the used catalyst 5NieZrO2eP, after a 50 h time-on-stream test was also performed in the laboratories of the National Institute of Chemistry, Slovenia.
ZrO2-C ZrO2-P ZrO2 5Ni-ZrO2-P
1.2 1.0 0.8 0.6 0.4 0.2 0.0 1
10 d, nm
100
Fig. 2 e Pore size distributions of ZrO2eC, ZrO2eP, ZrO2 and 5NieZrO2eP samples.
acidity of the synthesized materials was achieved by diffuse reflectance FTIR (DRIFTS) analysis of pyridine adsorbed samples, using the PerkineElmer Spectrum One FTIR instrument available at Middle East Technical University. For the characterization of Lewis and Brønsted acid sites, differences of the DRIFT spectra of pyridine adsorbed and non-adsorbed samples were used. To obtain information about the basic characteristics of the catalysts, temperature programmed desorption (CO2-TPD) measurements of the reduced materials were carried with a 20% CO2e80% He gas mixture, using a QuantaChrome Chembet 3000 instrument. Before the CO2 adsorption step, catalyst (100 mg) was degassed at 200 C for 1 h in He atmosphere (flowing at a rate of 20 ml/min at STP), to clean the surface. Then, the catalytic material was cooled to 50 C and CO2 adsorption was achieved by passing a stream of CO2eHe gas mixture over the catalytic material, at a flow rate of 25 ml/min, for 30 min. Subsequently, sample was purged under He flow at the same temperature. In the last step of CO2TPD analysis, catalytic material was heated at a heating rate of 10 C/min, from 50 C to 700 C, under He flow (25 ml/min). Amount and type of coke formed during dry reforming tests was analyzed by means of thermal gravimetric analysis (TGA) and differential thermal analysis (DTA) techniques, in
Catalytic activity tests Carbon dioxide reforming of methane was performed in a quartz flow reactor, which was placed into a tubular furnace. Pelletized catalysts were crushed to obtain particles of about 1 mm in size and 0.1 g of crushed catalyst was loaded into the center of quartz reactor by supporting with quartz wool. Catalysts were reduced with H2 gas for 3 h, prior to dry reforming tests. A feed mixture consisting of methane, carbon dioxide and argon (as an inert diluent) with CH4/CO2/Ar of 1/1/ 1 was used in these activity tests, which were performed at 600 C. Flow rate of the feed stream was adjusted to 60 ml/min, measured at room temperature. Composition of the reactor outlet stream was determined with a gas chromatograph (PerkineElmer Autosystem XL), which was connected on-line to the reactor outlet. This chromatograph was equipped with a Carbosphere column and a TCD detector. Fractional conversion values of CO2 and methane, as well as H2 and CO selectivities with respect to converted methane, were then evaluated. Selectivity values were defined as the ratios of moles of H2 or CO produced to the converted moles of CH4. All of these activity tests were performed for reaction periods extending over 4 h [33]. However, a time-on-stream activity test, extending up to 50 h, was also performed in the presence of 5NieZrO2eP catalyst, at the National Institute of Chemistry, Slovenia. In that experiment, 0.5 g of catalyst was mixed with 2.83 g of inert SiC and packed into the reactor. Total flow rate and the composition of the feed stream were 100 ml/min and CH4/CO2 ¼ 1/1, respectively.
Table 1 e Physical properties of pure zirconia and Ni incorporated zirconia catalysts prepared by different routes. Catalyst ZrO2eC 5Ni@ZrO2eC ZrO2eP 5Ni@ZrO2eP 5NieZrO2eP ZrO2 5Ni@ZrO2
Synthesis method
BET surface area (m2/g)
Average pore size (nm)
Pore volume (cc/g)
Hydrothermal (CTMABr) Impregnation Hydrothermal (P123) Impregnation One-pot Hydrothermal (NH3) Impregnation
75 31 192 55 142 153 41
9.0 11.3 3.5 5.1 3.7 7.5 24.3
0.19 0.15 0.16 0.14 0.17 0.38 0.32
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 3 2 1 7 e3 2 2 8
Results and discussion Characterization of catalysts Nitrogen adsorption-desorption analysis results of the synthesized materials were used to determine pore size distributions, pore volumes and BET surface area values. Fig. 1 shows N2 adsorption-desorption isotherms of pure zirconia materials prepared following different routes, as well as the isotherms of Ni incorporated catalysts. As shown in Fig. 1a, b and c, adsorption/desorption isotherms of all of these materials are similar to Type 4, according to IUPAC classification. A typical hysteresis loop was observed in these isotherms. In the case of mesoporous structures with long range order, sharp hysteresis loops with parallel adsorption and desorption branches are expected. The hysteresis loops observed for ZrO2eC and ZrO2 based materials (Fig. 1a and c) were distributed within a wide range of relative pressures of nitrogen, between 0.6 and 0.95. This indicated that the mesopores of these materials were not highly ordered and they were in a wide range of pore sizes. In the case of the materials prepared by using P123 as the surfactant (ZrO2eP, 5Ni@ZrO2eP
3221
and 5NieZrO2eP), the behavior of the adsorption/desorption isotherms were somewhat different than the isotherms of the other materials. In this case (Fig. 1b), desorption branches of the hysteresis loops were very sharp; however, adsorption branches were not as sharp. Much narrower pore size distributions are expected for these materials than the others. These types of hysteresis loops indicated ordered mesoporous structures with the presence of significant interconnectivity [35]. The isotherms of these materials showed some decrease of pore volume, after impregnation of Ni into the mesoporous zirconia supports. This is an indication of closure of some of the smaller pores by nickel clusters. Pore size distributions of the porous zirconia materials prepared by different routes also supported the conclusion that the materials prepared by using P123 as the surfactant had much narrower pore size distributions than the others (Fig. 2). BET surface area values, average mesopore diameters and pore volumes of all of the synthesized materials are presented in Table 1. Comparison of the results obtained for pure zirconia materials prepared by different procedures showed that ZrO2eP had the largest BET surface area (192 m2/g) and the smallest average pore diameter (3.5 nm) as compared to the corresponding values of ZrO2eC and ZrO2. Significant
Fig. 3 e XRD patterns of (a) ZrO2eC and 5Ni@ZrO2eC, (b) ZrO2eP, 5Ni@ZrO2eP and 5NieZrO2eP, (c) ZrO2 and 5Ni@ZrO2. Legend: T: Tetragonal-ZrO2, M: Monoclinic-ZrO2, C: Cubic-ZrO2, NiO: Nickel oxide, Ni: Nickel.
3222
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 3 2 1 7 e3 2 2 8
0.5
(a)
CH 4 Conversion
CO2 Conversion
5Ni@ZrO2-C
5Ni@ZrO2-C
5Ni@ZrO2
5Ni@ZrO2
Conversion
0.4
0.3
0.2
0.1
0.0 0 0.5
30
60
90 120 150 Time, min
(b)
Conversion
210
240
CO2 Conversion
CH 4 Conversion 0.4
180
5Ni@ZrO2-P
5Ni@ZrO2-P
5Ni-ZrO2-P
5Ni-ZrO2-P
0.3
0.2
corresponding to NiO were also observed in the XRD pattern of 5Ni@ZrO2. XRD patterns of calcined ZrO2eP and 5Ni@ZrO2eP are given in Fig. 3b. In the same figure, XRD pattern of 5NieZrO2eP, after the reduction step, is also given. All of these materials contain tetragonal and cubic phases of zirconia. No peaks corresponding to NiO were observed in the XRD pattern of calcined 5Ni@ZrO2eP. This indicated that nickel was very well dispersed within this mesoporous material and its cluster size was smaller than the detection limit of XRD. Such amorphous forms could remain undetected by the XRD instrument. To have small Ni clusters is a big advantage from the point of view of less coke formation and stability during dry reforming tests [36]. Higher coke deposition is expected as the cluster size of Ni becomes larger. The XRD pattern of the material prepared by the one-pot route (5NieZrO2eP) is also quite similar to the XRD pattern of ZrO2eP, indicating the formation of cubic and tetragonal phases (Fig. 3b). A very small peak corresponding to Ni was also observed at a 2q value of about 44.6 in the XRD pattern of reduced material, again indicating the presence of very small Ni clusters. ICP-MS analysis of the Ni incorporated mesoporous zirconia type materials showed that the Ni contents of all of these materials were very close to 5 wt%. For instance, for
0.1
4.0
0 30
60
90 120 Time, min
150
180
210
3.0
240
Fig. 4 e CH4 and CO2 fractional conversions of (a) 5Ni@ZrO2eC and 5Ni@ZrO2, (b) 5Ni@ZrO2eP and 5NieZrO2eP. Reaction temperature: 600 C, Ar/CH4/CO2: 1/ 1/1.
Selectivity
0
CO Selectivity
H 2 Selectivity
(a)
3.5
5Ni@ZrO2-C
5Ni@ZrO2-C
5Ni@ZrO2
5Ni@ZrO2
2.5 2.0 1.5 1.0 0.5 0.0 0 4.0
Selectivity
reduction of surface area values of these materials and some shift of average pore diameter to larger sizes were observed as a result of impregnation of Ni into these mesoporous support materials. This is an indication of closure of some of the smaller pores by the impregnated Ni clusters. The surface area of the Ni incorporated material prepared by the one-pot route (5NieZrO2eP) was also quite high (142 m2/g). It also has a very narrow pore size distribution with an average pore diameter of 3.7 nm. It was promising to observe that pore structure of mesoporous zirconia did not alter much when Ni was incorporated into the zirconia lattice following the one-pot route. XRD patterns of calcined ZrO2eC and 5Ni@ZrO2eC catalysts are shown in Fig. 3a. Peaks corresponding to monoclinic, tetragonal and cubic structure of zirconia were observed in these materials. In addition to these peaks, the main characteristic peaks of NiO structure were also observed at 2q values at 37.2 and 43.3 in the XRD pattern of calcined 5Ni@ZrO2eC. NiO cluster size was estimated as being smaller than 7 nm in this material, using the Scherrer equation. Fig. 3c showed the XRD patterns of ZrO2 and 5Ni@ZrO2 materials. Presence of monoclinic, tetragonal, and cubic zirconia phases were also indicated in the XRD patterns of these materials. Peaks
30
60
90
120 150 Time, min
H 2 Selectivity
(b)
180
210
240
CO Selectivity
3.5
5Ni@ZrO2-P
5Ni@ZrO2-P
3.0
5Ni-ZrO2-P
5Ni-ZrO2-P
2.5 2.0 1.5 1.0 0.5 0.0 0
30
60
90 120 Time, min
150
180
210
240
Fig. 5 e H2 and CO selectivities over (a) 5Ni@ZrO2eC and 5Ni@ZrO2, (b) 5Ni@ZrO2eP and 5NieZrO2eP. Reaction temperature: 600 C, Ar/CH4/CO2: 1/1/1.
3223
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 3 2 1 7 e3 2 2 8
Fig. 6 e (a) TGA profiles of used 5Ni@ZrO2eC, 5Ni@ZrO2eP, 5Ni@ZrO2 and 5NieZrO2eP catalysts, (b) DTA profile of used 5Ni@ZrO2 catalyst after 4 h reaction tests at 600 C. 5NieZrO2eP, which was synthesized by the one-pot route, ICP-MS analysis gave a Ni content of 4.9 wt%. Ni contents of the synthesized materials were also analyzed by the EDS technique. These analyses also gave Ni contents quite close to 5 wt%. For instance, EDS analyses of 5Ni@ZrO2eP and 5Ni@ZrO2 gave Ni contents of 5 wt% and 4.56 wt%, respectively. These results clearly showed that Ni was successfully incorporated into the zirconia lattice during the synthesis of these materials.
and 5NieZrO2eP are shown in Fig. 4b. With all of these catalysts, CO2 conversion values were higher than CH4 conversions. In fact, the ratio of CO2 conversion to CH4 conversion ranged between 1.4 and 1.7. This result is mainly due to the occurrence of reverse water gas shift reaction (Eq. (2)) together with dry reforming of methane (Eq. (1)). Comparison of conversion values obtained with different catalytic materials showed that the most active catalysts were 5Ni@ZrO2eC and 5Ni@ZrO2, while the lowest conversion values were obtained with the catalyst prepared by the one-pot route (5NieZrO2eP). In the case of this material, some of the Ni clusters may not be exposed to the surface but embedded within the zirconia lattice. Although the activity of the materials prepared by using P123 as the surfactant (5Ni@ZrO2eP and 5NieZrO2eP) were somewhat less than the other catalysts, these materials showed more stable catalytic performance. Hydrogen and CO selectivity values obtained with all of the catalysts synthesized in this work are shown in Fig. 5a and b. These selectivity values were evaluated based on moles of CH4 converted. According to the dry reforming reaction (Eq. (1)) both H2 and CO selectivity values should be 2.0. However, H2 selectivity values were less than two, while CO selectivities were higher than two, with all of these catalysts. This is again due to the contribution of reverse water gas shift reaction. The H2/CO ratio was about 0.55e0.60 with all of the catalytic materials synthesized in this work. In order to have more information about catalyst stabilities and coke formation, amounts of coke formed over these catalysts were determined by TGA analysis of the used materials, at the end of 4 h reaction tests (Fig. 6a). Weight loss observed after about 450 C in the TGA curves is due to oxidation of coke with air. Small weight loss values observed at lower temperatures are simply due to desorption of moisture from the catalyst surface. Results indicated the importance of the synthesis procedure on the coke resistance of the zirconia based Ni catalysts. The largest amount of coke was formed over 5Ni@ZrO2 (Table 2). In fact, this material was the least stable catalyst, showing significant decrease in CO2 and CH4 conversion values within the reaction period of 4 h. TGA analysis showed that most of deposited carbon was oxidized in a temperature range of 350e600 C. DTA profile of the used 5Ni@ZrO2 catalyst, which gave the highest coke formation, also supported that oxidation of coke took place mostly at temperatures lower than 600 C (Fig. 6b). DTA analysis gave a single broad exothermic peak having a maximum at about 500 C. Oxidation of carbon species at relatively low temperatures
Table 2 e CH4 and CO2 fractional conversions and amount of coke formed over the catalysts within a reaction period of 4 h. Reaction temperature: 600 C, Ar/CH4/CO2: 1/1/1. Catalysts
Catalytic activity tests for dry reforming of methane Variation of methane and carbon dioxide conversion values within a reaction period of 4 h is given in Fig. 4. Dry reforming test results obtained at a reaction temperature of 600 C with 5Ni@ZrO2eC and 5Ni@ZrO2 catalysts are given in Fig. 4a, while CO2 and CH4 conversion values obtained with 5Ni@ZrO2eP
5Ni@ZrO2eC 5Ni@ZrO2eP 5NieZrO2eP 5Ni@ZrO2
CH4 conversion
CO2 conversion
TGA weight loss due to coke oxidation (%)
0.24 0.15 0.12 0.17
0.34 0.25 0.20 0.29
14 4 None 38
3224
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 3 2 1 7 e3 2 2 8
Fig. 7 e (a) SEM and (b) TEM images of used 5Ni@ZrO2 catalyst.
could be attributed to the absence of graphitic carbon over the catalyst surface. As it was shown in our recent publication, graphitic carbon was expected to be oxidized at much higher temperatures (750e800 C) [15]. If both amorphous and graphitic carbonaceous species were present, a DTA profile giving two peaks at about 400e600 C and 750e800 C would be expected. In order to have better understanding of nature and morphology of carbonaceous species formed on the catalysts, further characterizations of the spent materials were performed using SEM and TEM techniques (Fig. 7). SEM image (Fig. 7a) of the used 5Ni@ZrO2 catalyst clearly showed the formation of carbon filaments at some places on the catalyst surface. This result indicated that most of the coke formed on the catalyst surface was in filamentous form. TEM images (Fig. 7b) of this sample also indicated the presence of carbon filaments, within a diameter range of 30e70 nm. Dark spots in the TEM images correspond to the nickel clusters. Coke formed over the catalysts which were prepared by using P123 as the surfactant, were quite low. In fact, with the catalyst prepared by the one-pot route (5NieZrO2eP), no coke
formation was observed. This was considered as a highly promising result. As it was discussed in the previous section, in the XRD pattern of this material Ni or NiO peaks were almost absent, indicating excellent dispersion of nickel clusters with cluster sizes less than the detection limit of XRD. However, for the other catalysts (5Ni@ZrO2 and 5Ni@ZrO2eC), XRD peaks corresponding to NiO were clearly observed. Apparently, Ni dispersion has a very important effect on coke deposition. In the catalysts containing larger nickel clusters, some of the Ni might even be deposited at the external surface. 5Ni@ZrO2 and 5Ni@ZrO2eC catalysts showed higher activity than the others. However, coke formation was also high over these catalysts. In the case of the catalysts containing well dispersed nickel clusters within the mesoporous lattice, activity was somewhat less. However, these materials were highly stable and coke formation was negligible. Importance of Ni dispersion on coke formation was also discussed by Wang et al. [36]. Besides the size of Ni clusters, surface acidity might also have some effect on coke formation. In order to test this possibility, FTIR analysis of pyridine adsorbed samples was
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 3 2 1 7 e3 2 2 8
performed. The bands observed at about 1445 cm1 and 1490 cm1 correspond to the Lewis acid sites of the synthesized materials (Fig. 8a). As shown in this figure, the intensities of these bands were the highest for 5Ni@ZrO2eC, which caused formation of large amount of coke. However, the intensities of these bands observed in the FTIR spectra of 5Ni@ZrO2 and 5NieZrO2eP were quite similar, although these
Fig. 8 e The study of acidity and basicitiy of reduced zirconia based catalysts (a) FTIR analysis of pyridine adsorbed samples (b) CO2-TPD profiles of samples.
3225
two materials had shown highly different coke formation behavior. There was no coke formation over 5NieZrO2eP, while coke formation was maximum over 5Ni@ZrO2. These results indicated that surface acidity of zirconia supported catalysts was not the most important factor for coke formation. The most important factor was considered as Ni cluster size and its dispersion within the mesopores of the catalyst. In order to have some information about the basicity of the synthesized catalysts, CO2-TPD analyses of the reduced catalysts were also carried out (Fig. 8b). As it was reported in the literature [37,38], CO2 desorption peaks located at temperature ranges of 50e200 C, 200e400 C, 400e650 C and at temperatures higher than 650 C were considered to be due to the presence of weak, intermediate, strong and very strong basic sites, respectively. As it is seen in Fig. 8b, all zirconia supported Ni catalysts have similar CO2-TPD profiles, i.e., the first small peak being located at around 100 C and the second broad peak located between 200 C and 600 C. These results indicated the presence of weak, intermediate and strong basic sites on the synthesized materials. However, the area of CO2 desorption peak of the material synthesized by the one-pot route (5NieZrO2eP) was higher than the desorption peaks observed with the impregnated materials, indicating the presence of more basic sites on 5NieZrO2eP. Presence of basic sites on the catalyst surface is expected to facilitate CO2 sorption during dry reforming of methane and has positive effect to minimize coke formation. In order to have more information about the structural stability of these catalysts, XRD analyses of the used catalysts were also made and compared with the XRD patterns of the fresh ones. As shown in Fig. 9a and b, no change was observed in the XRD patterns of 5Ni@ZrO2eP and 5NieZrO2eP samples after the reaction. However, in the case of 5Ni@ZrO2, some changes were observed with a new small peak at a 2q value of 26 . This peak corresponds to the formation of some crystalline carbon on the catalyst surface. As it was discussed in the previous section, coke formation was the highest over this catalyst. All the results obtained from TGA-DTA analysis, SEM-TEM images and XRD patterns of the used catalysts indicated the formation of mostly amorphous coke together with some filamentous carbon over the synthesized materials. In the case of 5Ni@ZrO2, formation of some crystalline carbon was also observed. As a result of the activity test experiments performed in a reaction period of 4 h, it was concluded that coke formation was negligible over the 5NieZrO2eP catalyst. In order to test the stability of this catalytic material for a longer reaction period, time-on-stream tests extending up to 50 h were performed at 800 C. These tests also showed very stable catalytic performance of this material (Fig. 10). As shown in this figure, CO mole fraction was higher than the mole fraction of H2 in the product stream, due to contribution of RWGS reaction. Methane mole fraction was very low, indicating close to complete conversion of it. In order to test whether there was any coke formation over this catalyst after the 50 h test period, temperature programmed oxidation (TPO) of the spent catalyst was carried out (Fig. 11) in the presence of dry air. TPO analysis showed slight increase of weight of the spent catalyst after 400 C. This was thought to be due to oxidation of Ni to
3226
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 3 2 1 7 e3 2 2 8
Fig. 9 e Comparison of XRD patterns of fresh and used catalysts: (a) 5NieZrO2eP, (b) 5Ni@ZrO2eP, (c) 5Ni@ZrO2. Legend: T: Tetragonal-ZrO2, M: Monoclinic-ZrO2, C: Cubic-ZrO2, Ni: Nickel, C: Carbon.
90
CO
80
H2
70
100
CH4
60 50
#TPO, %
Concentration of the product, %
100
40 30
95
5Ni-ZrO2-P
20 10 0
90
0
10
20
30
40
Time, h Fig. 10 e Time-on-stream activity test results of 5NieZrO2eP catalyst at 800 C.
50
0 100 200 300 400 500 600 700 800 900 Temperature, 0 C
Fig. 11 e TPO analysis of spent 5NieZrO2eP catalyst after 50 h time-on-stream test.
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 3 2 1 7 e3 2 2 8
NiO. Main conclusion reached from TPO analysis was that no coke formation took place over this catalyst during the timeon-stream test.
Conclusions It was demonstrated that mesoporous zirconia with a very narrow pore size distribution can be synthesized following a hydrothermal route, using Pluronic P123 as the surfactant. Synthesized materials using other procedures and surfactants had much wider pore size distributions. Synthesis of Ni incorporated mesoporous zirconia with a very narrow pore size distribution and high surface area was also achieved following a one-pot route. Nickel was shown to be very well dispersed within this material. Coke resistance of this material was also very high in dry reforming of methane. In fact, negligible coke formation was observed over 5NieZrO2eP, even during time-on-stream reaction tests, extending up to 50 h. This is a highly promising result. Catalysts containing larger clusters of Ni were shown to cause formation of large amount of coke, although the activities of these Ni impregnated materials were higher than the activity of 5NieZrO2eP. Results proved that mesoporous zirconia was a highly stable catalyst support for dry reforming of methane and the most stable catalytic performance was observed with the Ni incorporated mesoporous zirconia, which was prepared by the one-pot route, using P123 as the surfactant.
Acknowledgments TUBITAK (111M449) and collaboration with National Institute of Chemistry, Slovenia, as well as contributions of Dr. Huseyin Arbag and Dr. Ilja Gasan Osojnik Crnivec are gratefully acknowledged. The authors also thank the Central Laboratory of Middle East Technical University.
references
[1] Kambolis A, Matralis H, Trovarelli A, Papadopoulou Ch. Ni/ CeO2eZrO2 catalysts for the dry reforming of methane. Appl Catal A Gen 2010;377:16e26. [2] Saha B, Khan A, Ibrahim H, Idem R. Evaluating the performance of non-precious metal based catalysts for sulfur-tolerance during the dry reforming of biogas. Fuel 2014;120:202e17. [3] Arbag H, Yasyerli S, Yasyerli N, Dogu G. Activity and stability enhancement of Ni-MCM-41 catalysts by Rh incorporation for hydrogen from dry reforming of methane. Int J Hydrogen Energy 2010;35:2296e304. [4] Serrano-Lotina A, Daza L. Influence of the operating parameters over dry reforming of methane to syngas. Int J Hydrogen Energy 2014;39:4090e4. [5] Ozkara-Aydinoglu S, Ozensoy E, Aksoylu AE. The effect of impregnation strategy on methane dry reforming activity of Ce promoted Pt/ZrO2. Int J Hydrogen Energy 2009;34:9711e22.
3227
[6] Ghelamallah M, Granger P. Impact of barium and lanthanum incorporation to supported Pt and Rh on a-Al2O3 in the dry reforming of methane. Fuel 2012;97:269e76. [7] Garcı´a-Garcı´a FR, Soria MA, Mateos-Pedrero C, GuerreroRuiz A, Rodrı´guez-Ramos I, Li K. Dry reforming of methane using Pd-based membrane reactors fabricated from different substrates. J Memb Sci 2013;435:218e25. [8] Hou Z, Chen P, Fang H, Zheng X, Yashima T. Production of synthesis gas via methane reforming with CO2 on noble metal sand small amount of noble-(Rh-) promoted Ni catalysts. Int J Hydrogen Energy 2006;31:555e61. -Alonso D, Juan-Juan J, Illa n-Go mez MJ, Roma n[9] San-Jose Martı´nez MC. Ni, Co and bimetallic NieCo catalysts for dry reforming of methane. Appl Catal A Gen 2009;371:54. [10] Luisetto I, Tuti S, Bartolomeo ED. Co and Ni supported on CeO2 as selective bimetallic catalyst for dry reforming of methane. Int J Hydrogen Energy 2012;37:15992e9. [11] Yasyerli S, Filizgok S, Arbag H, Yasyerli N, Dogu G. Ru incorporated NieMCM-41 mesoporous catalysts for dry reforming of methane: effects of Mg addition, feed composition and temperature. Int J Hydrogen Energy 2011;36:4863e74. [12] Cheng J, Huang W. Effect of cobalt (nickel) content on the catalytic performance of molybdenum carbides in drymethane reforming. Fuel Process Technol 2010;91:185e93. [13] Nimwattanakul W, Luengnarumitchai A, Jitkarnka S. Potential of Ni supported on clinoptilolite catalysts for carbondioxide reforming of methane. Int J Hydrogen Energy 2006;31:93e100. [14] Sutthiumporn K, Maneerung T, Kathiraser Y, Kawi S. CO2 dry-reforming of methane over La0.8Sr0.2Ni0.8M0.2O perovskite (M ¼ Bi, Co, Cr, Cu, Fe): roles of lattice oxygen on CeH activation and carbon suppression. Int J Hydrogen Energy 2012;37:11195e207. [15] Arbag H, Yasyerli S, Yasyerli N, Dogu T, Dogu G. Coke minimization in dry reforming of methane by Ni based mesoporous alumina catalysts synthesized following different routes: effects of W and Mg. Top Catal 2013;56:1695e707. [16] Nagaoka K, Takanabe K, Aika K. Modification of Co/TiO2 for dry reforming of methane at 2 MPa by Pt, Ru or Ni. Appl Catal A Gen 2004;268:151e8. [17] Frontera P, Macario A, Aloise A, Antonucci PL, Giordano G, Nagy JB. Effect of support surface on methane dry-reforming catalyst preparation. Catal Today 2013;218e219:18e29. [18] Fakeeha AH, Khan WU, Al-Fatesh AS, Abasaeed AE. Stabilities of zeolite-supported Ni catalysts for dry reforming of methane. Chin J Catal 2013;34:764e8. [19] Damyanova S, Pawelec B, Arishtirova K, Fierro JLG, Sener C, Dogu T. MCM-41 supported PdeNi catalysts for dry reforming of methane. Appl Catal B: Env 2009;92:250e61. [20] Rezai M, Alavi SM, Sahebdelfar S, Bai P, Liu X, Yan Z. CO2 reforming of CH4 over nano crystalline zirconia-supported nickel catalysts. Appl Catal B: Env 2008;77:346e54. P, Osojnik Crnivec [21] Djinovic IG, Erjavec B, Pintar A. Influence of active metal loading and oxygen mobility on coke-free dry reforming of NieCo bimetallic catalysts. Appl Catal B: Env 2012;125:259e70. [22] Yao L, Zhu J, Peng X, Tong D, Hu C. Comparative study on the promotion effect of Mn and Zr on the stability of Ni/SiO2 catalyst for CO2 reforming of methane. Int J Hydrogen Energy 2013;38:7268e79. [23] Therdthianwong S, Therdthianwong A, Siangchin C, Yongprapat S. Synthesis gas production from dry reforming of methane over Ni/Al2O3 stabilized by ZrO2. Int J Hydrogen Energy 2008;33:991e9. [24] Barroso-Quiroga MM, Castro-Luna AE. Catalytic activity and effect of modifiers on Ni based catalysts for the dry reforming of methane. Int J Hydrogen Energy 2010;35:6052e6.
3228
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 3 2 1 7 e3 2 2 8
[25] Tao W, Cheng H, Yao W, Lu X, Zhu Q, Li G, et al. Syngas production by CO2 reforming of coke oven gas over Ni/ La2O3eZrO2 catalysts. Int J Hydrogen Energy 2014;39:18650e8. [26] Chen J, Yao C, Zhao Y, Jia P. Synthesis gas production from dry reforming of methane over Ce0.75Zr0.25O2-supported Ru catalysts. Int J Hydrogen Energy 2010;35:1630e42. [27] Goharshadi EK, Hadadian M. Effect of calcination temperature on structural, vibrational, optical, andrheological properties of zirconia nanoparticles. Ceram Int 2012;38:1771e7. [28] Chraska T, King AH, Berndt CC. On the size-dependent phase transformation in nano particulate zirconia. Mat Sci Eng 2000;A286:169e78. [29] Stoia M, Barvinschi P, Barbu-Tudoran L, Negrea A, Barvinschi F. Influence of thermal treatment on the formation of zirconia nano structured powder by thermal decomposition of different precursors. J Cryst Growth 2013;381:93e9. [30] McGuire NE, Kondamudi N, Petkovic LM, Ginosar DM. Effect of lanthan promoters on zirconia-based isosynthesis catalysts prepared by surfactant-assisted coprecipitation. Appl Catal A Gen 2012;429e430:59e66. [31] Santos V, Zeni M, Bergmann CP, Hohemberger JM. Correlation between thermal treatment and tetragonal/ monoclinic nano structured zirconia powder obtained by solgel process. Rev Adv Mater Sci 2009;17:62e70.
[32] Kumari L, Li WZ, Xu M, Leblanc RM, Wang DZ, Li Y, et al. Controlled of hydrothermal synthesis zirconium oxide nano structures and their optical properties. Cryst Growth Des 2009;9:3874e80. [33] Mustu H. New catalysts and activity tests for synthesis gas from fry reforming of methane reaction [MS thesis]. Turkish: Gazi University; Sept. 2014. [34] Ozaydin Z, Yasyerli S, Dogu G. Synthesis and activity comparison of copper incorporated MCM-41 type sorbents prepared by one-pot and impregnation procedures for H2S removal. Ind Eng Chem Res 2008;47:1035e42. [35] Ciftci A, Sezgi NA, Dogu T. Nafion-incorporated silicate structured nano composite mesoporous catalysts for dimethyl ether synthesis. Ind Eng Chem Res 2010;49:6753e62. [36] Wang N, Chu W, Zhang T, Zhao XS. Synthesis, characterization and catalytic performances of Ce-SBA-15 supported nickel catalysts for methane dry reforming to hydrogen and syngas. Int J Hydrogen Energy 2012;37:19e30. [37] Garcia V, Fernandez JJ, Ruiz W, Mondragon F, Moreno A. Effect of MgO addition on the basicity of Ni/ZrO2 and on its catalytic activity in carbon dioxide reforming of methane. Catal Commun 2009;11:240e6. [38] Naeem MA, Al-Fatesh AS, Fakeeha AH, Abasaeed A. Hydrogen production from methane dry reforming over nickel-based nanocatalysts using surfactant-assisted or polyol method. Int J Hydrogen Energy 2014;39:17009e23.
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/he
Effect of synthesis route of mesoporous zirconia based Ni catalysts on coke minimization in conversion of biogas to synthesis gas Hande Mustu a, Sena Yasyerli a, Nail Yasyerli a, Gulsen Dogu a,*, Timur Dogu b, Petar Djinovic c, Albin Pintar c a
Gazi University, Department of Chemical Engineering, Maltepe, 06570 Ankara, Turkey Middle East Technical University, Department of Chemical Engineering, Cankaya, 06800 Ankara, Turkey c Laboratory for Environmental Sciences and Engineering, National Institute of Chemistry, Hajdrihova 19, SI-1001 Ljubljana, Slovenia b
article info
abstract
Article history:
Mesoporous zirconia and nickel incorporated zirconia catalysts were prepared following
Received 10 November 2014
different routes. Synthesis of mesoporous zirconia and Ni incorporated zirconia with very
Received in revised form
narrow pore size distributions and high surface area was achieved. Ni incorporated mes-
4 January 2015
oporous zirconia materials showed high activity in carbon dioxide reforming of methane,
Accepted 7 January 2015
performed at 600 C. Coke formation during dry reforming was eliminated over the Ni
Available online 28 January 2015
incorporated zirconia catalyst prepared by the one-pot procedure, using Pluronic P123 as the surfactant. It was shown that Ni was very well distributed within this material with
Keywords:
cluster sizes smaller than the detection limit of XRD. This catalyst also showed highly
Biogas
stable catalytic performance. However, the catalysts prepared by the impregnation method
Syngas
showed higher activity but much higher coke formation than the catalyst prepared by the
Dry reforming
one-pot route.
Mesoporous zirconia
Copyright © 2015, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
Coke
Introduction Fast depletion of fossil resources and the related environmental concerns initiated significant research for the production of alternative fuels and petrochemicals from biowaste. Biogas contains large quantities of two of the most abundant greenhouse gases, namely methane and carbon dioxide. Catalytic reforming of methane with CO2 (Eq. (1)) is a promising method to convert biogas to synthesis gas [1e4].
CH4 þCO2 42CO þ 2H2
1
DHo298 ¼ 247 kJ mol
(1)
Synthesis gas, containing CO and H2 may then be converted to valuable chemicals through FischereTropsch synthesis, or may be used to produce methanol and/or dimethyl ether. Product distribution of CO2 reforming of methane and hence the ratio of CO to H2 in the product stream, are strongly influenced by the occurrence of number of side reactions, namely reverse water gas shift (RWGS) reaction (Eq. (2)),
* Corresponding author. Tel.: þ90 312 5823559. E-mail address: [email protected] (G. Dogu). http://dx.doi.org/10.1016/j.ijhydene.2015.01.023 0360-3199/Copyright © 2015, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
3218
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 3 2 1 7 e3 2 2 8
methanation (Eq. (3)) and steam reforming of methane (Eq. (4)). CO2 þH2 4CO þ H2 O
1
DHo298 ¼ 41:2 kJ mol
1
CO2 þ4H2 4CH4 þ 2H2 O DHo298 ¼ 164:9 kJ mol CH4 þH2 O4CO þ 3H2
DHo298
1
¼ 206 kJ mol
(2)
Synthesis of catalysts
(3)
Synthesis of mesoporous zirconia by different routes
(4)
Noble metals were reported to show high activity for dry reforming of methane [5e8]. Much cheaper Ni and Co based catalysts were also reported to show good activity for this reaction [9e16]. However, coke formation is the major problem of using Ni based catalysts in dry reforming of methane [14,15]. Coke formation through methane cracking (Eq. (5)) and/or Boudouard reaction (Eq. (6)) may cause catalyst deactivation, as well as clogging of the reactor. Coke formation at lower temperatures is mainly due to Boudouard reaction. However, endothermic decomposition reactions become more significant at higher temperatures. 1
CH4 /C þ 2H2
DHo298 ¼ 75:2 kJ mol
2CO/C þ CO2
DHo298 ¼ 173:0 kJ mol
(5) 1
Experimental
(6)
Coke formation was reported to be decreased by the modification of Ni based catalysts through incorporation of noble metals. As it was shown in our recent studies, performance of MCM-41 incorporated Ni based catalysts could be improved by the addition of Ru or Rh to the catalyst structure [3,11]. Another very important factor effecting coke formation is the type of support material used in the synthesis of Ni based catalytic materials [16,17]. Conventionally, g-alumina or zeolites [6,18] were used as the catalyst support for this reaction. It was reported in the literature that mesoporous catalyst supports, like MCM-41, were less susceptible to catalyst deactivation due to coke formation than the conventional microporous materials [19]. In a more recent publication of ours, it was reported that mesoporous alumina supported Ni based catalysts showed good catalytic performance and coke formation was significantly decreased by the modification of these catalysts by tungsten [15]. Zirconia is another attractive support material for Ni based catalysts. There are few studies for catalytic performances of zirconia modified catalysts in the recent literature [20e26]. Zirconia (ZrO2) presents three crystalline phases, which are monoclinic, tetragonal and cubic. It has excellent properties, such as high mechanical strength, high thermal stability, high fracture toughness and hardness. ZrO2 is very stable at high temperatures and is quite suitable to be used as a catalyst support for endothermic catalytic reactions, which are generally performed at high temperatures and require catalysts with high thermal stability [27,28]. In the present study, mesoporous zirconia supported Ni based catalysts were synthesized following different routes using different templates, and catalytic performances of these materials were tested in dry reforming of methane. Ni based catalysts were prepared by both impregnation and one-pot procedures and the best catalyst synthesis route for coke elimination was illustrated.
Zirconia can be synthesized by various methods, such as precipitation, hydrothermal synthesis, surfactant-assisted, co-precipitation and solegel [29e32]. Considering the diffusion resistance minimization advantages of mesoporous materials, zirconia catalyst supports were synthesized in the present study using different templates in order to obtain ordered mesoporous pore structures [33]. These mesoporous zirconia materials were synthesized following three different procedures. In the first two procedures, zirconia supports were synthesized using two different surfactants as the templates, namely, cetyl-trimethyl-ammonium bromide CTMABr (denoted as ZrO2eC) and Pluronic P123 (denoted as ZrO2eP). In the third procedure, mesoporous zirconia was prepared following an ammonia solution route. This material was denoted as ZrO2. For the synthesis of ZrO2eC, a hydrothermal route similar to the procedure described in a recent publication of ours, for the synthesis of MCM-41, was used [34]. In this procedure, cetyl-trimethyl-ammonium-bromide (CTMABr, C16H33(CH3)3NBr, Merck) was used as the surfactant. CTMABr was first dissolved in deionized water by continuous stirring at 40 C. Zirconium (IV)-oxynitrate-hydrate (ZrO(NO3)2$xH2O, Aldrich) was then added to this solution. This is a basic synthesis route and in order to have a solution pH of 11.5, NaOH was added drop-wise to the solution during mixing. Hydrothermal synthesis of mesoporous zirconia was then achieved in a Teflon lined stainless steel autoclave at 60 C for four days. The product was then filtered, washed with deionized water, dried and calcined at 600 C in a tubular furnace in a flow of dry air. In the case of synthesis of ZrO2eP, the procedure described by Rezai et al. [20] was used. In this method, Pluronic P123 ((EO)20(PO)70(EO)20, SigmaeAldrich) was used as the surfactant and 28e30% aqueous ammonia solution (Merck) was used as the precipitation agent. A solution containing zirconium (IV)-oxynitrate-hydrate (ZrO(NO3)2$xH2O, Aldrich) in deionized water was prepared and then P123 was added (P123:ZrO2 molar ratio ¼ 0.03), by continuous stirring. A precipitate was obtained by drop-wise addition of ammonia. The slurry was then refluxed at 88 C for 24 h with continuous stirring. The resultant mixture was then filtered, washed with deionized water and calcined at 600 C in a flow of dry air. For the synthesis of zirconia by the hydrothermal ammonia route (ZrO2), the procedure described by Djinovic et al. for the ceria/zirconia support was modified and used [21]. In this procedure, zirconium (IV)-oxynitrate-hydrate (ZrO(NO3)2$xH2O, Aldrich) was dissolved in deionized water. This solution was added into the 30% aqueous ammonia solution (Merck) by continuous stirring. The final mixture was placed into a Teflon lined stainless steel autoclave and aged at 120 C for 6 h. The final product was filtered, washed with deionized water and ethanol and then calcined.
3219
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 3 2 1 7 e3 2 2 8
140
(a) ZrO2 -C 5Ni@ZrO2 -C
100
(b) ZrO2-P 5Ni@ZrO2-P 5Ni-ZrO2-P
120
Volume Adsorbed, cc/g (STP)
Volume Adsorbed, cc/g (STP)
125
100
75
50
25
80 60 40 20
0
0
0.0
0.2 0.4 0.6 0.8 Relative Pressure [P/P0 ] 275
0.2 0.4 0.6 0.8 Relative Pressure [P/P0 ]
1.0
(c)
250 Volume Adsorbed, cc/g (STP)
0.0
1.0
ZrO2 5Ni@ZrO 2
225 200 175 150 125 100 75 50 25 0 0.0
0.2 0.4 0.6 0.8 Relative Pressure [P/P0 ]
1.0
Fig. 1 e N2 adsorption-desorption isotherms of (a) ZrO2eC and 5Ni@ZrO2eC, (b) ZrO2eP, 5Ni@ZrO2eP and 5NieZrO2eP, (c) ZrO2 and 5Ni@ZrO2.
Synthesis of Ni containing mesoporous zirconia catalysts Zirconia supported catalysts containing 5 wt% Ni were prepared by impregnation of Ni into mesoporous zirconia materials, which were synthesized by the procedures described in the previous section. In addition to the catalysts prepared by the impregnation procedure, another mesoporous zirconia based Ni incorporated catalyst was also synthesized by a onepot route. For the impregnation of Ni into mesoporous zirconia, a solution containing 0.1 M Ni was prepared by dissolving nickel (II) nitrate hexahydrate (Ni(NO3)2$6H2O) in deionized water. Ethanol was added to this solution to have an ethanol/water ratio of two. The support material (mesoporous zirconia) was then added to this solution. Mixing was continued for 48 h. Then, this suspension was evaporated to dryness at 80 C. Solid product was then calcined at 800 C in a flow of dry air. The catalysts synthesized by using ZrO2eC, ZrO2eP or ZrO2 as the support materials were denoted as 5Ni@ZrO2eC, 5Ni@ZrO2eP and 5Ni@ZrO2, respectively. Ni contents of all of the catalysts prepared in this work were adjusted as 5 wt%.
One-pot synthesis of Ni incorporated zirconia catalyst was achieved following a hydrothermal route using Pluronic P123 as the template. Synthesis procedure is quite similar to the procedure described in the previous section for the synthesis of ZrO2eP. However, in this case a solution of Ni salt (Ni(NO3)2$6H2O) was directly added to the solution containing zirconium(IV)-oxynitrate-hydrate and P123, before the hydrothermal synthesis step. Amount of Ni added to the solution was determined to achieve 5 wt% Ni in the final product. Catalyst obtained by this one-pot procedure is denoted as 5NieZrO2eP [33].
Characterization of catalysts Pore size distribution and surface area values of catalysts were analyzed by using N2 adsorption-desorption technique with a Quanta Chrome-Autosorb-1C Sorptometer. X-ray diffraction (XRD) patterns of fresh and used catalysts were measured using a Rigaku Ultima-IV instrument, with a CuKa radiation source (l ¼ 0.15406 nm). Characterization of the surface
3220
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 3 2 1 7 e3 2 2 8
1.8 1.6 1.4
dV/d(logd)
the presence of dry air, using Perkin Elmer Pyris and Setaram Labsys instruments, respectively. In order to determine the nature and morphology of carbon formed over catalyst surface, scanning electron microscope (SEM) and transmission electron microscopy (TEM) analysis of the used catalyst were also performed, using QUANTA 400F Field Emission and Jeol Jem-2100F 200 kV HRTEM instruments, respectively. Chemical compositions of the synthesized materials were evaluated by the ICP-MS analysis, using a Perkin Elmer DRCII instrument, as well as energy dispersive spectroscopy (EDS) analysis using a QUANTA 400F Field Emission instrument. Besides these characterization tests, which were performed in the Central Laboratory of Middle East Technical University, temperature programmed oxidation (TPO) analysis of the used catalyst 5NieZrO2eP, after a 50 h time-on-stream test was also performed in the laboratories of the National Institute of Chemistry, Slovenia.
ZrO2-C ZrO2-P ZrO2 5Ni-ZrO2-P
1.2 1.0 0.8 0.6 0.4 0.2 0.0 1
10 d, nm
100
Fig. 2 e Pore size distributions of ZrO2eC, ZrO2eP, ZrO2 and 5NieZrO2eP samples.
acidity of the synthesized materials was achieved by diffuse reflectance FTIR (DRIFTS) analysis of pyridine adsorbed samples, using the PerkineElmer Spectrum One FTIR instrument available at Middle East Technical University. For the characterization of Lewis and Brønsted acid sites, differences of the DRIFT spectra of pyridine adsorbed and non-adsorbed samples were used. To obtain information about the basic characteristics of the catalysts, temperature programmed desorption (CO2-TPD) measurements of the reduced materials were carried with a 20% CO2e80% He gas mixture, using a QuantaChrome Chembet 3000 instrument. Before the CO2 adsorption step, catalyst (100 mg) was degassed at 200 C for 1 h in He atmosphere (flowing at a rate of 20 ml/min at STP), to clean the surface. Then, the catalytic material was cooled to 50 C and CO2 adsorption was achieved by passing a stream of CO2eHe gas mixture over the catalytic material, at a flow rate of 25 ml/min, for 30 min. Subsequently, sample was purged under He flow at the same temperature. In the last step of CO2TPD analysis, catalytic material was heated at a heating rate of 10 C/min, from 50 C to 700 C, under He flow (25 ml/min). Amount and type of coke formed during dry reforming tests was analyzed by means of thermal gravimetric analysis (TGA) and differential thermal analysis (DTA) techniques, in
Catalytic activity tests Carbon dioxide reforming of methane was performed in a quartz flow reactor, which was placed into a tubular furnace. Pelletized catalysts were crushed to obtain particles of about 1 mm in size and 0.1 g of crushed catalyst was loaded into the center of quartz reactor by supporting with quartz wool. Catalysts were reduced with H2 gas for 3 h, prior to dry reforming tests. A feed mixture consisting of methane, carbon dioxide and argon (as an inert diluent) with CH4/CO2/Ar of 1/1/ 1 was used in these activity tests, which were performed at 600 C. Flow rate of the feed stream was adjusted to 60 ml/min, measured at room temperature. Composition of the reactor outlet stream was determined with a gas chromatograph (PerkineElmer Autosystem XL), which was connected on-line to the reactor outlet. This chromatograph was equipped with a Carbosphere column and a TCD detector. Fractional conversion values of CO2 and methane, as well as H2 and CO selectivities with respect to converted methane, were then evaluated. Selectivity values were defined as the ratios of moles of H2 or CO produced to the converted moles of CH4. All of these activity tests were performed for reaction periods extending over 4 h [33]. However, a time-on-stream activity test, extending up to 50 h, was also performed in the presence of 5NieZrO2eP catalyst, at the National Institute of Chemistry, Slovenia. In that experiment, 0.5 g of catalyst was mixed with 2.83 g of inert SiC and packed into the reactor. Total flow rate and the composition of the feed stream were 100 ml/min and CH4/CO2 ¼ 1/1, respectively.
Table 1 e Physical properties of pure zirconia and Ni incorporated zirconia catalysts prepared by different routes. Catalyst ZrO2eC 5Ni@ZrO2eC ZrO2eP 5Ni@ZrO2eP 5NieZrO2eP ZrO2 5Ni@ZrO2
Synthesis method
BET surface area (m2/g)
Average pore size (nm)
Pore volume (cc/g)
Hydrothermal (CTMABr) Impregnation Hydrothermal (P123) Impregnation One-pot Hydrothermal (NH3) Impregnation
75 31 192 55 142 153 41
9.0 11.3 3.5 5.1 3.7 7.5 24.3
0.19 0.15 0.16 0.14 0.17 0.38 0.32
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 3 2 1 7 e3 2 2 8
Results and discussion Characterization of catalysts Nitrogen adsorption-desorption analysis results of the synthesized materials were used to determine pore size distributions, pore volumes and BET surface area values. Fig. 1 shows N2 adsorption-desorption isotherms of pure zirconia materials prepared following different routes, as well as the isotherms of Ni incorporated catalysts. As shown in Fig. 1a, b and c, adsorption/desorption isotherms of all of these materials are similar to Type 4, according to IUPAC classification. A typical hysteresis loop was observed in these isotherms. In the case of mesoporous structures with long range order, sharp hysteresis loops with parallel adsorption and desorption branches are expected. The hysteresis loops observed for ZrO2eC and ZrO2 based materials (Fig. 1a and c) were distributed within a wide range of relative pressures of nitrogen, between 0.6 and 0.95. This indicated that the mesopores of these materials were not highly ordered and they were in a wide range of pore sizes. In the case of the materials prepared by using P123 as the surfactant (ZrO2eP, 5Ni@ZrO2eP
3221
and 5NieZrO2eP), the behavior of the adsorption/desorption isotherms were somewhat different than the isotherms of the other materials. In this case (Fig. 1b), desorption branches of the hysteresis loops were very sharp; however, adsorption branches were not as sharp. Much narrower pore size distributions are expected for these materials than the others. These types of hysteresis loops indicated ordered mesoporous structures with the presence of significant interconnectivity [35]. The isotherms of these materials showed some decrease of pore volume, after impregnation of Ni into the mesoporous zirconia supports. This is an indication of closure of some of the smaller pores by nickel clusters. Pore size distributions of the porous zirconia materials prepared by different routes also supported the conclusion that the materials prepared by using P123 as the surfactant had much narrower pore size distributions than the others (Fig. 2). BET surface area values, average mesopore diameters and pore volumes of all of the synthesized materials are presented in Table 1. Comparison of the results obtained for pure zirconia materials prepared by different procedures showed that ZrO2eP had the largest BET surface area (192 m2/g) and the smallest average pore diameter (3.5 nm) as compared to the corresponding values of ZrO2eC and ZrO2. Significant
Fig. 3 e XRD patterns of (a) ZrO2eC and 5Ni@ZrO2eC, (b) ZrO2eP, 5Ni@ZrO2eP and 5NieZrO2eP, (c) ZrO2 and 5Ni@ZrO2. Legend: T: Tetragonal-ZrO2, M: Monoclinic-ZrO2, C: Cubic-ZrO2, NiO: Nickel oxide, Ni: Nickel.
3222
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 3 2 1 7 e3 2 2 8
0.5
(a)
CH 4 Conversion
CO2 Conversion
5Ni@ZrO2-C
5Ni@ZrO2-C
5Ni@ZrO2
5Ni@ZrO2
Conversion
0.4
0.3
0.2
0.1
0.0 0 0.5
30
60
90 120 150 Time, min
(b)
Conversion
210
240
CO2 Conversion
CH 4 Conversion 0.4
180
5Ni@ZrO2-P
5Ni@ZrO2-P
5Ni-ZrO2-P
5Ni-ZrO2-P
0.3
0.2
corresponding to NiO were also observed in the XRD pattern of 5Ni@ZrO2. XRD patterns of calcined ZrO2eP and 5Ni@ZrO2eP are given in Fig. 3b. In the same figure, XRD pattern of 5NieZrO2eP, after the reduction step, is also given. All of these materials contain tetragonal and cubic phases of zirconia. No peaks corresponding to NiO were observed in the XRD pattern of calcined 5Ni@ZrO2eP. This indicated that nickel was very well dispersed within this mesoporous material and its cluster size was smaller than the detection limit of XRD. Such amorphous forms could remain undetected by the XRD instrument. To have small Ni clusters is a big advantage from the point of view of less coke formation and stability during dry reforming tests [36]. Higher coke deposition is expected as the cluster size of Ni becomes larger. The XRD pattern of the material prepared by the one-pot route (5NieZrO2eP) is also quite similar to the XRD pattern of ZrO2eP, indicating the formation of cubic and tetragonal phases (Fig. 3b). A very small peak corresponding to Ni was also observed at a 2q value of about 44.6 in the XRD pattern of reduced material, again indicating the presence of very small Ni clusters. ICP-MS analysis of the Ni incorporated mesoporous zirconia type materials showed that the Ni contents of all of these materials were very close to 5 wt%. For instance, for
0.1
4.0
0 30
60
90 120 Time, min
150
180
210
3.0
240
Fig. 4 e CH4 and CO2 fractional conversions of (a) 5Ni@ZrO2eC and 5Ni@ZrO2, (b) 5Ni@ZrO2eP and 5NieZrO2eP. Reaction temperature: 600 C, Ar/CH4/CO2: 1/ 1/1.
Selectivity
0
CO Selectivity
H 2 Selectivity
(a)
3.5
5Ni@ZrO2-C
5Ni@ZrO2-C
5Ni@ZrO2
5Ni@ZrO2
2.5 2.0 1.5 1.0 0.5 0.0 0 4.0
Selectivity
reduction of surface area values of these materials and some shift of average pore diameter to larger sizes were observed as a result of impregnation of Ni into these mesoporous support materials. This is an indication of closure of some of the smaller pores by the impregnated Ni clusters. The surface area of the Ni incorporated material prepared by the one-pot route (5NieZrO2eP) was also quite high (142 m2/g). It also has a very narrow pore size distribution with an average pore diameter of 3.7 nm. It was promising to observe that pore structure of mesoporous zirconia did not alter much when Ni was incorporated into the zirconia lattice following the one-pot route. XRD patterns of calcined ZrO2eC and 5Ni@ZrO2eC catalysts are shown in Fig. 3a. Peaks corresponding to monoclinic, tetragonal and cubic structure of zirconia were observed in these materials. In addition to these peaks, the main characteristic peaks of NiO structure were also observed at 2q values at 37.2 and 43.3 in the XRD pattern of calcined 5Ni@ZrO2eC. NiO cluster size was estimated as being smaller than 7 nm in this material, using the Scherrer equation. Fig. 3c showed the XRD patterns of ZrO2 and 5Ni@ZrO2 materials. Presence of monoclinic, tetragonal, and cubic zirconia phases were also indicated in the XRD patterns of these materials. Peaks
30
60
90
120 150 Time, min
H 2 Selectivity
(b)
180
210
240
CO Selectivity
3.5
5Ni@ZrO2-P
5Ni@ZrO2-P
3.0
5Ni-ZrO2-P
5Ni-ZrO2-P
2.5 2.0 1.5 1.0 0.5 0.0 0
30
60
90 120 Time, min
150
180
210
240
Fig. 5 e H2 and CO selectivities over (a) 5Ni@ZrO2eC and 5Ni@ZrO2, (b) 5Ni@ZrO2eP and 5NieZrO2eP. Reaction temperature: 600 C, Ar/CH4/CO2: 1/1/1.
3223
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 3 2 1 7 e3 2 2 8
Fig. 6 e (a) TGA profiles of used 5Ni@ZrO2eC, 5Ni@ZrO2eP, 5Ni@ZrO2 and 5NieZrO2eP catalysts, (b) DTA profile of used 5Ni@ZrO2 catalyst after 4 h reaction tests at 600 C. 5NieZrO2eP, which was synthesized by the one-pot route, ICP-MS analysis gave a Ni content of 4.9 wt%. Ni contents of the synthesized materials were also analyzed by the EDS technique. These analyses also gave Ni contents quite close to 5 wt%. For instance, EDS analyses of 5Ni@ZrO2eP and 5Ni@ZrO2 gave Ni contents of 5 wt% and 4.56 wt%, respectively. These results clearly showed that Ni was successfully incorporated into the zirconia lattice during the synthesis of these materials.
and 5NieZrO2eP are shown in Fig. 4b. With all of these catalysts, CO2 conversion values were higher than CH4 conversions. In fact, the ratio of CO2 conversion to CH4 conversion ranged between 1.4 and 1.7. This result is mainly due to the occurrence of reverse water gas shift reaction (Eq. (2)) together with dry reforming of methane (Eq. (1)). Comparison of conversion values obtained with different catalytic materials showed that the most active catalysts were 5Ni@ZrO2eC and 5Ni@ZrO2, while the lowest conversion values were obtained with the catalyst prepared by the one-pot route (5NieZrO2eP). In the case of this material, some of the Ni clusters may not be exposed to the surface but embedded within the zirconia lattice. Although the activity of the materials prepared by using P123 as the surfactant (5Ni@ZrO2eP and 5NieZrO2eP) were somewhat less than the other catalysts, these materials showed more stable catalytic performance. Hydrogen and CO selectivity values obtained with all of the catalysts synthesized in this work are shown in Fig. 5a and b. These selectivity values were evaluated based on moles of CH4 converted. According to the dry reforming reaction (Eq. (1)) both H2 and CO selectivity values should be 2.0. However, H2 selectivity values were less than two, while CO selectivities were higher than two, with all of these catalysts. This is again due to the contribution of reverse water gas shift reaction. The H2/CO ratio was about 0.55e0.60 with all of the catalytic materials synthesized in this work. In order to have more information about catalyst stabilities and coke formation, amounts of coke formed over these catalysts were determined by TGA analysis of the used materials, at the end of 4 h reaction tests (Fig. 6a). Weight loss observed after about 450 C in the TGA curves is due to oxidation of coke with air. Small weight loss values observed at lower temperatures are simply due to desorption of moisture from the catalyst surface. Results indicated the importance of the synthesis procedure on the coke resistance of the zirconia based Ni catalysts. The largest amount of coke was formed over 5Ni@ZrO2 (Table 2). In fact, this material was the least stable catalyst, showing significant decrease in CO2 and CH4 conversion values within the reaction period of 4 h. TGA analysis showed that most of deposited carbon was oxidized in a temperature range of 350e600 C. DTA profile of the used 5Ni@ZrO2 catalyst, which gave the highest coke formation, also supported that oxidation of coke took place mostly at temperatures lower than 600 C (Fig. 6b). DTA analysis gave a single broad exothermic peak having a maximum at about 500 C. Oxidation of carbon species at relatively low temperatures
Table 2 e CH4 and CO2 fractional conversions and amount of coke formed over the catalysts within a reaction period of 4 h. Reaction temperature: 600 C, Ar/CH4/CO2: 1/1/1. Catalysts
Catalytic activity tests for dry reforming of methane Variation of methane and carbon dioxide conversion values within a reaction period of 4 h is given in Fig. 4. Dry reforming test results obtained at a reaction temperature of 600 C with 5Ni@ZrO2eC and 5Ni@ZrO2 catalysts are given in Fig. 4a, while CO2 and CH4 conversion values obtained with 5Ni@ZrO2eP
5Ni@ZrO2eC 5Ni@ZrO2eP 5NieZrO2eP 5Ni@ZrO2
CH4 conversion
CO2 conversion
TGA weight loss due to coke oxidation (%)
0.24 0.15 0.12 0.17
0.34 0.25 0.20 0.29
14 4 None 38
3224
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 3 2 1 7 e3 2 2 8
Fig. 7 e (a) SEM and (b) TEM images of used 5Ni@ZrO2 catalyst.
could be attributed to the absence of graphitic carbon over the catalyst surface. As it was shown in our recent publication, graphitic carbon was expected to be oxidized at much higher temperatures (750e800 C) [15]. If both amorphous and graphitic carbonaceous species were present, a DTA profile giving two peaks at about 400e600 C and 750e800 C would be expected. In order to have better understanding of nature and morphology of carbonaceous species formed on the catalysts, further characterizations of the spent materials were performed using SEM and TEM techniques (Fig. 7). SEM image (Fig. 7a) of the used 5Ni@ZrO2 catalyst clearly showed the formation of carbon filaments at some places on the catalyst surface. This result indicated that most of the coke formed on the catalyst surface was in filamentous form. TEM images (Fig. 7b) of this sample also indicated the presence of carbon filaments, within a diameter range of 30e70 nm. Dark spots in the TEM images correspond to the nickel clusters. Coke formed over the catalysts which were prepared by using P123 as the surfactant, were quite low. In fact, with the catalyst prepared by the one-pot route (5NieZrO2eP), no coke
formation was observed. This was considered as a highly promising result. As it was discussed in the previous section, in the XRD pattern of this material Ni or NiO peaks were almost absent, indicating excellent dispersion of nickel clusters with cluster sizes less than the detection limit of XRD. However, for the other catalysts (5Ni@ZrO2 and 5Ni@ZrO2eC), XRD peaks corresponding to NiO were clearly observed. Apparently, Ni dispersion has a very important effect on coke deposition. In the catalysts containing larger nickel clusters, some of the Ni might even be deposited at the external surface. 5Ni@ZrO2 and 5Ni@ZrO2eC catalysts showed higher activity than the others. However, coke formation was also high over these catalysts. In the case of the catalysts containing well dispersed nickel clusters within the mesoporous lattice, activity was somewhat less. However, these materials were highly stable and coke formation was negligible. Importance of Ni dispersion on coke formation was also discussed by Wang et al. [36]. Besides the size of Ni clusters, surface acidity might also have some effect on coke formation. In order to test this possibility, FTIR analysis of pyridine adsorbed samples was
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 3 2 1 7 e3 2 2 8
performed. The bands observed at about 1445 cm1 and 1490 cm1 correspond to the Lewis acid sites of the synthesized materials (Fig. 8a). As shown in this figure, the intensities of these bands were the highest for 5Ni@ZrO2eC, which caused formation of large amount of coke. However, the intensities of these bands observed in the FTIR spectra of 5Ni@ZrO2 and 5NieZrO2eP were quite similar, although these
Fig. 8 e The study of acidity and basicitiy of reduced zirconia based catalysts (a) FTIR analysis of pyridine adsorbed samples (b) CO2-TPD profiles of samples.
3225
two materials had shown highly different coke formation behavior. There was no coke formation over 5NieZrO2eP, while coke formation was maximum over 5Ni@ZrO2. These results indicated that surface acidity of zirconia supported catalysts was not the most important factor for coke formation. The most important factor was considered as Ni cluster size and its dispersion within the mesopores of the catalyst. In order to have some information about the basicity of the synthesized catalysts, CO2-TPD analyses of the reduced catalysts were also carried out (Fig. 8b). As it was reported in the literature [37,38], CO2 desorption peaks located at temperature ranges of 50e200 C, 200e400 C, 400e650 C and at temperatures higher than 650 C were considered to be due to the presence of weak, intermediate, strong and very strong basic sites, respectively. As it is seen in Fig. 8b, all zirconia supported Ni catalysts have similar CO2-TPD profiles, i.e., the first small peak being located at around 100 C and the second broad peak located between 200 C and 600 C. These results indicated the presence of weak, intermediate and strong basic sites on the synthesized materials. However, the area of CO2 desorption peak of the material synthesized by the one-pot route (5NieZrO2eP) was higher than the desorption peaks observed with the impregnated materials, indicating the presence of more basic sites on 5NieZrO2eP. Presence of basic sites on the catalyst surface is expected to facilitate CO2 sorption during dry reforming of methane and has positive effect to minimize coke formation. In order to have more information about the structural stability of these catalysts, XRD analyses of the used catalysts were also made and compared with the XRD patterns of the fresh ones. As shown in Fig. 9a and b, no change was observed in the XRD patterns of 5Ni@ZrO2eP and 5NieZrO2eP samples after the reaction. However, in the case of 5Ni@ZrO2, some changes were observed with a new small peak at a 2q value of 26 . This peak corresponds to the formation of some crystalline carbon on the catalyst surface. As it was discussed in the previous section, coke formation was the highest over this catalyst. All the results obtained from TGA-DTA analysis, SEM-TEM images and XRD patterns of the used catalysts indicated the formation of mostly amorphous coke together with some filamentous carbon over the synthesized materials. In the case of 5Ni@ZrO2, formation of some crystalline carbon was also observed. As a result of the activity test experiments performed in a reaction period of 4 h, it was concluded that coke formation was negligible over the 5NieZrO2eP catalyst. In order to test the stability of this catalytic material for a longer reaction period, time-on-stream tests extending up to 50 h were performed at 800 C. These tests also showed very stable catalytic performance of this material (Fig. 10). As shown in this figure, CO mole fraction was higher than the mole fraction of H2 in the product stream, due to contribution of RWGS reaction. Methane mole fraction was very low, indicating close to complete conversion of it. In order to test whether there was any coke formation over this catalyst after the 50 h test period, temperature programmed oxidation (TPO) of the spent catalyst was carried out (Fig. 11) in the presence of dry air. TPO analysis showed slight increase of weight of the spent catalyst after 400 C. This was thought to be due to oxidation of Ni to
3226
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 3 2 1 7 e3 2 2 8
Fig. 9 e Comparison of XRD patterns of fresh and used catalysts: (a) 5NieZrO2eP, (b) 5Ni@ZrO2eP, (c) 5Ni@ZrO2. Legend: T: Tetragonal-ZrO2, M: Monoclinic-ZrO2, C: Cubic-ZrO2, Ni: Nickel, C: Carbon.
90
CO
80
H2
70
100
CH4
60 50
#TPO, %
Concentration of the product, %
100
40 30
95
5Ni-ZrO2-P
20 10 0
90
0
10
20
30
40
Time, h Fig. 10 e Time-on-stream activity test results of 5NieZrO2eP catalyst at 800 C.
50
0 100 200 300 400 500 600 700 800 900 Temperature, 0 C
Fig. 11 e TPO analysis of spent 5NieZrO2eP catalyst after 50 h time-on-stream test.
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 3 2 1 7 e3 2 2 8
NiO. Main conclusion reached from TPO analysis was that no coke formation took place over this catalyst during the timeon-stream test.
Conclusions It was demonstrated that mesoporous zirconia with a very narrow pore size distribution can be synthesized following a hydrothermal route, using Pluronic P123 as the surfactant. Synthesized materials using other procedures and surfactants had much wider pore size distributions. Synthesis of Ni incorporated mesoporous zirconia with a very narrow pore size distribution and high surface area was also achieved following a one-pot route. Nickel was shown to be very well dispersed within this material. Coke resistance of this material was also very high in dry reforming of methane. In fact, negligible coke formation was observed over 5NieZrO2eP, even during time-on-stream reaction tests, extending up to 50 h. This is a highly promising result. Catalysts containing larger clusters of Ni were shown to cause formation of large amount of coke, although the activities of these Ni impregnated materials were higher than the activity of 5NieZrO2eP. Results proved that mesoporous zirconia was a highly stable catalyst support for dry reforming of methane and the most stable catalytic performance was observed with the Ni incorporated mesoporous zirconia, which was prepared by the one-pot route, using P123 as the surfactant.
Acknowledgments TUBITAK (111M449) and collaboration with National Institute of Chemistry, Slovenia, as well as contributions of Dr. Huseyin Arbag and Dr. Ilja Gasan Osojnik Crnivec are gratefully acknowledged. The authors also thank the Central Laboratory of Middle East Technical University.
references
[1] Kambolis A, Matralis H, Trovarelli A, Papadopoulou Ch. Ni/ CeO2eZrO2 catalysts for the dry reforming of methane. Appl Catal A Gen 2010;377:16e26. [2] Saha B, Khan A, Ibrahim H, Idem R. Evaluating the performance of non-precious metal based catalysts for sulfur-tolerance during the dry reforming of biogas. Fuel 2014;120:202e17. [3] Arbag H, Yasyerli S, Yasyerli N, Dogu G. Activity and stability enhancement of Ni-MCM-41 catalysts by Rh incorporation for hydrogen from dry reforming of methane. Int J Hydrogen Energy 2010;35:2296e304. [4] Serrano-Lotina A, Daza L. Influence of the operating parameters over dry reforming of methane to syngas. Int J Hydrogen Energy 2014;39:4090e4. [5] Ozkara-Aydinoglu S, Ozensoy E, Aksoylu AE. The effect of impregnation strategy on methane dry reforming activity of Ce promoted Pt/ZrO2. Int J Hydrogen Energy 2009;34:9711e22.
3227
[6] Ghelamallah M, Granger P. Impact of barium and lanthanum incorporation to supported Pt and Rh on a-Al2O3 in the dry reforming of methane. Fuel 2012;97:269e76. [7] Garcı´a-Garcı´a FR, Soria MA, Mateos-Pedrero C, GuerreroRuiz A, Rodrı´guez-Ramos I, Li K. Dry reforming of methane using Pd-based membrane reactors fabricated from different substrates. J Memb Sci 2013;435:218e25. [8] Hou Z, Chen P, Fang H, Zheng X, Yashima T. Production of synthesis gas via methane reforming with CO2 on noble metal sand small amount of noble-(Rh-) promoted Ni catalysts. Int J Hydrogen Energy 2006;31:555e61. -Alonso D, Juan-Juan J, Illa n-Go mez MJ, Roma n[9] San-Jose Martı´nez MC. Ni, Co and bimetallic NieCo catalysts for dry reforming of methane. Appl Catal A Gen 2009;371:54. [10] Luisetto I, Tuti S, Bartolomeo ED. Co and Ni supported on CeO2 as selective bimetallic catalyst for dry reforming of methane. Int J Hydrogen Energy 2012;37:15992e9. [11] Yasyerli S, Filizgok S, Arbag H, Yasyerli N, Dogu G. Ru incorporated NieMCM-41 mesoporous catalysts for dry reforming of methane: effects of Mg addition, feed composition and temperature. Int J Hydrogen Energy 2011;36:4863e74. [12] Cheng J, Huang W. Effect of cobalt (nickel) content on the catalytic performance of molybdenum carbides in drymethane reforming. Fuel Process Technol 2010;91:185e93. [13] Nimwattanakul W, Luengnarumitchai A, Jitkarnka S. Potential of Ni supported on clinoptilolite catalysts for carbondioxide reforming of methane. Int J Hydrogen Energy 2006;31:93e100. [14] Sutthiumporn K, Maneerung T, Kathiraser Y, Kawi S. CO2 dry-reforming of methane over La0.8Sr0.2Ni0.8M0.2O perovskite (M ¼ Bi, Co, Cr, Cu, Fe): roles of lattice oxygen on CeH activation and carbon suppression. Int J Hydrogen Energy 2012;37:11195e207. [15] Arbag H, Yasyerli S, Yasyerli N, Dogu T, Dogu G. Coke minimization in dry reforming of methane by Ni based mesoporous alumina catalysts synthesized following different routes: effects of W and Mg. Top Catal 2013;56:1695e707. [16] Nagaoka K, Takanabe K, Aika K. Modification of Co/TiO2 for dry reforming of methane at 2 MPa by Pt, Ru or Ni. Appl Catal A Gen 2004;268:151e8. [17] Frontera P, Macario A, Aloise A, Antonucci PL, Giordano G, Nagy JB. Effect of support surface on methane dry-reforming catalyst preparation. Catal Today 2013;218e219:18e29. [18] Fakeeha AH, Khan WU, Al-Fatesh AS, Abasaeed AE. Stabilities of zeolite-supported Ni catalysts for dry reforming of methane. Chin J Catal 2013;34:764e8. [19] Damyanova S, Pawelec B, Arishtirova K, Fierro JLG, Sener C, Dogu T. MCM-41 supported PdeNi catalysts for dry reforming of methane. Appl Catal B: Env 2009;92:250e61. [20] Rezai M, Alavi SM, Sahebdelfar S, Bai P, Liu X, Yan Z. CO2 reforming of CH4 over nano crystalline zirconia-supported nickel catalysts. Appl Catal B: Env 2008;77:346e54. P, Osojnik Crnivec [21] Djinovic IG, Erjavec B, Pintar A. Influence of active metal loading and oxygen mobility on coke-free dry reforming of NieCo bimetallic catalysts. Appl Catal B: Env 2012;125:259e70. [22] Yao L, Zhu J, Peng X, Tong D, Hu C. Comparative study on the promotion effect of Mn and Zr on the stability of Ni/SiO2 catalyst for CO2 reforming of methane. Int J Hydrogen Energy 2013;38:7268e79. [23] Therdthianwong S, Therdthianwong A, Siangchin C, Yongprapat S. Synthesis gas production from dry reforming of methane over Ni/Al2O3 stabilized by ZrO2. Int J Hydrogen Energy 2008;33:991e9. [24] Barroso-Quiroga MM, Castro-Luna AE. Catalytic activity and effect of modifiers on Ni based catalysts for the dry reforming of methane. Int J Hydrogen Energy 2010;35:6052e6.
3228
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 3 2 1 7 e3 2 2 8
[25] Tao W, Cheng H, Yao W, Lu X, Zhu Q, Li G, et al. Syngas production by CO2 reforming of coke oven gas over Ni/ La2O3eZrO2 catalysts. Int J Hydrogen Energy 2014;39:18650e8. [26] Chen J, Yao C, Zhao Y, Jia P. Synthesis gas production from dry reforming of methane over Ce0.75Zr0.25O2-supported Ru catalysts. Int J Hydrogen Energy 2010;35:1630e42. [27] Goharshadi EK, Hadadian M. Effect of calcination temperature on structural, vibrational, optical, andrheological properties of zirconia nanoparticles. Ceram Int 2012;38:1771e7. [28] Chraska T, King AH, Berndt CC. On the size-dependent phase transformation in nano particulate zirconia. Mat Sci Eng 2000;A286:169e78. [29] Stoia M, Barvinschi P, Barbu-Tudoran L, Negrea A, Barvinschi F. Influence of thermal treatment on the formation of zirconia nano structured powder by thermal decomposition of different precursors. J Cryst Growth 2013;381:93e9. [30] McGuire NE, Kondamudi N, Petkovic LM, Ginosar DM. Effect of lanthan promoters on zirconia-based isosynthesis catalysts prepared by surfactant-assisted coprecipitation. Appl Catal A Gen 2012;429e430:59e66. [31] Santos V, Zeni M, Bergmann CP, Hohemberger JM. Correlation between thermal treatment and tetragonal/ monoclinic nano structured zirconia powder obtained by solgel process. Rev Adv Mater Sci 2009;17:62e70.
[32] Kumari L, Li WZ, Xu M, Leblanc RM, Wang DZ, Li Y, et al. Controlled of hydrothermal synthesis zirconium oxide nano structures and their optical properties. Cryst Growth Des 2009;9:3874e80. [33] Mustu H. New catalysts and activity tests for synthesis gas from fry reforming of methane reaction [MS thesis]. Turkish: Gazi University; Sept. 2014. [34] Ozaydin Z, Yasyerli S, Dogu G. Synthesis and activity comparison of copper incorporated MCM-41 type sorbents prepared by one-pot and impregnation procedures for H2S removal. Ind Eng Chem Res 2008;47:1035e42. [35] Ciftci A, Sezgi NA, Dogu T. Nafion-incorporated silicate structured nano composite mesoporous catalysts for dimethyl ether synthesis. Ind Eng Chem Res 2010;49:6753e62. [36] Wang N, Chu W, Zhang T, Zhao XS. Synthesis, characterization and catalytic performances of Ce-SBA-15 supported nickel catalysts for methane dry reforming to hydrogen and syngas. Int J Hydrogen Energy 2012;37:19e30. [37] Garcia V, Fernandez JJ, Ruiz W, Mondragon F, Moreno A. Effect of MgO addition on the basicity of Ni/ZrO2 and on its catalytic activity in carbon dioxide reforming of methane. Catal Commun 2009;11:240e6. [38] Naeem MA, Al-Fatesh AS, Fakeeha AH, Abasaeed A. Hydrogen production from methane dry reforming over nickel-based nanocatalysts using surfactant-assisted or polyol method. Int J Hydrogen Energy 2014;39:17009e23.