Reforming and cracking of CH4 over Al2O3 supported Ni, Ni-Fe and Ni-Co catalysts

Fuel Processing Technology 156 (2017) 195–203

Contents lists available at ScienceDirect

Fuel Processing Technology journal homepage: www.elsevier.com/locate/fuproc

Research article

Reforming and cracking of CH4 over Al2O3 supported Ni, Ni-Fe and Ni-Co catalysts Koustuv Ray a, Siddhartha Sengupta b, Goutam Deo a,⁎ a b

Department of Chemical Engineering, Indian Institute of Technology Kanpur, 208 016, India Department of Chemical Engineering, Indian Institute of Technology (Indian School of Mines), Dhandbad 826004, India

a r t i c l e

i n f o

Article history: Received 6 May 2016 Received in revised form 1 November 2016 Accepted 4 November 2016 Available online xxxx Keywords: Reforming Cracking Ni-Fe/Al2O3 Ni-Co/Al2O3 CH4 CO2

a b s t r a c t Supported Ni, Ni-Fe and Fe catalysts of the same total metal loading and different Ni to Fe ratios were studied for the dry reforming and cracking of methane (CH4). The supported Ni-Fe catalysts containing Ni and Fe in the ratio of 3:1 (75Ni25Fe/Al2O3) was the most active for both reactions and was slightly more active than the supported Ni catalyst. The same Ni to Co ratio of 3:1 was present in the most active Ni-Co catalyst (75Ni25Co/Al2O3). Characterization of 75Ni25Fe/Al2O3 revealed the formation of Ni3Fe alloy, whose surface properties were different from the Ni1-xCox alloy present in 75Ni25Co/Al2O3. The presence of Ni based alloys of specific composition seemed responsible for the enhanced activity of 75Ni25Fe/Al2O3 and 75Ni25Co/Al2O3 relative to supported Ni catalyst for both the reactions. Furthermore, 75Ni25Co/Al2O3 was the most active catalyst for both reactions though deactivation occurred. In contrast, lower deactivation occurred with 75Ni25Fe/Al2O3. The turnover frequency during reforming and cracking were closely related for the supported Ni, Ni-Fe and Ni-Co catalysts. The higher activity of the 75Ni25Co/Al2O3 for the dry reforming reaction appeared to be due to the higher turnover frequency of this catalyst for the cracking reaction. Thus, the formation of alloys with specific composition, which improved the CH4 cracking capability, seems to be the key factor for determining the best catalytic performance for the reforming reaction over the promoted Ni catalysts. © 2016 Elsevier B.V. All rights reserved.

1. Introduction The reforming of methane (CH4) with CO2 had been extensively investigated during the past several years [1–7]. This process is often referred to as Dry Reforming of Methane (DRM), in contrast to the industrially relevant steam reforming of methane. Some of the interesting features leading to the extensive investigations are: (i) The DRM reaction produces synthesis gas with a H2/CO ratio more favourable for the production of valuable synthetic liquid fuels and oxygenates [1–4,8]. (ii) The DRM reaction has the lowest operating cost compared to other methane reforming processes [5]. (iii) The DRM reaction allows exploitation of natural gas resources with high CO2 content, thereby avoiding the expensive and intricate gas separation process [3–5]. (iv) The DRM reaction offers the use of biogas, a renewable resource containing CH4 (40–70%) and CO2 (30–60%) produced by anaerobic digestion of biomass [5,6]. (v) The DRM reaction offers the best solution for simultaneous utilization of these two greenhouse gases, CH4 and CO2 [2,5,7–9]. ⁎ Corresponding author. E-mail address: [email protected] (G. Deo).

http://dx.doi.org/10.1016/j.fuproc.2016.11.003 0378-3820/© 2016 Elsevier B.V. All rights reserved.

Considering the diminishing nature of petroleum oil reserves, efficient upgrading of CH4 is necessary and CO2 being a major environmental concern requires its effective utilization. Therefore, investigations involving the DRM reaction are very important from an industrial and environmental standpoint. The DRM reaction is catalytic and the commonly investigated catalysts for this reaction are supported noble metals and non-noble metals [2–4,10–12]. The noble metals include Pt, Pd, Ru, Rh and Ir, and these metals are very active and resistant toward deactivation by carbon deposition [1–4]. However, the high cost and limited availability of noble metals restrict their use. Amongst the non-noble metals supported Ni and Ni based catalysts are widely used due to their favourable activity, availability and cost [2,5,7]. The major difficulty associated with supported Ni and Ni based catalysts is carbon deposition, which render these catalysts unstable over long periods of time-on-stream (TOS) [1, 6,11]. To improve the catalytic activity and enhance the stability of supported Ni based catalysts different approaches have been undertaken. These approaches include support modification, addition of promoter and change in preparation method [1,5,12]. The effect of using a promoter is particularly interesting. Previous studies reveal that the catalytic activity and stability for the DRM reaction was improved by introducing a small amount of noble metal

196

K. Ray et al. / Fuel Processing Technology 156 (2017) 195–203

(Ru, Rh, Pt, or Pd) to supported Ni catalysts [1,3,13,14]. Significantly improved performance was also achieved by adding non-noble metals, such as Co or Fe, to supported Ni catalysts [15–18]. Furthermore, the specific ratio of Ni to Co and Ni to Fe had an effect on the catalytic activity and performance of the DRM reaction [1,7]. Closely associated with the DRM reaction is the cracking of CH4 (CRM). During DRM the CRM reaction was suggested to be an important step [3,4,19–21]. In the DRM reaction CH4 dissociates to various CHx species and the ultimate formation of Cads was substantiated [22–26]. The Cads is subsequently oxidized by CO2. Furthermore, the dissociation of CH4 to CH3ads and Hads species was proposed as the rate determining step in DRM [22]. The dissociation of CH4 was also suggested as the rate determining step in CRM [27]. Therefore, examining the CRM reaction over various catalysts is expected to assist in the understanding and development of catalysts for DRM. In addition to the importance during the DRM reaction the CRM reaction independently possesses significant industrial importance as it produces COx free hydrogen [27–29]. In a previous study co-precipitated Ni-Co, Ni-Fe, Ni-Cu, and Ni-Mn catalysts of fixed Ni-M composition were compared for the DRM reaction [2]. Based on initial screening additional studies were carried out only for the Ni-Co catalyst. However, in co-precipitated catalysts the surface concentrations of the metals may be different from the bulk. Furthermore, determining the reasons for the higher activity of the Ni-Co catalyst, compared to other Ni-metal systems, were not pursued. The increase in DRM and CRM activity by alloying Ni with Co was also shown in our previous study using co-impregnated catalysts [7]. However, the most active supported Ni-Co catalyst deactivated during the course of the reaction. Thus, developing cheap, active and stable Ni based alloy catalyst apart from Ni-Co is highly desirable. Another recent study showed that a particular Fe/Ni ratio in a series of bimetallic Fe-Ni/MgAl2O4 catalysts deactivated to a lesser extent and had a better catalytic activity than Ni alone [1]. Furthermore, Ni-Fe alloy is one of the most promising catalysts for CO2 methanation [30], where CO2 is also one of the key reactants in DRM reaction. Therefore, bimetallic Ni-Fe appears to be a potential and economically viable option for the DRM reaction. Our present work attempts to understand the effect of using supported Ni-Fe catalysts for the DRM reaction and comparing the activity and deactivation with supported Ni and the most active supported Ni-Co catalyst. To achieve the above objective supported Ni, Ni-Fe and Fe catalysts were synthesized, characterized and tested for the two reactions. Alumina (Al2O3) was chosen as the support. The total metal loading was constant and the amount of Ni and Fe in the Ni-Fe/Al2O3 catalysts was varied similarly to our previous study. The supported Ni, Ni-Fe and Fe catalysts were characterized by H2-Temperature programmed reduction (H2-TPR), Xray diffraction (XRD) and H2-Temperature programmed desorption (H2-TPD). The surface areas of the catalysts were also determined. All the catalysts were tested for the DRM under similar operating conditions to enable proper comparison. Furthermore, a moderate temperature was used to test the catalytic activity. The use of moderate temperatures was

intentional since carbon formation was favoured and catalyst stability can be analyzed. Under such conditions the conversions and yields were determined. Additionally, the reactivity results were compared with the best Ni-Co catalyst established in our previous study [7]. Thermogravimetric analysis (TGA) was also carried out to quantify the amount of carbon in some of the spent catalysts. Finally, the CRM reaction was carried over the catalysts and the conversions were compared with those observed for the DRM reaction. Based on the results and comparing with the best Ni-Co catalyst the reason for the increased catalytic activity of the Ni-M (M = Fe or Co) catalysts was proposed. This would enable us to develop robust Ni based catalysts that are active and stable for the DRM reaction at moderate and higher reaction temperatures. 2. Experimental 2.1. Catalyst synthesis Alumina supported Ni, Fe, and Ni-Fe catalysts were synthesized by the incipient wetness impregnation or co-impregnation method. A total metal loading of 15 wt.% was maintained for each supported catalysts. The Al2O3 (SASOL) support was pretreated with known amounts of water following previously published procedures [31]. The required amounts of an aqueous solution of Ni, and Fe precursors (nickel (II) nitrate hexahydrate (Sigma-Aldrich, 99.99%) and iron (III) nitrate nonahydrate (Sigma-Aldrich, 99.99%)) were thoroughly mixed with this pretreated support. The mixture was then dried and calcined at increasing temperatures and finally at 773 K for 6 h. A fresh batch of alumina supported Ni-Co and Co catalysts was synthesized based on the procedure described elsewhere [7]. The calcined catalysts were reduced in a reactor at 823 K for 4 h under flowing H2 stream prior to DRM and CRM reactions. The sample nomenclature and nominal compositions are given in Table 1. 2.2. Catalyst characterization 2.2.1. Surface area measurement The surface areas of the catalysts were measured by the BET method using N2 adsorption data at 77 K. The instrument used for surface area measurement was SMART SORB 92/93 surface area analyzer. A 30% N2-He gas mixture was used for adsorption. All the samples were degassed at 498 K for 8 h prior to the measurement. 2.2.2. Hydrogen-temperature programmed reduction (H2-TPR) The H2-TPR experiment was performed using an Altamira AMI-200 setup, which was equipped with a thermal conductivity detector (TCD). The reduction was analyzed in the temperature range of 323 to 1223 K using a constant flow (30 ml/min) of 10% H2-Ar gas mixture. Hydrogen consumption during reduction was measured by the TCD and a H2-TPR profile was obtained. The H2 consumption was obtained by integrating the area under the TPR profile and calibration amount used. This

Table 1 Sample nomenclature and characterization information of 15Ni/Al2O3, Ni-Fe/Al2O3 and 15Fe/Al2O3. Supported Ni-Co, 75Ni25Co/Al2O3 and 15Co/Al2O3 have been included for comparison. Samples were reduced at 823 K for 4 h prior to analysis. Sample nomenclature

15Ni/Al2O3 75Ni25Fe/Al2O3 50Ni50Fe/Al2O3 25Ni75Fe/Al2O3 15Fe/Al2O3 75Ni25Co/Al2O3 15Co/Al2O3 a b c

% Metal loading Ni Nominal

Co/Fe Nominal

15 11.25 7.5 3.75 – 11.25 –

– 3.75 7.5 11.25 15 3.75 15

Calcined catalysts. Reduced catalysts. Desorbed amount from Al2O3 has been subtracted.

Surface areaa (m2/g)

Metal-oxide reductiona (%)

Metal crystallite size (nm)b

H2 desorbed (μmole/g)b,c

164 162 167 169 162 163 150

75 73 55 40 23 80 78

34 15 17 16 40 19 23

82 101 170 508 376 62 39

K. Ray et al. / Fuel Processing Technology 156 (2017) 195–203

was further used to calculate the percentage of metal oxide reduced, which was referred to as metal oxide reduction (%). The calculation procedure was similar to those previously used [32]. 2.2.3. X-ray diffraction (XRD) of reduced catalysts The X-ray diffraction (XRD) patterns of the reduced catalysts were obtained on a PANalytical X'Pert PRO Diffractometer, using Ni filtered Kα radiation from a Cu target (λ = 1.541841 Å). The patterns were recorded for a 2θ range of 30–80°, a sweep of 3°min−1 and a time constant of 3 s. The Pearson's Crystal Data (PCD) library was used for phase identification. The crystallite size values of the reduced catalysts were determined from the highest intensity peak of Ni or Ni-M alloy or M by the X'Pert HighScore Plus software available with the above instrument. Prior to XRD measurements, the calcined catalysts were reduced at 823 K for 4 h in presence of hydrogen flow. 2.2.4. Hydrogen-temperature programmed desorption (H2-TPD) The chemisorbed H2 on the reduced catalysts was analyzed from the H2-TPD profile, which was obtained using the above mentioned AMI200 setup. It was assumed that the amount of chemisorbed and desorbed H2 were equal. Each catalyst was initially reduced in a down flow tubular reactor for 4 h at 823 K in a flow of H2. The reactor was then cooled to 323 K in the same H2 environment. About 0.22 g of reduced sample was loaded in the U-tube reactor of the AMI-200 setup for the H2-TPD studies. The sample was reduced again for 2 h at 823 K. The sample was then cooled to 313 K under the flow of H2. The temperature of the sample was held at 313 K for 0.5 h under the flow of Ar to remove weakly bound physisorbed hydrogen [7]. The temperature of the sample was then ramped to 773 K at 10 K/min under the flow of Ar and held at the same temperature for 0.5 h. During the temperature ramp the effluent gas was monitored by the TCD and a H2-TPD profile was obtained. The number of moles of desorbed H2 was calculated by integrating the area of the H2-TPD profile and using calibration pulses of premixed H2 in Ar of known quantities. A stoichiometry of 1 H atom to 1 surface metal site was considered for calculating the number of surface metal sites. The flow rate of different gases used during H2-TPD study was kept constant at 30 ml/min.

197

number of surface metal sites was determined from the H2-TPD data. The TOFDRM or TOFCRM for each catalyst was calculated using the following formula [7]:
TOFDRM or TOFCRM s−1 ¼ ðFlow rate CH4 =g cat Þ  ðCH4 conversion ð%Þ=100Þ  ð1=Ms Þ where, the flow rate of CH4 was expressed in mol/s, the weight of catalyst, gcat, was expressed in g and the surface metal sites of the catalyst, Ms., was expressed as mol/g. The value of Ms. was determined from H2-TPD studies mentioned below. The stability of the catalysts for the DRM and CRM reactions was calculated as the ratio of difference in activity to the initial activity using the following notation, Stability ¼ ðinitial activity−final activityÞ=initial activity where, initial activity is the conversion or TOF at 0.5 h and final activity is the conversion or TOF at 3.0 h of the reaction run. 2.3.2. Carbon measurement The amount of carbon deposited on the catalyst after 3 h of the DRM and CRM reaction was measured by TGA using a SDT Q600 unit (TA instruments). About 5 to 10 mg of the spent catalyst was heated from ambient temperature to 1273 K at 15 K/min in a flow of air. The carbon deposited on the spent catalyst was calculated from the weight loss data assuming negligible weight gain (if any) due to metal oxidation. Carbon deposited ðgc =gcat Þ ¼ ðwi –w f Þ=w f gc ¼ wt:of carbon gcat ¼ wt:of catalyst without carbon wi ¼ initial wt:of spent catalyst w f ¼ final wt:of spent catalyst

2.3. Reaction studies

3. Results and discussions

The synthesized catalysts were tested for the DRM and CRM reactions. The synthesized catalysts included the alumina supported Ni, Ni-Fe, Ni-Co, and Fe as mentioned above. Multiple runs were carried out on each catalyst so that error bars (in terms of standard deviations) could be provided.

3.1.1. Metal loading and Surface area analysis The surface area of the synthesized samples was determined and tabulated in Table 1 along with the nominal metal content in the synthesized catalysts. Data in Table 1 revealed that the surface area of the supported catalysts was lower than the support. The decrease in surface area was attributed to partial blocking of pores of the support by the deposited metals [34]. Furthermore, there was no specific trend in changes of surface area upon changing the Ni to Fe ratio. Similar results were reported previously [7,35].

2.3.1. Catalytic activity test for DRM and CRM reactions The DRM and CRM reactions were performed in a down flow tubular quartz reactor at 873 K and 1 atm, unless mentioned otherwise. A thermocouple was placed inside the reactor and just above the catalyst bed to measure the reactor temperature. The catalyst bed contained the supported catalyst and quartz particles of uniform size. Individual mass flow controllers were used to maintain the flow rates of the reactant gases, CO2, CH4 and N2. For the DRM reaction the volumetric ratio of CH4:CO2:N2 was maintained at 1:1:3 and for the CRM reaction the CH4:N2 ratio was 1:4. Such compositions were chosen so that the inlet partial pressure of CH4 for both reactions was the same. The gaseous products were analyzed using a gas chromatograph (GC) (Nucon 5765) equipped with a TCD and a Carbosphere column. Additional details of the reactor can be found in our earlier work [7]. The conversions of CH4 and CO2 were calculated using the previously reported formulae [7,33,34]. The turn over frequency (TOF) of CH4 for the DRM and CRM reactions, TOFDRM and TOFCRM, were also calculated for the different catalysts. The TOFDRM and TOFCRM were defined as the number of CH4 molecules converted over each surface metal site per second. The

3.1.2. Temperature programmed reduction analysis of calcined catalysts The H2-TPR profiles of 15Ni/Al2O3, Ni-Fe/Al2O3 and 15Fe/Al2O3were obtained and are shown in Fig. 1. Additionally, the data for 75Ni25Co/ Al2O3 and 15Co/Al2O3 catalysts were also obtained and included in the same figure for comparison. For 15Ni/Al2O3 (trace a) three reduction features were observed. The most intense peak at ~820 K was attributed to the reduction of NiO to metallic Ni [7,36]. The two shoulders at 735 K and 1040 K were assigned to the different levels of interaction of the NiO phase with the Al2O3 support in accordance with previous studies [35, 37]. The 15Co/Al2O3 sample (trace g) possessed two peaks at 723 K and 853 K, which were attributed to the reduction of Co3O4 to CoO and CoO to metallic Co [7,38]. The H2-TPR profile for 15Fe/Al2O3 (trace e) possessed three reduction peaks at 676 K, 880 K and 1118 K. The first peak centered at 676 K is ascribed to the reduction of α Fe2O3 to α Fe3O4. The second peak at ~ 880 K was of smaller intensity and showed overlapping features, which might be due to the simultaneous

198

K. Ray et al. / Fuel Processing Technology 156 (2017) 195–203

Fig. 1. H2-TPR profile of calcined Al2O3 supported Ni, Ni-Fe and Fe samples: (a) 15Ni/ Al2O3; (b) 75Ni25Fe/Al2O3; (c) 50Ni50Fe/Al2O3 (d) 25Ni75Fe/Al2O3 (e) 15Fe/Al2O3. Supported Ni-Co and Co catalysts e.g. (f) 75Ni25Co/Al2O3; (g) 15Co/Al2O3 have been included for comparison.

reduction of iron oxides into metallic iron. Similar arguments were proposed previously for SiO2 supported Fe catalyst [39]. The H2-TPR profile of the calcined 75Ni25Co/Al2O3 catalyst was different from 15Ni/Al2O3 and 15Co/Al2O3, and possessed three peaks at 620 K, 716 K and 840 K. Interestingly, the reduction of this sample starts at an earlier temperature compared to 15Ni/Al2O3 and 15Co/Al2O3 catalysts. The overlapping peaks observed for this catalyst suggested simultaneous reduction of the oxides of Ni and Co. The reduction features of the two Ni-Fe catalysts, 75Ni25Fe/Al2O3 and 50Ni50Fe/ Al2O3, were different from their monometallic counterparts. The reduction profile of 75Ni25Fe/Al2O3 revealed a single peak at about 678 K along with a small shoulder at 647 K. Relative to 15Ni/Al2O3 and 15Fe/ Al2O3 the reduction temperature for this Ni-Fe sample was shifted to lower temperatures. The reduction profile of 50Ni50Fe/Al2O3 catalyst was not very different from that of 75Ni25Fe/Al2O3 catalyst. The 25Ni75Fe/Al2O3 catalyst possessed three broad reduction features, which were similar to that of 15Fe/Al2O3, with slight differences in the positions of the peak maxima. Similar to the Al2O3 supported Ni-Co catalyst the shift to lower temperature reflected the ease in reducibility of the sample. Thus, the presence of Co or Fe decreased the reduction temperature of the surface metal oxide species. The metal oxide reduction (%) was determined from the H2-TPR studies and was tabulated in Table 1. The metal oxide reduction (%) of the calcined 15Ni/Al2O3 (75%) and 15Co/Al2O3 (78%) samples were similar, and significantly greater than that of the calcined 15Fe/Al2O3 (23%) sample. The metal oxide reduction (%) of the 75Ni25Co/Al2O3 and 75Ni25Fe/Al2O3 catalysts was 80% and 73%, respectively. The slightly lower reduction temperature of 75Ni25Fe/Al2O3 may be related to the lower metal oxide reduction (%) achieved for 15Fe/Al2O3. In fact, the metal oxide reduction (%) of the two other Ni-Fe catalysts were significantly lower than 75Ni25Fe/Al2O3. Although major differences in terms of metal oxide reduction (%) of the two 75Ni25M catalysts relative to pure Ni are not noticed, but shifting of reduction features to lower temperature suggests the beneficial role of substituting Ni with Co or Fe in the Ni based catalysts. Substituting Ni with Co or Fe in the Ni based catalyst facilitates the formation of the metal at lower temperatures. 3.1.3. X-ray diffraction patterns of the reduced catalysts The XRD patterns of reduced Al2O3 supported bimetallic catalysts were obtained and shown from 40° to 50° in Fig. 2. The XRD patterns of the reduced supported monometallic catalysts are also shown for comparison. The XRD patterns of all the catalysts and Al2O3 support are shown from 30° to 80° in the supplementary section as Fig. S1. The

Fig. 2. XRD patterns of reduced Al2O3 supported Ni, Ni-Fe and Fe samples: (a) 15Ni/Al2O3; (b) 75Ni25Fe/Al2O3; (c) 50Ni50Fe/Al2O3 (d) 25Ni75Fe/Al2O3 (e) 15Fe/Al2O3. Supported Ni-Co and Co catalysts e.g. (f) 75Ni25Co/Al2O3; (g) 15Co/Al2O3 have been included for comparison.

40° to 50° region was highlighted to show changes in the most intense Ni metal peak in the Ni-M catalysts. The XRD patterns of the 15Ni/Al2O3 and 15Co/Al2O3 samples revealed metallic Ni and Co peaks at 2θ of 44.52° and 44.3°. The two peaks were assigned to the (111) facet of metallic Ni and Co [40–42]. Nickel or cobalt oxide features were absent indicating significant reduction of the metal oxide phases. The 15Fe/Al2O3 catalyst revealed a metallic Fe peak at 44.73° in the highlighted region. Unlike the 15Ni/Al2O3 and 15Co/Al2O3 catalysts, peaks for the oxide phase (Fe3O4) were also observed in the XRD patterns of this catalyst, as shown in Fig. S1 (provided as Supplementary information), indicating incomplete reduction. The presence of Fe3O4 in the reduced 15Fe/ Al2O3 sample was consistent with the low metal oxide reduction (%) observed in the H2-TPR studies discussed above. The 75Ni25Fe/Al2O3 catalyst possessed a peak at 44.22°, which was slightly shifted from the metallic Ni peak at 44.52°. Such a shift was previously reported [35], and attributed to the formation Ni3Fe alloy [43]. A further shift in peak position was noticed for 50Ni50Fe/Al2O3 catalyst suggesting formation of a Ni-Fe alloy, whose chemical composition and structure would be different than the alloy present in 75Ni25Fe/Al2O3. Furthermore, the PCD library also confirmed such shifting in Ni-Fe alloys. The 2θ position of the crystalline phase (111) contained in this catalyst closely matched with those available [44], and the composition of such a phase was determined to be that of the Ni3Fe alloy. The 75Ni25Co/Al2O3 catalyst revealed a peak in between the 2θ locations of metallic Ni and Co. Similar observations were previously reported for supported Ni-Co catalysts with different loading [15,16,45]. In these previous studies the identified Ni-Co phase with the corresponding 2θ position was attributed to the formation of a homogeneous Ni-Co alloy. The 2θ positions of the alloy phase formed in this work were also consistent with Ni-Co alloys of the PCD Library [46,47]. The Ni-Co alloy formed in the present study could be of the type Ni1-xCox consistent with previous studies [45]. Thus, two different types of Ni-M alloy were evidenced from the XRD patterns of the Ni-M/Al2O3 catalysts with the following characteristic differences between them. The peak shift observed in 75Ni25Fe/ Al2O3 relative to 15Ni/Al2O3 and 15Fe/Al2O3 is non-monotonic and relatively larger compared to the monotonic and smaller shift observed for the 75Ni25Co/Al2O3 catalysts. The crystallite sizes of the reduced Al2O3 supported catalysts were determined from the XRD patterns by applying the Scherrer equation [7,35]. The most prominent XRD peak was used to calculate the crystallite size and the results are reported in Table 1. The results showed that the Ni metal in 15Ni/Al2O3 possessed a crystallite size of 34 nm, which

K. Ray et al. / Fuel Processing Technology 156 (2017) 195–203

199

was larger than that of Co metal in 15Co/Al2O3 (23 nm) and smaller than Fe metal in 15Fe/Al2O3 (40 nm). The crystallite sizes of the 75Ni25Co/Al2O3 and 75Ni25Fe/Al2O3 alloy catalysts were 19 and 15 nm respectively, which were smaller than the crystallite size of the supported Ni catalyst. The crystallite sizes of the other two Ni-Fe catalysts, 50Ni50Fe/Al2O3 and 25Ni75Fe/Al2O3, were also smaller than their monometallic counterparts. The calculated crystallite sizes of the metal and metal alloys in the present study were in the range of those reported in previous studies [39,48–50]. 3.1.4. H2-TPD of the reduced catalysts The H2-TPD profiles of the Al2O3 support and supported catalysts were obtained from 323 K to 773 K. The H2-TPD profile of Al2O3 was subtracted from the profile of the supported catalysts and the resulting profiles are compared in Fig. 3. The H2-TPD profiles of 15Ni/Al2O3, 75Ni25Co/Al2O3 and 15Co/Al2O3 catalysts shown in Fig. 3 revealed that the there was a slight monotonic shift of desorption temperature from 15Ni/Al2O3 (368 K) to 15Co/Al2O3 (382 K). Furthermore, the shape of the H2-TPD profile of the 75Ni25Co/Al2O3 catalyst and its monometallic counterparts were similar. The H2 desorption temperature for 15Fe/Al2O3, occurring at 509 K, was at a much higher temperature relative to 15Ni/Al2O3. A small shoulder at 426 K was also observed for 15Fe/Al2O3. Furthermore, the H2-TPD profile of the 75Ni25Fe/Al2O3 catalyst shown in Fig. 3 was noticeably different than pure Ni and Fe. Moreover, the H2-TPD profile of 75Ni25Fe/Al2O3 contained a single desorption peak at 401 K, shifted from the desorption peak of 15Ni/Al2O3. Therefore, the nature of surface sites present in the two alloy catalysts appears to be qualitatively different. The profile for 50Ni50Fe/Al2O3 catalyst remained similar to that of 75Ni25Fe/Al2O3. However, the 25Ni75Fe/Al2O3 catalyst contained two desorption peaks. The low temperature peak was of smaller intensity and might be associated with sites that predominantly contained Ni, whereas the high temperature peak might reflect the sites that predominantly contained Fe. The amount of chemisorbed H2 on the supported catalysts was determined from the H2-TPD studies and is reported in Table 1. Amongst the single metal supported catalysts the amount of chemisorbed H2 was the highest for 15Fe/Al2O3 followed by 15Ni/Al2O3, and was the least for 15Co/Al2O3. The amount of chemisorbed H2 on 75Ni25Fe/ Al2O3 was larger than that for 15Ni/Al2O3 and on 75Ni25Co/Al2O3 was lower than the 15Ni/Al2O3 catalyst. A previous study dealing with CO chemisorption on titania supported Ni-Co catalysts also reported a lowered amount of chemisorbed CO on Ni-Co catalyst compared to Ni/

Fig. 3. H2-TPD profile of reduced Al2O3 supported Ni, Ni-Fe and Fe samples: (a) 15Ni/ Al2O3; (b) 75Ni25Fe/Al2O3; (c) 50Ni50Fe/Al2O3 (d) 25Ni75Fe/Al2O3 (e) 15Fe/Al2O3. Supported Ni-Co and Co catalysts e.g. (f) 75Ni25Co/Al2O3; (g) 15Co/Al2O3 have been included for comparison.

Fig. 4. CH4 conversions during DRM versus time-on-stream (TOS) over 15Ni/Al2O3, and Ni-Fe/Al2O3 catalysts: (a) 15Ni/Al2O3 (■); (b) 75Ni25Fe/Al2O3 ( ); (c) 50Ni50Fe/Al2O3 ( ). Supported Ni-Co catalyst (d) 75Ni25Co/Al2O3 ( ) has been added for comparison. Reaction conditions: CH4:CO2:N2 = 1:1:3 (vol ratio), W/FA0,CH4 = 0.14 kgcat·h/kg CH4 and reaction temperature = 873 K.

Al2O3 [15]. In contrast, an increased amount of chemisorbed H2 on NiFe catalysts compared to Ni/Al2O3 was evidenced from H2-TPD studies previously [35]. The amount of chemisorbed H2 on 25Ni75Fe/Al2O3 was the highest and even more than that for 15Fe/Al2O3. It is worthwhile noting that the lower crystallite sizes of the metal alloy observed in the two 75Ni25M alloy catalysts relative to 15Ni/Al2O3 metal do not correlate with the amount of chemisorbed H2. These differences further substantiate the dissimilarity of the Al2O3 supported 75Ni25Co and 75Ni25Fe alloys relative to 15Ni/Al2O3. 3.1.5. Catalytic performance for DRM reaction Initially the DRM reaction was carried out in a blank reactor tube and a reactor tube containing the Al2O3 support and no conversion was detected in either case. Then the DRM reaction was carried out using the Al2O3 supported Ni, Ni-Fe and Fe catalysts at 873 K for 3 h and the results are shown in Figs. 4, 5 and 6. The error bars associated with each conversion value were also included in these figures. To enable

Fig. 5. CO2 conversions during DRM versus TOS over 15Ni/Al2O3 and Ni-Fe/Al2O3 catalysts: (a) 15Ni/Al2O3 (■); (b) 75Ni25Fe/Al2O3 ( ); (c) 50Ni50Fe/Al2O3 ( ). Supported Ni-Co catalyst (d) 75Ni25Co/Al2O3 ( ) has been included for comparison. Reaction conditions: CH4:CO2:N2 = 1:1:3 (vol ratio), W/FA0,CH4 = 0.14 kgcat·h/kg CH4 and reaction temperature = 873 K.

200

K. Ray et al. / Fuel Processing Technology 156 (2017) 195–203

Fig. 6. H2:CO ratio during DRM versus TOS over 15Ni/Al2O3 and Ni-Fe/Al2O3 catalysts: (a) 15Ni/Al2O3 (■); (b) 75Ni25Fe/Al2O3 ( ); (c) 50Ni50Fe/Al2O3 ( ). Supported Ni-Co catalyst (d) 75Ni25Co/Al2O3 ( ) has been included for comparison. Reaction conditions: CH4:CO2:N2 = 1:1:3 (vol ratio), W/FA0,CH4 = 0.14 kgcat·h/kg CH4 and reaction temperature = 873 K.

meaningful comparison with the most active supported Ni-Co catalyst (75Ni25Co/Al2O3) the reactivity data along with the error bars of the specified Ni-Co catalyst were obtained and are also included in Figs. 4, 5, and 6. The data for 25Ni75Fe/Al2O3 and 15Fe/Al2O3 were not included since no significant conversions were obtained for these two catalysts. Supported Fe catalysts are known to be relatively inactive under these conditions [52]. Furthermore, the reactivity data presented here were devoid of any internal and/or external mass transfer effects based on analysis done previously [7]. Thus, the data presented in Figs. 4, 5, and 6 not suffer any thermodynamic and mass transfer limitations and were useful for comparing the supported catalysts. The CH4 conversions (%) during DRM as a function of time-onstream (TOS) for the active Ni-Fe catalysts shown in Fig. 4 revealed that the initial conversion of CH4 for 75Ni25Fe/Al2O3 (16 ± 1.1%) was slightly more than that for 15Ni/Al2O3 (13 ± 1.1%). However, the initial conversion for 50Ni50Fe/Al2O3 (4.5 ± 1.0%) was the lowest. In contrast, the initial conversion of 23 ± 1.6% for 75Ni25Co/Al2O3 was the highest. Thus, an optimum amount of Fe substitution gave rise to a slight enhancement of conversion. Furthermore, the optimum amount of substitution was similar for the Ni-Fe and Ni-Co catalysts. The conversion decreased with TOS for all catalysts. The decrease in conversion was about 86% for 50Ni50Fe/Al2O3 compared to 36% for 15Ni/Al2O3 and 26% for 75Ni25Fe/Al2O3 during 3 h of reaction. Though 75Ni25Co/Al2O3 was the most active, the conversion also decreased with TOS by about 33% during 3 h of reaction. Therefore the most active Ni-Fe catalyst (75Ni25Fe/Al2O3) shows slightly improved initial conversion and higher stability relative to the monometallic Ni (15Ni/Al2O3) catalyst. The CO2 conversions during DRM for the active Ni-Fe catalysts were also calculated and shown along with 15Ni/Al2O3 and 75Ni25Co/Al2O3

catalysts in Fig. 5. Similar to CH4 the CO2 conversions also decreased with TOS for the catalysts shown in Fig. 5. Furthermore, the CO2 conversion was higher than the CH4 conversion for all the catalysts. Earlier studies also reported a higher CO2 conversion and attributed this higher conversion to the reverse water gas shift (RWGS) reaction [3,4]. However, a recent and more detailed study suggested that the CO2 conversion over supported Ni-Fe catalysts was lower during the first 10 to 15 min of the reaction due to the higher rates for CH4 cracking on the surface Ni sites [1]. In the same study, the CO2 conversions were higher only after 0.25 h due to the deactivation of the surface Ni sites and the increased rates for carbon oxidation by CO2. Thus, the CO2 conversions being higher than the CH4 conversions may not only be due to the RWGS reaction, and the relative rates of the remaining surface Ni sites and carbon oxidation by CO2 also need to be considered. The initial CO2 conversion of 75Ni25Fe/Al2O3 (24.0 ± 2.5%) was similar to that of 15Ni/Al2O3 (23.6 ± 1.3) and was much higher than that of 50Ni50Fe/Al2O3 (10 ± 1.2%). In comparison, the initial CO2 conversion of 31.0 ± 2.6% for 75Ni25Co/Al2O3 was the highest. Interestingly, the percentage drop in CO2 conversion for 15Ni/Al2O3, 75Ni25Co/Al2O3 and 75Ni25Fe/Al2O3 were similar and ranged from 40 to 43%. For 50Ni50Fe/Al2O3 the CO2 conversion also decreased by 61%. A similar increase in CH4 and CO2 conversions for supported Ni-Co catalysts relative to supported Ni catalyst were previously reported [2,7,15,48,50,52,53]. Thus, the Al2O3 supported Ni-Co catalyst (Ni:Co = 3:1) was found to possess superior catalytic activity relative to supported Ni and Ni-Fe catalysts for the DRM reaction, though a better stability for the DRM reaction was observed for the 75Ni25Fe/Al2O3 catalysts. The number of surface metal sites as estimated from the H2-TPD profiles was used to calculate the TOFDRM for the three relatively active catalysts, 15Ni/Al2O3, 75Ni25Co/Al2O3 and 75Ni25Fe/Al2O3. Accordingly, the TOFDRM for the catalysts at 0.5 h and 3 h TOS were tabulated in Table 2. Table 2 also contains data of TOFCRM and gc/gcat, which are discussed later. The TOFDRM for 15Ni/Al2O3 decreased from 5.40 ± 0.23 to 3.60 ± 0.16 s− 1 during 3 h of reaction. The TOFDRM for the 75Ni25Fe/Al2O3 catalyst decreased from 5.80 ± 0.21 to 4.20 ± 0.12 s− 1 and was relatively higher than 15Ni/Al2O3. The TOFDRM for 75Ni25Co/Al2O3 at 0.5 h and 3 h were significantly higher than the TOFDRM for the 15Ni/Al2O3 and 75Ni25Fe/Al2O3 catalysts. The TOFDRM of supported Ni and Ni-Co catalysts have been compared and discussed previously [7]. However, data on the TOFDRM for supported Ni-Fe catalysts appear to be limited. The closest study was that of MgAl2O4 supported Ni and Ni-Fe catalysts, where the site time yields (STY) were reported [1]. The STY was given in terms of the formation (CO and H2) or consumption (CH4 and CO2) of the compound per unit time per site. Furthermore, the dispersion of the metal(s) was based on the crystallite size determined from the full scan XRD pattern. For a MgAl2O4 supported Ni and the most active Ni-Fe catalyst the STY after 0.5 h was about 9.2 × 10−2 and 11 × 10−2 mol CH4 (mol Ni.s)−1, respectively. However, in that previous study the total metal loading was not the same and the STY appeared to be obtained at 1023 K. In the present study the total metal loading was 15 wt.% for the 15Ni/Al2O3 and 75Ni25Fe/Al2O3 catalysts, the reaction temperature was 873 K, and number of surface sites were measured by H2-TPD. Under these conditions the TOFDRM for 15Ni/Al2O3 and 75Ni25Fe/

Table 2 Reactivity data for DRM and CRM reactions for the 15Ni/Al2O3 and Ni-Fe/Al2O3 catalysts after 0.5 and 3 h TOS. Supported Ni-Co catalyst, 75Ni25Co/Al2O3, has been included for comparison. Reaction conditions: CH4:CO2:N2 = 1:1:3 for DRM and CH4:N2 = 1:4 for CRM, W/FA0,CH4 = 0.14 kgcat·h/kg CH4 and reaction temperature = 873 K. Sample nomenclature

15Ni/Al2O3 75Ni25Fe/Al2O3 50Ni50Fe/Al2O3 75Ni25Co/Al2O3

TOFDRM (s−1)

TOFCRM (s−1)

gc/gcat after 3 h reaction

After 0.5 h

After 3.0 h

After 0.5 h

After 3.0 h

For DRM

For CRM

5.40 ± 0.23 5.80 ± 0.21 0.54 ± 0.12 15.30 ± 0.42

3.60 ± 0.16 4.20 ± 0.12 0.10 ± 0.07 10.20 ± 0.29

0.75 1.01 0.43 1.52

0.18 0.37 0.06 0.42

0.17 0.18 n.d. 0.20

0.25 0.26 n.d. 0.29

± ± ± ±

0.02 0.03 0.01 0.03

± ± ± ±

0.01 0.01 0.01 0.01

K. Ray et al. / Fuel Processing Technology 156 (2017) 195–203

Al2O3 decreased during 3 h of the reaction from 5.40 ± 0.23 to 3.60 ± 0.16 s− 1 (for 15Ni/Al2O3) and from 5.80 ± 0.21 to 4.20 ± 0.12 s− 1 (for 75Ni25Fe/Al2O3). Though the specific values of the STY in the previous study and TOFDRM in the present study are different the ratio of the parameters for the Ni and Ni-Fe catalysts were similar. The higher activity of the most active Ni-Fe catalyst relative to the Ni catalyst in the previous study [1] may be related to the difference in the total metal loading considered. The H2:CO ratio during the DRM reaction is also an important parameter that is used to compare catalysts. The H2:CO ratio for the catalysts was determined from the reactivity data and presented in Fig. 6. A H2:CO ratio of one corresponds to the stoichiometric value for the DRM reaction occurring in isolation. However, the H2:CO ratios for all the catalysts shown in Fig. 6 were less than one due to the higher yield of CO compared to H2. The higher CO yield is favoured due to the RWGS and oxidation of deposited carbon reactions. The 75Ni25Fe/Al2O3 catalyst showed H2:CO ratio similar to that of 15Ni/Al2O3 and a much higher H2:CO ratio with TOS than the 50Ni50Fe/Al2O3 catalyst. In contrast, the H2:CO ratio with TOS was the highest for 75Ni25Co/Al2O3 during 3 h of reaction. Interestingly, the H2:CO ratio with TOS did not vary much for 75Ni25Fe/Al2O3 compared to 15Ni/Al2O3 and 75Ni25Co/ Al2O3 catalysts, which re-iterates the relatively more stable behaviour of 75Ni25Fe/Al2O3 during 3 h of DRM reaction. The average H2:CO ratio was the highest for 75Ni25Co/Al2O3 (0.52), and comparable for 15Ni/Al2O3 (0.46) and 75Ni25Fe/Al2O3 (0.44) catalysts. Previous studies reported higher H2:CO ratios for Ni-Co based catalysts, but at reaction temperatures ranging from 1023 to 1073 K [2,34,50]. High H2:CO ratios were also reported for supported Ni-Fe catalysts at 1023 K [1]. Considering that the reaction temperature used in our present study was 873 K a lower H2:CO ratio is expected. In a previous study the H2:CO ratio increased from about 0.9 to more than 1.0 when the reaction temperature was increased from 873 to 1073 K [49]. Partial validation of such an approach of increasing the H2:CO ratio by increasing the reaction temperature was successfully tested in our previous study on supported Ni-Co catalysts [7]. In this previous study the average H2:CO ratio increased from 0.52 to 0.71 for the 75Ni25Co/Al2O3 catalyst when the reaction temperature was increased from 873 to 973 K. In the present study the average H2:CO ratio also increased from 0.44 to 0.52 for the 75Ni25Fe/Al2O3 catalyst when the reaction temperature was increased from 873 to 923 K. For completeness, the amount of carbon deposited on the spent catalysts was also measured and reported as gc/gcat in Table 2. The data in Table 2 showed that the amount of carbon deposited on the spent 75Ni25Co/Al2O3 catalyst was more than that on the other catalysts. The maximum conversions and TOFDRM were also found for this catalyst during the DRM reaction. The amount of carbon deposited on spent 75Ni25Fe/Al2O3 and 15Ni/Al2O3 were similar, which was consistent with the small differences in the conversions observed on these two catalysts. The reactivity data for DRM revealed that the CH4 and CO2 conversions, H2:CO ratio and amount of carbon deposited were the highest for the 75Ni25Co/Al2O3 catalyst. However, the 75Ni25Co/Al2O3 catalyst deactivated during the 3 h of reaction. In contrast, 75Ni25Fe/ Al2O3 catalyst was slightly more active than supported Ni and was relatively more stable than 75Ni25Co/Al2O3 and 15Ni/Al2O3. Thus, the two supported Ni-M/Al2O3 (M = Co, Fe) catalyst with Ni:M of 3:1 resulted in an overall improved catalytic activity relative to the supported Ni catalyst. 3.1.6. Reasons for the better catalytic performance in DRM As discussed above the reduction temperature of the catalysts prior to the reaction was 823 K, and it appears unlikely that the shift in the reduction temperature and the small differences in metal oxide reduction (%) relative to 15Ni/Al2O3 were related to the improved catalytic activity of 75Ni25Fe/Al2O3 and 75Ni25Co/Al2O3. The XRD patterns revealed the presence of alloys that were formed with a specific Ni to M ratio of 3:1. The XRD patterns also revealed that the presence of Fe or Co had an

201

effect on the crystallite size. However, the more relevant parameter to correlate with the catalytic activity would the particle size of the supported metals and metal alloys. The particle size is closely related to the amount of H2 desorbed, which was determined from H2-TPD. However, comparing the amount of H2 desorbed in Table 1 with the TOFDRM in Table 2 reveals no correlation. Furthermore, the surface properties of the Ni-M alloys formed were different as suggested from the H2-TPD results. Thus, the presence of the alloys with specific composition appears to give rise to the improved catalytic activity of the supported Ni-Co and Ni-Fe catalysts for the DRM reaction and the best catalytic activity of the supported Ni-Co catalysts appears to be due to the specific surface properties of the Ni-Co alloy. 3.1.7. Catalytic performance for CRM reaction As mentioned above the CRM reaction has been proposed to be an important step during the DRM reaction. To examine the importance and relation between the CRM and DRM reactions the catalytic activity of supported Ni, Ni-Co and Ni-Fe catalysts for the CRM reaction was carried out under the same operating conditions as for the DRM. The CH4 conversion data during cracking were used to calculate the TOFCRM, and the TOFCRM was plotted as a function of TOS in Fig. 7. The TOFCRM for 15Fe/Al2O3 are not shown as the conversions were not significant. The data in Fig. 7 revealed that the initial TOFCRM for 75Ni25Fe/Al2O3 was slightly higher than the TOFCRM for 15Ni/Al2O3. Furthermore, the TOFCRM for 75Ni25Fe/Al2O3 was higher than that for 15Ni/Al2O3 over the entire 3 h of reaction. The 50Ni50Fe/Al2O3 catalyst had a lower TOFCRM during 3 h reaction, and the TOFCRM for 25Ni75Fe/Al2O3 was the least. In contrast, the TOFCRM of 75Ni25Co/Al2O3 was significantly higher than that of 75Ni25Fe/Al2O3 and 15Ni/Al2O3. The TOFCRM of all catalysts decreased with TOS. The decrease in TOFCRM was about 91% and 78% for 50Ni50Fe/Al2O3 and 25Ni75Fe/Al2O3 respectively during 3 h of reaction. The decrease in TOFCRM was 70% for 75Ni25Fe/Al2O3 compared to 83% for 15Ni/Al2O3 during 3 h of reaction. Though 75Ni25Co/Al2O3 was the most active, the TOFCRM also decreased with TOS by about 77% during 3 h of reaction. Therefore, the most active Ni-Fe catalyst (75Ni25Fe/Al2O3) shows slightly improved initial TOFCRM and higher stability relative to the monometallic Ni (15Ni/Al2O3) catalyst. Thus, the two Ni-M supported catalysts possessing Ni and M in the ratio of 3:1 exhibited an improved catalytic activity for the CRM reaction relative to the 15Ni/Al2O3 catalyst, and 75Ni25Co/Al2O3 was the most active amongst the catalysts studied. Interestingly, a similar

Fig. 7. TOFCRM versus TOS for the Al2O3 supported Ni and Ni-Fe catalysts: (a) 15Ni/Al2O3 (■); (b) 75Ni25Fe/Al2O3 ( ); (c) 50Ni50Fe/Al2O3 ( ); (d) 25Ni75Fe/Al2O3 ( ). Supported Ni-Co catalyst (e) 75Ni25Co/Al2O3 ( ) has been included for comparison. Reaction conditions: CH4:N2 = 1:4 (vol ratio), W/FA0,CH4 = 0.14 kgcat·h/kg CH4 and reaction temperature = 873 K.

202

K. Ray et al. / Fuel Processing Technology 156 (2017) 195–203

trend was observed for the DRM reaction and discussed above. This improved CRM activity of 75Ni25Co/Al2O3, and the relatively slight improved activity and stability of 75Ni25Fe/Al2O3 closely matched the trends observed for the DRM reaction. Similar to the arguments above for the DRM reaction the improved CRM activity for the 75Ni25Fe/ Al2O3 and 75Ni25Co/Al2O3 relative to 15Ni/Al2O3 appear to be due to the presence of alloys of a specific composition. Furthermore, the better CRM activity for 75Ni25Co/Al2O3 appears to be due to the difference in surface properties of the Ni-Co alloy formed. Wei et al. revealed an identical forward CH4 reaction rates, rate constants, kinetic isotopic effect and activation energies for CH4 decomposition and CH4-CO2 reforming over Ni-MgO catalysts [23]. In a recent study the DRM reaction was proposed to occur by a Mars van Krevelan mechanism over supported Ni-Fe catalysts, where the CRM reaction occurred and carbon was deposited on the catalyst [1]. Simultaneously, the deposited carbon was oxidized by CO2 to produce CO. It was also proposed that the rate of carbon gasification needs to be in excess to the rate of carbon formation to avoid net carbon being deposited on the catalyst surface [54]. An imbalance of the two reactions would give rise to carbon deposition and deactivation of the supported Ni based catalysts [3,4,15,26,48,53,55]. These studies substantiate that the catalytic activity of CRM and DRM are closely linked. The close link between the DRM and CRM activities is further substantiated in the present study for supported Ni-Co and Ni-Fe catalysts. The amount of carbon deposited during 3 h of the CRM reaction on 15Ni/Al2O3, 75Ni25Co/Al2O3 and 75Ni25Fe/Al2O3 was also measured and reported in Table 2. The carbon deposited on 75Ni25Co/Al2O3 was slightly more than that on 75Ni25Fe/Al2O3 and 15Ni/Al2O3. Furthermore, the amounts of carbon deposited on the three catalysts during the CRM reaction were greater than that deposited during the DRM reaction suggesting that CO2 plays an important role in removing the deposited carbon. Additionally, the difference in carbon deposited during DRM and CRM reactions in Table 2 were similar for the 75Ni25Co/ Al2O3, 75Ni25Fe/Al2O3 and 15Ni/Al2O3 catalysts. Such a similarity suggests that the rate of carbon removal by CO2 was independent of the particular catalyst used for the DRM reaction. Thus, the more active 75Ni25Co/Al2O3 catalyst had more carbon deposited during the DRM reaction due to the higher amount of carbon deposited during the CRM reaction, and the higher activity of 75Ni25Co/Al2O3 appeared to be due to the higher CH4 cracking capability of this catalyst. 4. Conclusions The DRM and CRM reactions were studied over 15Ni/Al2O3 and NiFe/Al2O3 catalysts, with the total metal loading of 15 wt.% and the catalytic activity and stability of the two reactions were slightly enhanced for 75Ni25Fe/Al2O3 relative to 15Ni/Al2O3. Furthermore, 75Ni25Co/ Al2O3 showed the maximum promotional effect for the two reactions. Characterization of the catalysts revealed that 15Fe/Al2O3 was significantly less reduced compared to 15Ni/Al2O3 and 15Co/Al2O3. The lower extent of reduction of Fe had an effect on the reducibility of the supported Ni-Fe catalysts compared to the supported Ni-Co catalysts. However, the extent of reduction of 75Ni25Fe/Al2O3, 75Ni25Co/Al2O3 and 15Ni/Al2O3 were similar and appear not to be responsible for the improved catalytic activity. Furthermore, the amount of H2 desorbed, which is closely related to the particle size of the metal/alloy, did not correlate well with the changes in catalytic activity. The formation of Ni based alloys in the reduced supported Ni-Co and Ni-Fe catalysts containing a specific Ni to M ratio of 3:1 was substantiated from the XRD study. Furthermore, the surface properties of these specific alloys were different as suggested by the H2-TPD results. Specific changes in the surface properties would require extensive electronic property calculations, which is beyond the scope of the present work. The presence of these alloys appears to be the reason for the improved catalytic activity for CRM and DRM reactions of these two catalysts relative to 15Ni/ Al2O3. Furthermore, the 75Ni25Co/Al2O3 catalyst was found to be

superior to all catalysts for the DRM and CRM reactions. However, the amount of carbon deposited on 75Ni25Co/Al2O3 during the two reactions was the highest. In contrast, the 75Ni25Fe/Al2O3 catalyst was slightly more active and relatively more stable for the DRM and CRM reactions compared to 15Ni/Al2O3. It appears that the superior catalytic activity of supported Ni-Co and Ni-Fe for the DRM reaction relative to Ni was due to the improved cracking activity for these catalysts. It is the CH4 cracking capability that determines the overall catalytic performance for DRM and the methane cracking ability of the supported catalysts is affected by the presence of the alloys. Other strategies, such as a change in operating conditions, may assist in further decreasing the amount of carbon deposited on 75Ni25Co/Al2O3 during the DRM reaction, which may be pursued in the future. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.fuproc.2016.11.003. Acknowledgements The authors gratefully acknowledge the support from the Council of Scientific and Industrial Research (CSIR), Project no. 22(0634)/13/EMR-II. References [1] S.A. Theofanidis, V.V. Galvita, H. Poelman, G.B. Marin, Enhanced carbon-resistant dry reforming Fe-Ni catalyst: role of Fe, ACS Catal. 5 (2015) 3028–3039. [2] J. Zhang, H. Wang, A.K. Dalai, Development of stable bimetallic catalysts for carbon dioxide reforming of methane, J. Catal. 249 (2007) 300–310. [3] M.-S. Fan, A.Z. Abdullah, S. Bhatia, Catalytic technology for carbon dioxide reforming of methane to synthesis gas, ChemCatChem 1 (2009) 192–208. [4] M.C.J. Bradford, M.A. Vannice, CO2 reforming of CH4, Catal. Rev. 41 (1999) 1–42. [5] D. San-Jose-Alonso, J. Juan-Juan, M.J. Illan-Gomez, M.C. Roman-Martinez, Ni, Co and bimetallic Ni–Co catalysts for the dry reforming of methane, Appl. Catal. A Gen. 371 (2009) 54–59. [6] J. Zhang, H. Wang, A.K. Dalai, Effects of metal content on activity and stability of NiCo bimetallic catalysts for CO2 reforming of CH4, Appl. Catal. A Gen. 339 (2008) 121–129. [7] S. Sengupta, K. Ray, G. Deo, Effects of modifying Ni/Al2O3 catalyst with cobalt on the reforming of CH4 with CO2 and cracking of CH4 reactions, Int. J. Hydrog. Energy 39 (2014) 11462–11472. [8] M.A. Naeem, A.S. Al-Fatesh, A.E. Abasaeed, A.H. Fakeeha, Activities of Ni-based nano catalysts for CO2–CH4 reforming prepared by polyol process, Fuel Process. Technol. 122 (2014) 141–152. [9] L. Yao, J. Shi, H. Xu, W. Shen, C. Hu, Low-temperature CO2 reforming of methane on Zr-promoted Ni/SiO2 catalyst, Fuel Process. Technol. 144 (2016) 1–7. [10] H. Wu, G. Pantaleo, V. Parola, A.M. Venezia, X. Collard, C. Aprile, Bi- and trimetallic Ni catalysts over Al2O3 and Al2O3-MOx (M = Ce or Mg) oxides for methane dry reforming: Au and Pt additive effects, Appl. Catal. B Environ. 156–157 (2014) 350–361. [11] S.M.d. Lima, J.M. Assaf, Ni–Fe catalysts based on perovskite-type oxides for dry reforming of methane to syngas, Catal. Lett. 108 (2006) 63–70. [12] A. Djaidja, H. Messaoudi, D. Kaddeche, A. Barama, Study of Ni-M/MgO and Ni-M-Mg/ Al (M = Fe or Cu) catalysts in the CH4-CO2 and CH4-H2O reforming, Int. J. Hydrog. Energy 40 (2015) 4989–4995. [13] S.R.d. Miguel, I.M.J. Vilella, S.P. Maina, D.S. José-Alonso, M.C. Román-Martínez, M.J. Illán-Gómez, Influence of Pt addition to Ni catalysts on the catalytic performance for long term dry reforming of methane, Appl. Catal. A Gen. 435–436 (2012) 10–18. [14] C. Crisafulli, S. Scire, R. Maggiore, S. Minico, S. Galvagno, CO2 reforming of methane over Ni–Ru and Ni–Pd bimetallic catalysts, Catal. Lett. 59 (1996) 21–26. [15] K. Takanabe, K. Nagaoka, K. Nariai, K.-I. Aika, Titania-supported cobalt and nickel bimetallic catalysts for carbon dioxide reforming of methane, J. Catal. 232 (2005) 268–276. [16] L. Wang, D. Li, M. Koike, H. Watanabe, YaXu, Y. Nakagawa, K. Tomishige, Catalytic performance and characterization of Ni–Co catalysts for the steam reforming of biomass tar to synthesis gas, Fuel 112 (2013) 654–661. [17] M. Koike, D. Li, Y. Nakagawa, K. Tomishige, A highly active and coke-resistant steam reforming catalyst comprising uniform nickel–iron alloy nanoparticles, ChemSusChem 5 (2012) 2312–2314. [18] L. Wang, D. Li, M. Koike, S. Koso, Y. Nakagawa, YaXu, K. Tomishige, Catalytic performance and characterization of Ni-Fe catalysts for the steam reforming of tar from biomass pyrolysis to synthesis gas, Appl. Catal. A Gen. 392 (2011) 248–255. [19] Z. JSM, S. KJ, Kinetics of CH4 decomposition on supported cobalt catalysts, J. Catal. 176 (1998) 115–124. [20] J. Guo, H. Lou, X. Zheng, The deposition of coke from methane on a Ni/MgAl2O4 catalyst, Carbon 45 (2007) 1314–1321. [21] N.A. Hermes, M.A. Lansarin, O.W. Perez-Lopez, Catalytic decomposition of methane over M-Co-Al catalysts (M = Mg, Ni, Zn, Cu), Catal. Lett. 141 (2011) 1018–1025. [22] E.D. German, M. Sheintuch, Predicting CH4 dissociation kinetics on metals: trends, sticking coefficients, H tunneling, and kinetic isotope effect, J. Phys. Chem. C 117 (2013) 22811–22826.

K. Ray et al. / Fuel Processing Technology 156 (2017) 195–203 [23] J. Wei, E. Iglesia, Isotopic and kinetic assessment of the mechanism of reactions of CH4 with CO2 or H2O to form synthesis gas and carbon on nickel catalysts, J. Catal. 224 (2004) 370–383. [24] Z. Wang, X.-M. Cao, J. Zhu, P. Hu, Activity and coke formation of nickel and nickel carbide in dry reforming: a deactivation scheme from density functional theory, J. Catal. 311 (2014) 469–480. [25] D. Pakhare, V. Schwartz, V. Abdelsayed, D. Haynes, D. Shekhawat, J. Poston, J. Spivey, Kinetic and mechanistic study of dry (CO2) reforming of methane over Rh-substituted La2Zr2O7 pyrochlores, J. Catal. 316 (2014) 78–92. [26] Y. Kathiraser, U. Oemar, E.T. Saw, Z. Li, S. Kawi, Kinetic and mechanistic aspects for CO2 reforming of methane over Ni based catalysts, Chem. Eng. J. 278 (2015) 62–78. [27] A.M. Amin, E. Croiset, W. Epling, Review of methane catalytic cracking for hydrogen production, Int. J. Hydrog. Energy 36 (2011) 2904–2935. [28] T.V. Choudhary, C. Sivadinarayana, C.C. Chusuei, A. Klinghoffer, D.W. Goodman, Hydrogen production via catalytic decomposition of methane, J. Catal. 199 (2001) 9–18. [29] C. Tanggarnjanavalukul, W. Donphai, T. Witoon, M. Chareonpanich, J. Limtrakul, Deactivation of nickel catalysts in methane cracking reaction: effect of bimodal meso– macropore structure of silica support, Chem. Eng. J. 262 (2015) 364–371. [30] J. Sehested, K.E. Larsen, A.L. Kustov, A.M. Frey, T. Johannessen, T. Bligaard, M.P. Andersson, J.K. Nørskov, C.H. Christensen, Discovery of technical methanation catalysts based on computational screening, Top. Catal. 45 (2007) 9–13. [31] B. Mitra, X. Gao, I. Wachs, A. Hirt, G. Deo, Characterization of supported rhenium oxide catalysts: effect of loading, support and additives, Phys. Chem. Chem. Phys. 3 (2001) 1144–1152. [32] S. Logdberg, D. Tristantini, Ø. Borg, L. Ilver, B. Gevert, S. Jaras, E.A. Blekkan, A. Holmen, Hydrocarbon production via Fischer–Tropsch synthesis from H2-poor syngas over different Fe-Co/gamma-Al2O3 bimetallic catalysts, Appl. Catal. B Environ. 89 (2009) 167–182. [33] A. Nandini, K.K. Pant, S.C. Dhingra, Kinetic study of the catalytic carbon dioxide reforming of methane to synthesis gas over Ni-K/CeO2-Al2O3 catalyst, Appl. Catal. A Gen. 308 (2006) 119–127. [34] M.-S. Fan, A.Z. Abdullah, S. Bhatia, Utilization of greenhouse gases through carbon dioxide reforming of methane over Ni-Co/MgO-ZrO2: preparation, characterization and activity studies, Appl. Catal. B Environ. 100 (2010) 365–377. [35] D. Pandey, G. Deo, Promotional effects in alumina and silica supported bimetallic Ni–Fe catalysts during CO2 hydrogenation, J. Mol. Catal. A Chem. 382 (2014) 23–30. [36] R. Brown, M.E. Cooper, D.A. Whan, Temperature programmed reduction of aluminasupported iron, cobalt and nickel bimetallic catalysts, Appl. Catal. 3 (1982) 177–186. [37] L.D. Rogatis, T. Montini, A. Cognigni, L. Olivi, P. Fornasiero, Methane partial oxidation on NiCu-based catalysts, Catal. Today 145 (2009) 176–185. [38] Y. Ji, Z. Zhao, A. Duan, G. Jiang, J. Liu, Comparative study on the formation and reduction of bulk and Al2O3-supported cobalt oxides by H2-TPR technique, J. Phys. Chem. C 113 (2009) 7186–7199. [39] S. Sitthisa, W. An, D.E. Resasco, Selective conversion of furfural to methylfuran over silica-supported Ni-Fe bimetallic catalysts, J. Catal. 284 (2011) 90–101.

203

[40] D.P. Rodionov, I.V. Gervas'eva, Y.V. Khlebnikova, V.A. Sazonova, B.K. Sokolov, Effect of alloying and heat treatment on the formation of recrystallization cube texture in nickel alloys, Fiz. Met. Metalloved. 99 (2005) 1–88. [41] F. Hofer, Thermodynamic properties of solid rhodium-nickel alloys, J. Solid State Chem. 45 (1982) 303–308. [42] S. Zoriasatain, F. Azarkharman, A.S. Sebt, M. Akhavan, Magnetic anisotropies in FeCo fine particles, J. Magn. Magn. Mater. 300 (2006) 525–531. [43] I. Chicinas, V. Pop, O. Isnard, J.M.L. Breton, J. Juraszek, Synthesis and magnetic properties of Ni3Fe intermetallic compound obtained by mechanical alloying, J. Alloys Compd. 352 (2003) 34–40. [44] E.A. Ibragimov, Y.P. Selisskiy, M.N. Sorokin, Investigation of the effects of ordering in ternary Ni-Fe-Mn solid solutions, Russ. Metall. 5 (1966) 82–87. [45] C. Resini, M.C.H. Delgado, S. Presto, L.J. Alemany, P. Riani, R. Marazza, G. Ramis, G. Busca, Yttria-stabilized zirconia (YSZ) supported Ni–Co alloys (precursor of SOFC anodes) as catalysts for the steam reforming of ethanol, Int. J. Hydrog. Energy 33 (2008) 3728–3735. [46] S. Ochiai, Y. Mishima, T. Suzuki, Lattice parameter data of Ni(γ), Ni3Al(γ') and Ni3Ga(γ') solid solutions, Bull. Res. Lab. Precis. Mach. Electron. (Tokyo Inst. Technol.) 53 (1984) 15–28. [47] A. Taylor, Lattice parameters of binary nickel-cobalt alloys, J. Inst. Met. 77 (1950) 585–594. [48] O. Gonzalez, J. Lujano, E. Pietri, M.R. Goldwasser, New Co-Ni catalyst systems used for methane dry reforming based on supported catalysts over an INT-MM1 mesoporous material and a perovskite-like oxide precursor LaCo0·4Ni0.6O3, Catal. Today 107-108 (2005) 436–443. [49] I. Luisetto, S. Tuti, E.D. Bartolomeo, Co and Ni supported on CeO2 as selective bimetallic catalyst for dry reforming of methane, Int. J. Hydrog. Energy 37 (2012) 15992–15999. [50] J. Xu, W. Zhou, Z. Li, J. Wang, J. Ma, Biogas reforming for hydrogen production over nickel and cobalt bimetallic catalysts, Int. J. Hydrog. Energy 34 (2009) 6646–6654. [51] S.d. Llobet, J.L. Pinilla, R. Moliner, I. Suelves, Relationship between carbon morphology and catalyst deactivation in the catalytic decomposition of biogas using Ni, Co and Fe based catalysts, Fuel 139 (2015) 71–78. [52] S.Y. Foo, C. CK, N. T-H, A. AA., Oxidative CO2 reforming of methane on alumina-supported Co-Ni catalyst, Ind. Eng. Chem. Res. 49 (2010) 10450–10458. [53] V.M. Gonzalez-delaCruz, R. Pereniguez, F. Ternero, J.P. Holgado, A. Caballero, In situ XAS study of synergic effects on NiCo/ZrO2 methane reforming catalysts, J. Phys. Chem. C 116 (2012) 2919–2926. [54] P. Littlewood, X. Xie, M. Bernicke, A. Thomas, R. Schomäcker, Ni0.05 Mn0·95O catalysts for the dry reforming of methane, Catal. Today 242 (2015) 111–118. [55] S. Wang, G.Q.M. Lu, G.J. Millar, Carbon dioxide reforming of methane to produce synthesis gas over metal-supported catalysts: state of the art, Energy Fuel 10 (1996) 896–904.