Fuel Processing Technology 186 (2019) 40–52

Contents lists available at ScienceDirect

Fuel Processing Technology

journal homepage: www.elsevier.com/locate/fuproc

Research article

Effect of support materials on the performance of Ni-based catalysts in tri- T

reforming of methane
Rohit Kumara, Kritagya Kumara, N.V. Choudaryb, K.K. Panta,
⁎

a
Department of Chemical Engineering, Indian Institute of Technology Delhi, Hauz Khas, New Delhi 110016, India
b
HP Green R&D Centre, KIADB Industrial Area, Tarabanahalli, Devanagundi, Hoskote, Bengaluru 560067, Karnataka, India

ARTICLE INFO ABSTRACT

Keywords: Tri-reforming of methane (TRM) utilizing CO2, H2O and O2 is an emerging technique for the production of
Tri-reforming synthesis gas and CO2 abatement. Nevertheless, formulation and design of low-cost, active and stable catalyst for
Ni catalysts TRM is a major challenge due to its complex reaction network and rapid catalyst deactivation. The support effect
Catalyst deactivation is a critical factor influencing the activity and stability of supported Ni catalyst. Thus, a series of supported Ni
Synthesis gas
catalysts over different metal oxides such as Al2O3, ZrO2, TiO2, SBA-15, MgO and CeO2-ZrO2 were prepared. The
CO2 conversion
catalysts were characterized by using N2 - physisorption, XRD, EDX, SEM, TEM, TPR, CO2 – TPD, NH3 – TPD, H2
Support effect
– pulse chemisorption and TGA. The performance of these catalysts was evaluated at 800 °C, 1 atm, WHSV
17220 mL h−1 g−1 and molar feed composition CH4:CO2:H2O:O2:N2 = 1:0.23:0.46:0.07:0.28 in a fixed-bed re-
actor. The initial activity of these catalysts followed the order: Ni/Al2O3 > Ni/SBA-15 > Ni/ZrO2 > Ni/CeO2-
ZrO2 > Ni/TiO2 > Ni/MgO. Ni/Al2O3 derived from spinel precursor NiAl2O4 resulted in well-dispersed smaller
Ni particles, stronger metal-support interaction, a higher degree of reducibility and higher basic sites con-
centration, which led to its superior activity with 8.72 × 10−2 mmol/gcat·s, 2.31 × 10−2 mmol/gcat·s and
4.2 × 10−2 mmol/gcat·s CH4, CO2, and H2O conversion rate respectively. On the contrary, the stability of SBA-15
and ZrO2 supported Ni catalysts was remarkably high. Hexagonal porous network of SBA-15 confined Ni par-
ticles that resulted in high resistance against metal sintering and carbon deposition whereas Ni/ZrO2 offered
resistance to Ni re-oxidation owing to its oxophilic property. Experimental investigations reveal that Ni/TiO2
and Ni/MgO resulted in poor CH4, CO2 and H2O conversion rates for TRM due to lower degree of reducibility.
Furthermore, Ni/CeO2-ZrO2 catalyst had weak metal-support interaction as observed from TPR studies. Due to
larger Ni crystallite size, it exhibited lower conversion compared to Ni/Al2O3, Ni/SBA-15, and Ni/ZrO2.

1. Introduction power plant's net electricity output by as much as 20%. Reforming of

methane in presence of CO2, H2O, and O2, known as “Tri-reforming of
Methane has been projected as a potential feedstock for fuels and methane” (TRM), can convert CO2 without its pre-separation from flue
chemicals [1]. Its importance has increased significantly since the re- gas [6]. It is a synergetic combination of steam reforming (Eq. (1)); dry
cent advancement of hydraulic fracturing technology and exploitation reforming (Eq. (2)) and partial oxidation (Eq. (3)) in a single reforming
of shale gas [2]. Methane reforming is a well-known process for the reactor.
production of synthesis gas (CO + H2) and hydrogen. Synthesis gas acts
CH 4 + H2 O CO + 3H2 H °298 = + 206 kJ/mol (steam reforming)
as a building block for the production of synthetic fuels like methanol
and dimethyl ether and other bulk chemicals [3]. The commercial (1)
technique for hydrogen production is presently methane steam re- CH 4 + CO2 2CO + 2H2 H °298 = +247 kJ/mol (dry reforming) (2)
forming. However, it generates a large amount of CO2. In this context,
dry reforming is especially interesting since it utilizes greenhouse gas CH 4 + 1/2O2 CO + 2H2 H °298 = 38 kJ/mol (partial oxidation)
CO2 to reform methane [4]. Flue gas from electric power generation (3)
plants is the largest source of CO2 worldwide [5]. Separation of CO2
CO + H2 O CO2 + H2 H °298 = 41 kJ/mol (water gas shift reaction)
from flue gas is highly energy intensive step and it alone reduces a
(4)

⁎
Corresponding author.
E-mail address: kkpant@chemical.iitd.ac.in (K.K. Pant).

https://doi.org/10.1016/j.fuproc.2018.12.018
Received 21 September 2018; Received in revised form 21 November 2018; Accepted 20 December 2018
Available online 31 December 2018
0378-3820/ © 2018 Elsevier B.V. All rights reserved.
R. Kumar et al. Fuel Processing Technology 186 (2019) 40–52

CH 4 C + 2H2 H °298 = + 75 kJ/mol (methane decomposition) (5) and MgO (Merck, ≥98% assay) were used for this purpose. While SBA-
15 and CeO2-ZrO2 (CeO2/ZrO2 mass ratio = 1.0) catalyst supports were
2CO C + CO2 H °298 = 172 kJ/mol (CO disproportionation) (6) synthesized as per the methods described elsewhere with little mod-
ification [17,18]. Tetraethylorthosilicate (TEOS) as a silica source and
In tri-reforming of methane, syngas of desired H2/CO ratio can be triblock co-polymer poly (poly ethylene glycol)-block-poly (propylene
obtained by adjusting the feed gas composition. Various downstream glycol)-block-poly (ethylene glycol) EO20PO70EO20 (Pluronic P123)
processes such as methanol synthesis, Fischer-Tropsch, Oxo synthesis (Sigma-Aldrich) as a templating agent were used to prepare mesoporous
and DME synthesis require different H2/CO ratio. The presence of siliceous SBA-15. 4 g of EO20PO70EO20 polymer was dissolved in 80 g of
oxygen in feed can make the process thermo-neutral/mild endothermic 2 M HCl and 20 g of H2O under constant stirring. Afterwards, 8.8 g of
as it can provide the necessary heat required for endothermic steam and TEOS was added dropwise to make a final solution of EO20PO70EO20:
dry reforming reactions. Also, it can eliminate coke deposition on re- HCl: TEOS: H2O = 2:60:4.25:12 (mass ratio). The solution was stirred at
forming catalyst and hence can increase catalyst life [6]. The occur- 40 °C for 20 h and then aged at 100 °C for 48 h. It was cooled to room
rence of two endothermic reactions would avoid hot spot formation and temperature. The resulting solid was filtered, washed several times in
local over heating in the oxidative reaction environment. In addition to deionized water and dried under ambient conditions. Subsequently, the
this, it ensures that no feed would be lost even in case CH4 oxidized to solid was calcined at 500 °C for 5 h to remove the template. The meso-
CO2 and H2O as both will be utilized by steam and dry reforming re- porous SBA-15 material was thus obtained [17]. The CeO2-ZrO2 (CeO2/
actions in the tri-reforming process [7]. Halmann et al. examined TRM ZrO2 mass ratio = 1.0) was prepared by co-precipitation method. A
process and reported that TRM route is more beneficial than methane certain amount of cerium (III) nitrate hexahydrate [Ce(NO3)3·6H2O] and
steam reforming route for synthesis gas generation for methanol pro- zirconyl nitrate hydrate [ZrO(NO3)2·xH2O] (Central Drug House Pvt.
duction as TRM result in 50% CO2 emission avoidance, 32% fuel saving Ltd., India) were dissolved in deionized water to form a mixed aqueous
and 73% exergy efficiency [8]. Despite these advantages, large-scale solution. It was stirred continuously for 1 h at 50 °C. Aqueous ammonia
implementation of methane tri-reforming is impeded by the lack of was added to the solution to maintain pH 10, which facilitates pre-
active and stable catalysts under harsh reaction conditions. Unlike cipitation of cerium and zirconium ions as their hydroxides. Afterwards,
single methane reforming reactions, the presence of CH4, CO2, H2O and the resulting precipitate was aged overnight. It was washed thoroughly
O2 in feed-stream makes the reaction environment very complex and using deionized water and dried overnight at 90 °C. The pre-dried solid
hence it tends to become challenging for a reforming catalyst to adsorb was calcined at 750 °C for 6 h to form CeO2-ZrO2 [18]. Nickel was loaded
and activate all the reactants efficiently. In addition to this, another to these metal oxides by the wet impregnation method. In this study, Ni
major challenge is catalyst deactivation such as carbon accumulation, (NO3)2·6H2O (Fisher Scientific, ≥98% assay) was used as a precursor. To
metal sintering, and metal re-oxidation during reforming reaction prepare the Ni-based catalyst, an appropriate amount of Ni(NO3)2·6H2O
which causes catalyst instability and ultimately leads to loss of activity and metal oxide was dissolved in water. The mixture was stirred for 1 h.
[9]. Considerable efforts have been made to develop a suitable catalyst Afterwards, excess water from the sample was evaporated using a rotary
for TRM. Most of the studies have utilized Ni as active phase of catalyst vacuum evaporator. The sample was dried overnight at 110 °C in an
owing to its high activity for conversion and low cost; however, there is oven. Finally, the solid was calcined at 900 °C for 5 h in the stationary air
no agreement over the role of support. Al2O3, CeO2, SiO2, β–SiC and with a heating ramp of 10 °C/min. The desired Ni loading was 10 wt%.
ZrO2 were attempted as catalyst support materials for TRM [7,10–13],
whereas, several other combinations of metal oxides like MgxTi1−xO,
2.2. Catalyst characterization
(Ce-Zr)O2 and YSZ-CeO2 were also employed [14–16]. However, de-
spite these studies, there is no clear relationship between the physico-
N2 physisorption test was carried out using a Micromeritics ASAP
chemical properties of the catalysts and their activity and stability. The
2010 apparatus at −196 °C for determination of BET surface area, pore
aim of the present study is to formulate Ni-based catalysts by employing
volume, and average pore diameter of the catalysts. Prior to gas ad-
potential metal oxides Al2O3, ZrO2, TiO2, SBA-15, MgO and CeO2-ZrO2
sorption measurements, the sample was degassed by heating it at
as support phase and to demystify the role of support materials for
200 ± 1 °C under vacuum for a period of 5 h. The elemental compo-
TRM. These metal oxides are strong candidates for support phase of
sition of reduced catalyst and surface morphology of samples was in-
TRM catalyst owing to their high hydrothermal stability, a high re-
vestigated using a ZEISS EVO Series Scanning Electron Microscope
sistance against oxidation and average to high surface area. The cata-
equipped with energy dispersive X-ray (EDX) analyzer. For both pur-
lysts were synthesized by wet-impregnation method and support effect
poses, the powdered sample was mounted on carbon tape attached to
on its performance has been investigated in detail. The content of CO2,
the sample holder. TEM was performed with FEI TecnaiTF20 micro-
H2O, and O2 in feed for the tri-reforming experiment was set to match
scope operated at an accelerating voltage of 100 kV. The images of
that of flue gas from gas-fired power plants to ensure flue gas utiliza-
reduced and spent catalysts were captured in order to observe Ni par-
tion. Also, the tri-reforming reaction was operated in a sub-stoichio-
ticle size. X-ray diffraction patterns were recorded on Rigaku Miniflex
metric condition (i.e. more methane than oxidants) in order to make
600 at ambient temperature in the 2θ range from 10° to 90° using Cu Kα
reaction conditions more severe towards coking to better judge dif-
radiation generated at 40 kV and 15 mA. Scans were taken at a speed of
ferent Ni-supported catalysts' stability against the most common cata-
4°/min with step width 0.02°. The crystalline phases were identified
lyst deactivation phenomena in methane reforming processes i.e.
using a database compiled by the Joint Committee on Powder
carbon deposition. The present work, particularly, examined spent
Diffraction Standards (JCPDS). Ni crystallite size was calculated by the
sample of each catalyst in detail in order to probe any sign of catalyst
Scherer equation (d = kλ / βcosθ); where d is the mean crystallite
deactivation, coking, sintering, and re-oxidation. diameter, k is shape factor (0.9), λ is the X-ray wavelength (1.54 Å), β is
the full width at half maximum (FWHM) of the diffraction peak at 44.5°
2. Experimental and θ is the peak position. The small-angle X-ray scattering diffraction
pattern was recorded using SAXSess mc2 Anton Paar Gmbh instrument
2.1. Catalyst preparation equipped with CCD detector having Cu Kα radiation source operating at
40 kV and 50 mA. The samples were scanned over a 2θ range from 0.5°
Six different batches of supported Ni catalysts with metal oxides to 4°. Reduction behavior of the catalyst was examined by the tem-
Al2O3, ZrO2, TiO2, SBA-15, MgO and CeO2-ZrO2 (CeO2/ZrO2 mass perature programmed reduction (TPR), using Micromeritics Chemisorb
ratio = 1.0) as support phase were prepared. Pure Al2O3 (Sud Chemie), 2720 instrument equipped with thermal conductivity detector (TCD).
ZrO2 (Loba Chemie, ≥97% assay), TiO2 (Sigma-Aldrich, ≥99.5% assay) The fresh calcined catalyst was taken in a U-shaped tube and degassed

41
R. Kumar et al. Fuel Processing Technology 186 (2019) 40–52

at 200 °C for 2 h under a flow of argon gas to remove moisture and other rDeact.=Drop in CH 4 conversion rate after 10 h of TOS
volatile compounds present in the sample and then cooled down to
= rCH 4 0.3 h rCH 410 h
room temperature. Following this, the gas flow was switched to a
stream of 10.2 vol% H2 in Ar at 30 ml/min and the temperature of the FH2, out
sample was raised to 900 °C at a ramp rate of 10 °C/min. The change in Molar H2 /CO ratio =
FCO, out
H2 content was measured by TCD and plotted against temperature. The
acidity of the reduced catalysts was measured by temperature pro- where, Fi, in/out corresponds to the molar flow rate of each component at
grammed desorption (TPD) of ammonia using the same equipment. The the inlet (in) and at the outlet (out) of the reactor. Wcat. is the weight (g)
sample pre-treatment procedure was the same as mentioned for TPR. of the catalyst loaded in the reactor.
An ammonia–helium mixture (5:95) was passed through the sample at
50 °C till complete saturation. Physically bound ammonia was removed 3. Results and discussion
first by flowing helium for 45 min at the same temperature. After that,
TPD was carried out by heating the sample from ambient temperature 3.1. Catalyst characterization
to 800 °C at a ramp rate of 10 °C/min under the flow of helium at 30 ml/
min. NH3 desorption was observed using a TCD and the total amount of 3.1.1. Characterization of calcined catalysts
desorbed NH3 was calculated from the integrated peak area relative to Fig. 1 shows XRD patterns of calcined catalysts. For Ni/ZrO2, Ni/
the calibration curve. The basicity of the reduced catalysts was mea- SBA-15 and Ni/CeO2-ZrO2 catalysts, cubic NiO peaks corresponding to
sured by temperature programmed desorption (TPD) of carbon dioxide (1 1 1), (2 0 0), (2 2 0), (3 1 1) and (2 2 2) crystal planes are observed at
using the same equipment. The procedure was also similar to NH3-TPD 37.2°, 43.3°, 62.8°, 75.4° and 79.2° respectively (JCPDS 78-0643). On
expect that the samples were exposed to a carbon dioxide –helium the other hand, Ni/MgO, Ni/Al2O3 and Ni/TiO2 do not feature any NiO
mixture (5:95). Ni metal dispersion of reduced catalyst was determined peak. The diffraction peaks of NiO and MgO appear at nearly same 2θ
by standard H2 – pulse chemisorption using the same apparatus. First, values (JCPDS 78-0643 and 78-0430 respectively). Both MgO and NiO
the sample was pre-treated in a manner as discussed above for TPR. are face-centered-cubic oxides and have very close ionic radii (0.070
Pulses of 1 cm3 pure H2 were injected from syringe till saturation. The and 0.065 nm for Ni2+ and Mg2+), lattice parameters (4.196 and
area of peaks was used to calculate the total volume of H2 chemisorbed; 4.212 Å for NiO and MgO respectively) and bond distances (2.10 and
which was further used to calculate Ni metal dispersion under the as- 2.11 Å for NiO and MgO respectively) [19]. At higher calcination
sumption that H2/Ni stoichiometry is equal to 2. TGA analysis on TA temperature, NiO usually diffuses into MgO lattice and forms a bulk
instrument (SDT Q600) was carried out to quantify carbon deposition NiO-MgO solid solution [20]. Therefore, all the five peaks appeared in
on spent catalyst. The measurement was performed in flowing air in the XRD pattern of Ni/MgO are assigned to NiO-MgO solid solution. Cal-
temperature range of 20 °C–800 °C with a heating rate of 10 °C/min. cined Ni/Al2O3 displayed a set of diffraction peaks centered at 19.5°,
31.8°, 37.2°, 45.3°, 60.1°, and 66.2°, which are assignable to NiAl2O4
2.3. Catalyst testing cubic spinel-type structure (JCPDS 78-1601). This crystalline phase
usually forms due to the solid-solid reaction between NiO and Al2O3 at
The activity of the catalysts for TRM was studied using a stainless higher calcination temperature [21]. XRD pattern of Ni/TiO2, not fea-
steel fixed-bed reactor (i.d. = 8 mm) at 800 °C temperature, weight turing any specific peak corresponding to NiO, represents reflections
hourly space velocity (WHSV) of 17,220 mL h−1 g−1 under atmospheric corresponding to ilmenite-type NiTiO3 having rhombohedral structure
pressure. Prior to the reaction, the as-synthesized catalyst was reduced in and rutile TiO2 (JCPDS file no. 85-0451 and 78-1510). Calcination
situ under a flowing gaseous mixture of 50% H2 and 50% N2 (total flow temperature above 500 °C facilitated a solid-solid reaction between NiO
rate = 40 ml/min) for 4 h. The reduction temperature for each of the and TiO2 to form phase-pured yellow NiTiO3 [22]. Another noteworthy
catalysts was identified from its TPR profile. For the TRM reaction, water feature of its XRD spectra is the absence of reflections corresponding to
was pumped into a vaporizer at 200 °C and allowed to mix with a gaseous anatase TiO2 which was present predominantly than rutile TiO2 in XRD
mixture of CH4, CO2, O2 and N2 in a pre-mixer maintained at 300 °C. The pattern (Fig. S1) of pure TiO2. Probably, either only anatase form of
final mixture was fed into the reactor placed in a tubular furnace. A TiO2 would have reacted with NiO during calcination [22] or remaining
heating tape was used to heat reactant line from vaporizer to reactor inlet anatase TiO2 might have transformed into more stable rutile form [23].
to prevent steam condensation. The molar composition of gaseous re- The XRD spectra show only one broad diffraction peak at 22.7° in
actants was taken as CH4:CO2:H2O:O2:N2 = 1:0.23:0.46:0.07:0.28. The Fig. 1(d) that corresponds to the amorphous silica phase of Ni/SBA-15
nitrogen was used as an internal standard for subsequent gas chroma- [24].
tograph analysis. The catalyst was placed between two pieces of quartz SAXS patterns shown in Fig. 2 represent characteristic of the 2D
wool in the center of the reactor. Effluents were passed through a con- hexagonal arrangement of cylindrical mesopores of SBA-15 based cat-
denser maintained at ~5 °C. The gaseous products were quantitatively alysts. The intense peak at 0.9° and two well-resolved small peaks at
analyzed using a gas chromatograph equipped with TCD and Porapak Q 1.6° and 1.9° correspond to the (1 0 0), (1 1 0) and (2 0 0) diffraction
and Molecular Sieve 13× columns in series. The condensed water was planes respectively for SBA-15 support material [25]. SAXS spectrum of
collected in a gas-liquid separator and its amount was measured at the the calcined sample indicates a decrease in intensity of diffraction peaks
end of TRM reaction. The quantitative analysis of gaseous effluents was at 1.6° and 1.9°; suggesting the partial collapse of the SBA-15 structure
carried out in every 20 min; typical duration of each test was 10 h. The during calcination. It could be attributed to partial dissolution of the
mass balance for all the runs was carried out and experiments, where C silica wall as its thickness decreased and intensity of amorphous silica
balance was < 97%, were discarded. Gaseous reactant's conversion rate increased as observed in high angle XRD pattern [25]. The position of
(rCH4/CO2/O2, mmol/gcat·s), average water conversion rate (rH2O, (1 0 0) diffraction peak shifted to higher angle; which may be attributed
mmol/gcat·s), catalyst deactivation rate (rDeact., mmol/gcat·s) and molar to the decrease in pore size due to the incorporation of Ni species in
H2/CO ratio were calculated according to following equations: mesoporous silica. The TPR profiles, shown in Fig. 3, reveal reduction
properties of calcined catalysts. These are indicative of temperatures at
FCH4/CO2/O2, in FCH4/CO2/O2, out which in situ reduction of calcined catalysts was carried out before
rCH 4 /CO2 /O2 =
Wcat . catalyst activity test. The reduction temperature of each catalyst has
been mentioned in Table 1.
Total H2 O fed in 10 h Total H2 O collected after 10 h
rH2 O = Generally, the weaker interaction between metal and support cause
Wcat . reduction of nickel oxide species at a lower temperature and stronger

42
R. Kumar et al. Fuel Processing Technology 186 (2019) 40–52

Fig. 1. XRD patterns of calcined catalysts: (a) Ni/Al2O3, (b) Ni/CeO2-ZrO2, (c) Ni/MgO, (d) Ni/SBA-15, (e) Ni/TiO2 and (f) Ni/ZrO2.

metal-support interaction lead to a reduction of nickel oxide species at a Ni/ZrO2 can be attributed to the reduction of NiO species. The presence
higher temperature. TPR spectrum of pure NiO shows its reduction at of NiO in calcined Ni/CeO2-ZrO2 and Ni/ZrO2 has already been con-
427 °C (Fig. S2). firmed by XRD results. The results of Table 1 indicate that both catalysts
Based on these information, the metal-support interactions (MSI) attained a high degree of reducibility. TPR profile of Ni/MgO shows a
have been categorized here as: weak (WMSI, reduction reduction peak at 743 °C with reduction occurring over a temperature
temperature < 500 °C), medium (MMSI, 500 °C < reduction range from 600 °C to 800 °C. It was mainly NiO species reducing in this
temperature < 600 °C) and strong (SMSI, reduction temperature range. The possibility of MgO reduction in above-specified
temperature > 600 °C). The degree of reducibility of each catalyst has temperature range is ruled out because of its high heat of formation
been reported in Table 1. Based on the TPR profiles of the catalysts above 420 kJ/mol [27]. Unlike Ni/CeO2-ZrO2 and Ni/ZrO2, NiO of
shown in Fig. 3, it can be established that Ni/CeO2-ZrO2 and Ni/ZrO2 calcined Ni/MgO required very high temperature for reduction because
had WMSI and MMSI respectively. While, Ni/Al2O3, Ni/MgO, Ni/SBA- of its strong interaction with MgO, which originated from NiO - MgO
15 and Ni/TiO2 pursued SMSI. TPR spectrum of Ni/ZrO2 shows an in- solid solution. In this way, TPR result corroborates XRD spectra of this
tense reduction peak at around 542 °C. The medium interaction could catalyst. The formation of solid solution rendered their surface segre-
contribute to the formation of Ni-ZrO2 interfaces [26]. TPR profile of gation and the consequent reduction very difficult [20]. It also validates
Ni/CeO2-ZrO2 displays a very sharp reduction peak having a maxima at the very low degree of reducibility of Ni/MgO as reported in Table 1.
490 °C. These peaks observed in the TPR profile of Ni/CeO2-ZrO2 and Similarly, Ni/TiO2 also attained a low degree of reducibility. TPR

Fig. 2. SAXS patterns of SBA-15 based samples: (a) SBA-15, (b) calcined Ni/ Fig. 3. TPR profiles of calcined catalysts: (a) Ni/Al2O3, (b) Ni/CeO2-ZrO2, (c)
SBA-15, (c) reduced Ni/SBA-15 and (d) spent Ni/SBA-15. Ni/MgO, (d) Ni/SBA-15, (e) Ni/TiO2 and (f) Ni/ZrO2.

43
R. Kumar et al. Fuel Processing Technology 186 (2019) 40–52

Table 1
Physico-chemical properties of catalysts.
Ni/Al2O3 Ni/CeO2-ZrO2 Ni/MgO Ni/SBA-15 Ni/TiO2 Ni/ZrO2

Reduction temperature (°C) 930 700 800 870 960 700

Degree of reducibility (%) 85 94 16 82 43 89
Ni loading (wt%) 9.2 13.1 10.5 9.0 10.4 10.4
BET surface area (m2/g) 97 4 2 36 n.o. 1
Ni crystallite size from XRD (nm) 17.0 64.1 n.o. 13.2 31.9 32.4
Ni particle diameter from TEM (nm) 16.6 59.5 21.0 14.4 n.a. 32.6
Ni dispersion (%) 2.4 1.5 0.6 2.2 1.7 2.0
Basic sites (μmol/g) 31.4 0.8 7.9 3.7 n.o. 2.4
Acidic sites (μmol/g) 12.4 0.6 n.o. n.o. n.o. 0.8

n.o.: not observed, n.a.: not analyzed.

profile of Ni/TiO2 revealed a small reduction peak centered at 888 °C, might be possible that SMSI could have provided strong resistance
clearly indicating very strong metal-support interaction. It could be against Ni metal sintering during in situ catalyst reduction prior to
attributed to the reduction of perovskite NiTiO3 [22]. The absence of reaction. On the contrary, WMSI could not be able to avoid Ni metal
any other reduction peak confirms that Ni was present in form of particle agglomeration and hence resulted in larger Ni crystals. The
NiTiO3 only as complemented by XRD result. A reduction peak centered highest Ni dispersion was achieved over Al2O3 followed by SBA-15
at 850 °C can be observed in the TPR profile of NiO/Al2O3. This high- owing to smaller Ni particle size. While Ni/ZrO2 had satisfactory dis-
temperature peak could be attributed to the reduction of spinel NiAl2O4 persion value due to a higher degree of reducibility. The low Ni dis-
[21]. Like Ni/MgO, NiO species of calcined Ni/SBA-15 also shows re- persion of Ni/TiO2 and Ni/MgO could be attributed to their low degree
duction peak with maxima appearing at a higher temperature, 767 °C. It of reducibility. The larger Ni particles with lower surface area and
could be attributed to the reduction of NiO species occluded inside the limited porosity of CeO2-ZrO2 might have led to relatively lower Ni
channels of SBA-15. NiO species might have diffused into the channels dispersion in Ni/CeO2-ZrO2. The size of Ni crystals as measured from
during calcination and remained confined inside the channels of silica, TEM images (Fig. 4) are comparable to those calculated using XRD.
reduction of such confined NiO species generally requires high tem- Fig. 4(d) displays a well-ordered hexagonal array of cylindrical pores of
perature [28]. Although, it could be also possible that some metal- SBA-15; complementing its SAXS patterns for mesoporous structure.
support compound might have formed during high-temperature calci- The histograms reveal a narrower distribution of Ni particle size over
nation as was the case for Ni/Al2O3 and Ni/TiO2. Nickel-silica com- SBA-15 compared to other catalyst supports. It could be attributed to
pound like 2:1 nickel phyllosilicate or Ni2SiO4 are highly thermostable high surface area and confinement effect of SBA-15. The Ni particle size
and their reduction also require very high temperature [28,29]. How- of Ni/TiO2 could not be measured using its TEM images. The shape of
ever, since no peak other than corresponding to NiO is observed in its Ni particles appears very far from circularity because of which precise
XRD pattern; this possibility is ruled out here. Table 1 reveals the de- measurement of Ni particle size is not possible. The non-circularity of
gree of reducibility of Ni/Al2O3 and Ni/SBA-15 was sufficiently high Ni particles could be due to their partial blockage by TiOx. In high-
but less than that of Ni/CeO2-ZrO2 and Ni/ZrO2. Based on the TPR temperature H2 – reduction environment, TiO2 tends to undergo partial
results for metal-support interaction and degree of reducibility, it can reduction to TiOx (x < 2) and then migrate onto the metal leading to
be concluded that weak/medium interaction allows easier reduction of its partial coverage [30]. Fig. 5 shows XRD patterns of reduced cata-
Ni2+ to Ni0 and hence relatively higher degree of reducibility can be lysts. The XRD spectra of Ni/Al2O3, Ni/CeO2-ZrO2, Ni/SBA-15, Ni/TiO2
achieved. On the contrary, strong metal-support can result in either and Ni/ZrO2 display metallic Ni peaks at about 44.6°, 51.9°, and 76.4°
high or low degree of reducibility depending on the compound in which corresponding to (1 1 1), (2 0 0) and (2 2 0) crystal planes respectively
Ni2+ phase exists. (JCPDS 89–7128). On the contrary, no diffraction peak of Ni is ob-
served for Ni/MgO in Fig. 5(c). Ni might be present in traces, which
could be below the detection limit of XRD. The TPR profile of Ni/MgO
3.1.2. Characterization of reduced catalysts has already indicated very low degree of reducibility. On combining
Table 1 summarizes the physicochemical properties of all the re- TPR and XRD results, it can be concluded that the reduction of calcined
duced catalysts. The final Ni content of the catalyst was 10.5%, 9.2%, Ni/MgO catalyst sample generated a quite few of metallic Ni particles.
10.4%, 13.1%, 10.4% and 9.0% in Ni/MgO, Ni/Al2O3, Ni/ZrO2, Ni/ The peaks corresponding to Ni can be found in XRD spectra of Ni/
CeO2-ZrO2, Ni/TiO2 and Ni/SBA-15 respectively. BET surface area of Al2O3; suggesting reduction of NiAl2O4 to Ni and Al2O3. However, it is
Ni/Al2O3 and Ni/SBA-15 was higher than the rest of the samples. It difficult to assign peaks corresponding to Al2O3 in Fig. 5(a). It is be-
could be ascribed to the high surface area of Al2O3 and SBA-15. BET cause γ-Al2O3 peaks, appear very close to the diffraction peaks of
surface area, pore volume and average pore diameter of catalyst sup- NiAl2O4 [21]. Also, a reflection at 33.0° can be observed which corre-
ports are reported in Table S1. The surface area of Ni/TiO2 vanished sponds to neither γ-Al2O3 nor Ni. It is the characteristic peak of θ-Al2O3
completely, as reported in Table 1. The relatively low surface area of (JCPDS 35-0121). The occurrence of this phase can be attributed to
the catalysts was due to the high calcination and reduction temperature high reduction temperature, 930 °C. Although the literature suggests its
incorporated in the catalyst preparation. The Ni crystallite size obtained formation require 1000 °C, it could be the reason for very low intensity
from Debye-Scherrer equation using line broadening of (1 1 1) reflec- of this peak [31].
tion from XRD patterns was found to vary with catalyst support from CO2 - TPD of reduced catalysts were carried out in order to quantify
13.2 nm to 64.1 nm. The smallest one corresponds to SBA-15 supported the surface basicity of reduced catalysts and the results are given in
catalyst and the largest one for the CeO2-ZrO2 supported catalyst. The Fig. 6. Since one of the reactants was CO2, which is mildly acidic, the
Ni crystallite size has following order: Ni/SBA-15 < Ni/Al2O3 < Ni/ presence of basic sites would be helpful in terms of CO2 adsorption and
MgO < Ni/TiO2 < Ni/ZrO2 < Ni/CeO2-ZrO2. It is noteworthy to activation. Weaker basic sites cause desorption of adsorbed CO2 at
mention here that Ni crystallite size of Ni/CeO2-ZrO2 and Ni/ZrO2 was lower temperatures, whereas stronger basic sites lead to desorption of
larger despite their relatively lower reduction temperature. It appears CO2 at relatively higher temperatures [32]. The strength of basic sites
that MSI strongly influences Ni crystallite size. The WMSI generates shown in Fig. 6 is categorized here as: weak basic sites (< 300 °C),
larger Ni crystals, whereas, SMSI leads to smaller Ni crystallite size. It

44
R. Kumar et al. Fuel Processing Technology 186 (2019) 40–52

Fig. 4. TEM images of reduced catalysts: (a) Ni/Al2O3, (b) Ni/CeO2-ZrO2, (c) Ni/MgO, (d) Ni/SBA-15, (e) Ni/TiO2 and (f) Ni/ZrO2.

appeared in temperature zone corresponding to medium basic strength.

These basic sites can be assigned to metal-oxygen pairs (such as Al-O
and Zr-O) [36]. Ni/CeO2-ZrO2, Ni/SBA-15, and Ni/Al2O3 pursued
weak-strength basic sites also, as observed in Fig. 6. No desorption peak
is observed for Ni/TiO2 catalyst. The absence of surface basicity of TiO2
supported Ni catalyst can be related to its high electronegativity [31]. It
could also be resulted owing to complete loss of its BET surface area.
Although only Ni/MgO had strong-strength basic sites, Ni/Al2O3 pre-
sented the highest concentration of basic sites. Han et al. also reported
higher amount of desorbed CO2 for Al2O3 overlayer on Ni compared to
MgO overlayer on Ni in their study [37]. As reported in Table 1, the
decreasing order of basic sites concentration is as follows: Ni/
Al2O3 > Ni/MgO > Ni/SBA-15 > Ni/ZrO2 > Ni/CeZrO2 > Ni/
TiO2.
NH3-TPD profiles of all the reduced catalysts are presented in Fig. 7.
NH3-TPD gives information about the acidic properties of catalysts.
Among the samples, Ni/MgO, Ni/SBA-15 and Ni/TiO2 did not feature
any desorption peak; indicating the absence of acid sites. Although the
literature does suggest Lewis acidity associated with Ti4+ centers in
Fig. 5. XRD patterns of reduced catalysts: (a) Ni/Al2O3, (b) Ni/CeO2-ZrO2, (c)
TiO2-based samples [38]; but, in the present study, BET surface area of
Ni/MgO, (d) Ni/SBA-15, (e) Ni/TiO2 and (f) Ni/ZrO2. reduced Ni/TiO2 vanished which might have led to the absence of
surface acidity. The generation of surface acidity in Ni/CeO2-ZrO2
catalyst is attributed to the incorporation of Zr4+ into the CeO2 matrix
medium basic sites (> 300 °C– < 550 °C) and strong basic sites
according to Kung's model. Lewis acidity is assigned to electronically
(> 550 °C). The basic sites concentration of each catalyst is reported in
deficient Zr4+ centers [39]. Ni/Al2O3 presented a broad desorption
Table 1. The support materials used in this study for Ni-based catalyst
peak centered at 216 °C along with a shoulder centered at around
formulation significantly influence the strength and concentration of
360 °C. These have been assigned to surface hydroxyl group and un-
basic sites. CO2-TPD profile of Ni/Al2O3 shows a small desorption peak
dercoordinated Lewis acidic Al cations respectively [40]. Likewise, Ni/
corresponding to weak- strength basic sites appearing at 194 °C. This
ZrO2 catalyst also displayed peaks ascribed to surface hydroxyl group
weak- strength basic sites can be traced as bicarbonates, which are
and Lewis acidic Zr cation at 222 °C and 420 °C respectively. Ni/ZrO2
formed because of the interaction between surface hydroxyl groups and
had slightly higher number of acid sites compared to Ni/CeO2-ZrO2.
CO2 [33]. Among all the samples, Ni/MgO was the only catalyst having
While, Ni/Al2O3 presented the highest concentration of acid sites, as
strong-strength basic sites, which appears at 644 °C. Unidentate car-
reported in Table 1.
bonates, which formed from low-coordination isolated surface O2– an-
ions, were the source of high-strength basic sites. Such low-coordina-
tion O2– ions are usually present on the surface of MgO [34]. However, 3.1.3. Characterization of spent catalysts
a major desorption peak for Ni/MgO is observed at 363 °C along with The main properties of catalysts after TRM reaction are given in
shoulder centered at 459 °C. This peak, corresponding to medium basic Table 2. Ni crystallite size of most of the catalysts was found to increase
strength, can be attributed to Mg-O metal-oxygen pairs [35]. The after run time of 10 h. However, it increased significantly in the case of
shoulder could be related to those Mg-O pairs, which would have in- Ni/CeO2-ZrO2 from 64.1 nm to 94.2 nm.
volved in solid solution formation with NiO species. The main deso- High-temperature metal sintering and agglomeration of Ni particles
rption peak for Ni/CeO2-ZrO2, Ni/SBA-15, Ni/Al2O3, and Ni/ZrO2 during tri-reforming reaction probably resulted in much larger Ni
crystals over CeO2-ZrO2 support. On the other hand, SBA-15 supported

45
R. Kumar et al. Fuel Processing Technology 186 (2019) 40–52

Fig. 6. CO2 – TPD profiles of reduced catalysts: (a) Ni/Al2O3, (b) Ni/CeO2-ZrO2, (c) Ni/MgO, (d) Ni/SBA-15, (e) Ni/TiO2 and (f) Ni/ZrO2.

catalyst resulted in a strong resistance against Ni sintering as the neg- Ni/SBA-15, and Ni/TiO2 underwent Ni re-oxidation to some extent. Ni/
ligible increase was found in its Ni crystallite size. It appears that the SBA-15 shows reflections corresponding to NiO, as seen in Fig. 8(d).
confinement of Ni particles in the cylindrical pores of SBA-15 avoided While XRD spectra of Ni/TiO2 shows peaks at 37.4° and 43.5° corre-
sintering of Ni particles. Ni/Al2O3, Ni/TiO2, and Ni/ZrO2 underwent Ni sponding to NiO (1 1 1) and NiO (2 0 0) and one very small reflection at
metal sintering to some extent as 5 to 7 nm increase in their Ni metal 33.3° which is a characteristic peak of NiTiO3 suggesting that some of
crystallite size were observed. Results from TPR, XRD, and TEM reveal re-oxidized Ni species converted to NiTiO3 during the reaction. The
that metallic Ni generated upon reduction of calcined catalysts, which peaks corresponding to NiAl2O4 reappeared at 37.1°, 45.1°, and 60.2° in
had SMSI and MMSI, were more resistant to sintering. On the other XRD pattern of Ni/Al2O3, indicating the formation of this spinel com-
hand, catalysts with WMSI were prone to metal sintering. The Ni pound during the reaction. On the other hand, Ni/ZrO2 and Ni/CeO2-
crystallite size of Ni/MgO could not be calculated owing to the dis- ZrO2 did not undergo Ni re-oxidation. Another noteworthy observation
appearance of Ni (1 1 1) reflection at 44.6°, as shown in Fig. 8. Ni is the emergence of a new diffraction peak at 26.5° in XRD spectra of
species would have re-oxidized during the reaction. However, the other Ni/Al2O3, Ni/SBA-15, and Ni/ZrO2. This peak is an indicative of crys-
spent catalysts display reflections corresponding to Ni (1 1 1), (2 0 0) talline graphite [41] and suggests the formation and deposition of
and (2 2 0) crystal planes in their XRD spectra; suggesting metallic Ni graphitic carbon over these catalysts during the reaction. The rate of
species remained intact after a run time of 10 h. Although, Ni/Al2O3, carbon deposition over each catalyst, as calculated from TGA, is

Fig. 7. NH3 – TPD profiles of reduced catalysts: (a) Ni/Al2O3, (b) Ni/CeO2-ZrO2, (c) Ni/MgO, (d) Ni/SBA-15, (e) Ni/TiO2 and (f) Ni/ZrO2.

46
R. Kumar et al. Fuel Processing Technology 186 (2019) 40–52

Table 2
Physico-chemical properties of spent catalysts after 10 h of time on stream study.
Ni/Al2O3 Ni/CeO2-ZrO2 Ni/MgO Ni/SBA-15 Ni/TiO2 Ni/ZrO2

Ni crystallite size from XRD (nm) 23.2 94.2 n.o. 13.7 38.4 37.1
Ni particle diameter from TEM (nm) 23.8 89.3 n.o. 14.2 39.1 36.0
Carbon deposition rate (gc/gcat.h) 14.0E−03 n.o. 1.4E−03 0.1E−03 0.02E−03 3.6E−03

n.o.: not observed.

Fig. 8. XRD patterns of spent catalysts after 10 h of time on stream study: (a) Ni/Al2O3, (b) Ni/CeO2-ZrO2, (c) Ni/MgO, (d) Ni/SBA-15, (e) Ni/TiO2 and (f) Ni/ZrO2.
+graphitic carbon, #NiAl2O4, *NiO and −NiTiO3.

reported in Table 2. It was the highest over Ni/Al2O3 followed by Ni/ average H2O conversion rate and molar H2/CO ratio. Fig. 11 displays
ZrO2. In comparison, Ni/MgO, Ni/SBA-15 and Ni/TiO2 exhibited very H2 and CO formation rates with time on stream. The trends of CH4 and
low carbon deposition rate, suggesting negligible coke formation during CO2 conversion rates and those of H2 and CO conversion rates are
the reaction. On the contrary, Ni/CeO2-ZrO2 did not display any weight nearly same; hence, experimental data of conversion rates and forma-
loss and hence no carbon accumulation occurred during the reaction. tion rates complement each other. Approximately complete conversion
There was an increase in its weight during TGA analysis, which could of O2 was observed for all the catalysts. Fig. 10 and Table 3 show a
be attributed to the re-oxidation of metallic Ni in Ni/CeO2-ZrO2 [42]. pronounced difference in the activity of Ni-supported catalysts; sug-
Among the number of TEM images of spent Ni/Al2O3 sample shown gesting the nature of catalyst supports strongly influences catalyst
in Fig. S3, quite a few of them have shown filamentous carbon. By performance in TRM. The catalyst activity results reveal that Ni/Al2O3
combining this observation with XRD result and TGA value of spent Ni/ had higher CH4 and CO2 conversion rates whereas the performance of
Al2O3 it can be concluded that most of the carbon accumulated over Ni/ MgO supported Ni catalyst was poor. Based on CH4 conversion rate, the
Al2O3 during the reaction was of amorphous nature. TEM images of order of initial activity for TRM was as follows: Ni/Al2O3 > Ni/SBA-
spent catalysts as shown in Fig. 9 were used to measure the average Ni 15 > Ni/ZrO2 > Ni/CeO2-ZrO2 > Ni/TiO2 > Ni/MgO. Ni/Al2O3
crystallite size of the catalysts after a run time of 10 h. The data are catalyst's initial activity was the highest with CH4, CO2, and H2O con-
presented in Table 2 and are in good agreement with Ni crystallite size version rate 8.72 × 10−2 mmol/gcat·s, 2.31 × 10−2 mmol/gcat·s and
obtained from the XRD spectrum. The particle size histograms of Ni/ 4.2 × 10−2 mmol/gcat·s respectively. The superior activity of Ni/Al2O3
MgO and Ni/TiO2 are not provided in Fig. 9. No Ni species was found in catalyst can be attributed to smaller Ni particle size, higher degree of
TEM images of Ni/MgO owing to its re-oxidation during the reaction; reducibility, higher Ni dispersion, and stronger metal-support interac-
while in case of Ni/TiO2 the shape of Ni particles still remained non- tion. Also, the literature suggests that Ni, with particle size < 25 nm,
circular even after reaction; suggesting that its partial coverage by TiOx consists of edge and corner atoms more predominantly than face atoms
persisted during the reaction. as surface atoms. While the share of face atoms (111) and (100) in the
total number of surface atoms becomes maximum only when Ni particle
3.2. Comparison of different catalysts for TRM size is above 25 nm [43]. In our study, Ni crystallite size of Ni/Al2O3
catalyst was 17 nm as evident from XRD and TEM results and remained
3.2.1. Catalyst activity below 25 nm even after an experimental run of 10 h, indicating a good
The activity of all the catalysts was compared under identical re- share of edge and corner atoms in the total number of surface atoms for
action conditions of TRM and results are shown in Fig. 10 in terms of Ni/Al2O3. These low-coordinated (edge and corner) surface atoms favor
CH4 and CO2 conversion rates with time on stream. Table 3 reports the catalyst activity in structure sensitive methane reforming reactions

47
R. Kumar et al. Fuel Processing Technology 186 (2019) 40–52

Fig. 9. TEM images of spent catalysts after 10 h of time on stream study: (a) Ni/Al2O3, (b) Ni/CeO2-ZrO2, (c) Ni/MgO, (d) Ni/SBA-15, (e) Ni/TiO2 and (f) Ni/ZrO2.

[43]. All these factors resulted in higher activity of Ni/Al2O3 compared 1.77 × 10−2 mmol/gcat·s respectively. The relatively lower perfor-
to other tested catalysts. mance of this catalyst can be attributed to its larger Ni particles
In the first study on the effect of catalyst support for TRM, Song 64.1 nm. The concentration of basic sites was also low as evident from
et al. also reported high activity of Ni/Al2O3 [6]. On the contrary, the CO2-TPD profile, justifying its low CO2 conversion. The average
García-Vargas et al. reported poor performance of Al2O3 as Ni catalyst H2O conversion rate was approximately 3.3 × 10−2 mmol/gcat·s. The
support while studying the influence of Al2O3, CeO2, β-SiC and YSZ as probable reason for relatively lower water conversion could be that
support on catalytic behavior of Ni catalyst for TRM process [44]. They ceria, initially present in its reduced state, might have interacted more
cited low reducibility of Ni/Al2O3 owing to the formation of NiAl2O4. with oxygen than steam owing to the high reactivity of oxygen. Oxygen
NiAl2O4 has formed in the present study also. It is because in both the activation might have filled oxygen vacancy of CeO2-ZrO2 and hence
studies the catalyst was calcined at 900 °C. However, García-Vargas not allowed it to remain in the reduced state. Kundakovic et al. reported
et al. reduced Ni/Al2O3 catalyst at 700 °C prior to reaction. The 700 °C that dissociation of water happens only in presence of Ce3+ in Rh/CeO2
temperature appears insufficient for reduction as per their TPR profile. [48]. Song et al. also reported poor catalytic performance of Ni/CeZrO
On the other hand, the catalyst was reduced at 930 °C in the present in TRM reaction [6]. Ni/TiO2 and Ni/MgO indicated SMSI similar to
study, which resulted in sufficiently high reducibility to catalyze TRM Ni/Al2O3 and Ni/SBA-15 catalysts, but even though, failed to catalyze
reaction effectively. Although, Ni/SBA-15 had Ni particles smaller than TRM to a larger extent. This leads to an important conclusion that SMSI
that of Ni/Al2O3. However, the difference between Ni/Al2O3 and Ni/ is not the only necessary recipe for highly active catalyst development
SBA-15 in CH4 and CO2 conversion rates can be explained by the dif- for TRM. Both the catalysts had a low degree of reducibility and hence
ference in concentration of surface basic sites for CO2 activation. Al2O3 lower number of accessible Ni species. Ni/TiO2 had no basic sites for
supported catalyst had higher number of basic sites, which were re- CO2 activation as evident from the results of CO2-TPD. H2O conversion
sponsible for high CO2 conversion. Moreover, acidic supports also help rate was also low. In addition, the partial blockage of active Ni metal
in CO2 activation. Although, CO2 activation is weaker on acidic sup- species by TiOx made Ni particles encapsulated, as evident from its TEM
ports than that on basic supports. It occurs mainly via formation of images, which further confirms its low activity. On the other hand,
formates by CO2 reaction with the surface hydroxyls [45]. In addition, despite Ni/MgO being an only catalyst having strong basic sites, it could
CO2 adsorbs via a η2-configuration in which one oxygen atom bonded not perform good enough in our experimental study. It appears that the
to a Lewis-acidic surface metal atom of support and the carbon atom strong basicity of Ni/MgO could not compensate its low NiO re-
bonded to the Ni particles and thus the metal-oxide interface provides ducibility and hence resulted in poor catalytic performance. However,
active sites for CO2 activation [46]. Ni/Al2O3 had the highest number of among the several catalysts tested in this study, it attained the lowest
surface Lewis acid sites as revealed by NH3-TPD result. The low average H2/CO ratio, 1.81, as reported in Table 3, which could be due to its
water conversion rate, 3.1 × 10−2 mmol/gcat·s, over Ni/SBA-15 in- strong basicity, which would have favored CO2 activation and hence
dicates its relatively weaker ability for steam activation. Ni/ZrO2 ex- dry reforming (Eq. (2)) over other (steam and oxidative) reforming
hibited good performance for TRM; although, its CH4 and CO2 con- reactions (Eqs. (1) and (3)). However, Song et al. found MgO very
versions were lower than that of Ni/Al2O3 and Ni/SBA-15, probably suitable Ni-catalyst support as it displayed higher ability to enhance the
due to larger Ni particles and relatively weaker MSI. In addition, the CO2 conversion among Ni/ZrO2, Ni/Al2O3, Ni/MgO/CeZrO, and Ni/
basic site's concentration was relatively lower compared to Ni/Al2O3 CeO2 catalysts [6]. On the contrary, Jiang et al. and Singha et al. re-
and Ni/SBA-15 catalysts. The higher H2O conversion rate ported low activity of Ni/MgO because of low reducibility due to the
(3.6 × 10−2 mmol/gcat·s) over this catalyst may be attributed to the formation of NiO-MgO solid solution [14,49]. In the present study also,
high ability of ZrO2 for steam activation. Rossetti et al. also reported the low activity of Ni/MgO was attributed to its low degree of reducibility.
higher ability of Ni/ZrO2 to activate steam when they observed very
low CO selectivity for steam reforming of glycerol [47]. Also, its high 3.2.2. Catalyst deactivation
steam activation ability resulted in high H2/CO ratio, 2.24, as reported The deactivation rates (rDeact.) of the all tested catalysts are
in Table 3, due to the occurrence of steam reforming reaction (Eq. (1)). equivalent to drop in CH4 conversion rates over 10 h time-on-stream
Activity results reveal that the CH4 and CO2 conversion rates over Ni/ experimental run and are reported in Table 3. Ni/MgO and Ni/TiO2
CeO2-ZrO2 catalyst were 6.05 × 10−2 mmol/gcat·s and suffered relatively higher activity loss. While Ni/SBA-15 and Ni/ZrO2

48
R. Kumar et al. Fuel Processing Technology 186 (2019) 40–52

Fig. 10. Reactant conversion rates of different catalysts: (a) Ni/MgO, (b) Ni/TiO2, (c) Ni/CeO2-ZrO2, (d) Ni/ZrO2, (e) Ni/SBA-15 and (f) Ni/Al2O3: CH4 conversion
rate (A) and CO2 conversion rate (B) as function of time on stream (reaction conditions: CH4:CO2:H2O:O2:N2 = 1:0.23:0.46:0.07:0.28, WHSV = 17,220 mL h−1 g−1,
Temp. 800 °C at 1 bar).

exhibited remarkable stability against catalyst deactivation with time- catalyst had shown high resistance against coking owing to its basic
on-stream. Ni/MgO underwent Ni re-oxidation during TRM as evident nature which helped in gasifying deposited coke [4]. The re-oxidation
from TEM results and lost metallic Ni species active for TRM and this and sintering of Ni particles observed over Ni/TiO2 resulted in its ac-
could be the main reason for its activity loss. Although, the Ni/MgO tivity loss, however, the carbon deposition rate was negligible due to

Table 3
Activity data of different catalysts for TRM (reactions conditions: CH4:CO2:H2O:O2:N2 = 1:0.23:0.46:0.07:0.28, WHSV = 17,220 mL h−1 g−1, Temp. 800 °C at 1 bar).
Ni/Al2O3 Ni/CeO2-ZrO2 Ni/MgO Ni/SBA-15 Ni/TiO2 Ni/ZrO2

Average H2O conversion rate (mmol/gcat·s) × 102 4.2 3.3 1.4 3.1 2.8 3.6
rDeact. (mmol/gcat·s) × 102 0.74 0.89 1.1 0.33 1.0 0.64
H2/CO molar ratio 2.27 2.08 1.81 2.22 1.94 2.24

49
R. Kumar et al. Fuel Processing Technology 186 (2019) 40–52

Fig. 11. Product formation rates of different catalysts: (a) Ni/MgO, (b) Ni/TiO2, (c) Ni/CeO2-ZrO2, (d) Ni/ZrO2, (e) Ni/SBA-15 and (f) Ni/Al2O3: H2 formation rate
(A) and CO formation rate (B) as function of time on stream (reaction conditions: CH4:CO2:H2O:O2:N2 = 1:0.23:0.46:0.07:0.28, WHSV = 17,220 mL h−1 g−1, Temp.
800 °C at 1 bar).

the negligible acidity of Ni/TiO2 catalyst. Ni re-oxidation also occurred 0.64 × 10−2 mmol/gcat·s, which was lower than that of Ni/Al2O3, Ni/
in Ni/SBA-15 as evident from XRD results, but this catalyst displayed TiO2, Ni/MgO and Ni/CeO2-ZrO2. Although, size of its Ni particles in-
the least deactivation rate 0.33 × 10−2 mmol/gcat·s due to its strong creased from 32.4 nm to 37.1 nm and approximately 3.6E-03 gc/gcat.h
resistance to metal sintering and coking. The remarkable stability of Ni/ carbon deposition rate was observed after a run time of 10 h. However,
SBA-15 can also be realized by observing approximately constant H2 it was able to retain Ni in metallic state and no Ni re-oxidation occurred
and CO formation rates as shown in Fig. 11. Ni metal confinement in its during the reaction as evident from its XRD pattern, which could be
hexagonal pores seemed to resist metal sintering and the smaller Ni attributed to the oxophilic nature of ZrO2. The incompletely co-
particles suppressed carbon deposition. Since carbon formation reac- ordinated Zr4+ cations in the vicinity of Ni metal particles represents
tions are structure sensitive, therefore, it could be postulated that Ni oxophilic sites [51]. These oxophilic sites usually prefer to interact with
particle size below 15 nm is less prone to coking. Bukhari et al. also oxygen and hence did not allow O2 present in the feed to re-oxidize
reported that SBA-15 supported catalysts are stable in methane re- metallic Ni. Ni/CeO2-ZrO2 catalyst also had Ni species remained intact
forming processes [50]. The catalyst deactivation rate of Ni/ZrO2 was in metallic state during TRM. Despite this, it showed high deactivation

50
R. Kumar et al. Fuel Processing Technology 186 (2019) 40–52

rate 0.89 × 10−2 mmol/gcat·s owing to Ni metal particle sintering. The Acknowledgments
size of Ni particles increased from 64.1 nm to 94.2 nm during the re-
action. Compared to other catalysts, it represents very broad distribu- The authors thank the Federation of Indian Petroleum Industry
tion of Ni particle size as shown in Fig. 9. However, the redox property (FIPI), New Delhi, India for providing financial support. Rohit Kumar is
and oxygen storage capacity of CeO2-ZrO2 helped the catalyst not only grateful to Confederation of Indian Industry (CII), Gurugram, India &
in avoiding the Ni re-oxidation but also in gasifying the coke deposited Science and Engineering Research Board (SERB) for awarding him
during the reaction. The small peak in NH3-TPD indicates the presence Prime Minister's Fellowship for Doctoral Research.
of acid sites, which might have favored carbon deposition; nevertheless,
the stored oxygen in its lattice structure favored carbon gasification and Appendix A. Supplementary data
hence resulted in effective carbon removal. The redox nature of CeO2-
ZrO2 appears to protect metallic Ni from re-oxidation in oxygen-rich Supplementary data to this article can be found online at https://
reaction environment. The most active catalyst of this study Ni/Al2O3 doi.org/10.1016/j.fuproc.2018.12.018.
also underwent a few degree of activity loss as CH4 conversion rate
dropped by 0.74 × 10−2 mmol/gcat·s. This activity loss could mainly be References
attributed to coking over catalyst surface as evident from its TGA result.
The higher carbon deposition rate also led to higher H2/CO ratio 2.27, [1] S. Majhi, P. Mohanty, H. Wang, K.K. Pant, Direct conversion of natural gas to higher
as reported in Table 3. The acidic nature of Al2O3 may be responsible hydrocarbons: a review, J. Energy Chem. 22 (2013) 543–554.
[2] T. Yabe, K. Mitarai, K. Oshima, S. Ogo, Y. Sekine, Low-temperature dry reforming of
for the carbon formation [52]. The sintering and phase change from methane to produce syngas in an electric field over La-doped Ni/ZrO2 catalysts,
active metallic Ni to inactive NiAl2O4, as observed in characterization Fuel Process. Technol. 158 (2017) 96–103.
of its spent sample, were other possible reasons for deactivation. One [3] K.K. Sonal, S. Upadhyayula Pant, Synthesis of C5+ hydrocarbons from low H2/CO
ratio syngas over silica supported bimetallic Fe-Co catalyst, Catal. Today 291
noteworthy observation about Ni/Al2O3 catalyst is that it was still able (2017) 133–145.
to maintain CH4 and CO2 conversion rates higher than all other tested [4] A. Nandini, K.K. Pant, S.C. Dhingra, K-, CeO2-, and Mn-promoted Ni/Al2O3 catalysts
catalysts. This catalytic behavior of Ni/Al2O3 was probably due to the for stable CO2 reforming of methane, Appl. Catal. A Gen. 290 (2005) 166–174.
[5] M. Halmann, A. Steinfeld, Fuel saving, carbon dioxide emission avoidance, and
formation of amorphous carbon instead of filamentous carbon as con-
syngas production by tri-reforming of flue gases from coal- and gas-fired power
firmed by their TEM images (Fig. S3). It has been reported that amor- stations, and by the carbothermic reduction of iron oxide, Energy 31 (2006)
phous carbon is less stable and more reactive and thus does not affect 3171–3185.
[6] C. Song, W. Pan, Tri-reforming of methane: a novel concept for catalytic production
the catalyst's performance adversely [53]. The Ni particles are quite
of industrially useful synthesis gas with desired H2/CO ratios, Catal. Today 98
visible in the SEM image of spent Ni/Al2O3 catalyst (Fig. S4) indicating (2004) 463–484.
Ni particles would have remained well exposed and available for re- [7] R.K. Singha, A. Shukla, A. Yadav, S. Adak, Z. Iqbal, N. Siddiqui, R. Bal, Energy
actants. Filamentous carbon usually covers Ni particles and hence efficient methane tri-reforming for synthesis gas production over highly coke re-
sistant nanocrystalline Ni-ZrO2 catalyst, Appl. Energy 178 (2016) 110–125.
makes them less exposed and available for reaction [53]. Bitter et al. [8] M. Halmann, A. Steinfeld, Thermoneutral tri-reforming of flue gases from coal- and
also postulated that carbon formed on the metal causes catalyst deac- gas-fired power stations, Catal. Today 115 (2006) 170–178.
tivation rather than the carbon formed on the support [54]. SMSI, as [9] Y. Wang, L. Yao, Y. Wang, S. Wang, Q. Zhao, D. Mao, C. Hu, Low-temperature
catalytic CO2 dry reforming of methane on Ni-Si/ZrO2 catalyst, ACS Catal. 8 (2018)
evident from its TPR profile, did not let sintering of Ni particles to a 6495–6506.
greater extent. Sintered Ni/Al2O3 catalyst still had Ni particles small [10] J. Yoo, Y. Bang, S.J. Han, S. Park, J.H. Song, I.K. Song, Hydrogen production by tri-
enough to catalyze methane tri-reforming reaction effectively. reforming of methane over nickel-alumina aerogel catalyst, J. Mol. Catal. A Chem.
410 (2015) 74–80.
[11] L. Pino, A. Vita, F. Cipitì, M. Laganà, V. Recupero, Hydrogen production by methane
tri-reforming process over Ni-ceria catalysts: effect of La-doping, Appl. Catal. B
4. Conclusions Environ. 104 (2011) 64–73.
[12] A.J. Majewski, J. Wood, Tri-reforming of methane over Ni@SiO2 catalyst, Int. J.
Hydrog. Energy 39 (2014) 12578–12585.
The activity of Ni catalysts with different support materials Al2O3, [13] J.M. García-Vargas, J.L. Valverde, J. Díez, P. Sánchez, F. Dorado, Influence of al-
ZrO2, TiO2, SBA-15, MgO and CeO2-ZrO2 was assessed for TRM and kaline and alkaline-earth cocations on the performance of Ni/β-SiC catalysts in the
methane tri-reforming reaction, Appl. Catal. B Environ. 148–149 (2014) 322–329.
their stability towards carbon accumulation, metal sintering, and metal
[14] H. Jiang, H. Li, H. Xu, Y. Zhang, Preparation of Ni/MgxTi1−xO catalysts and in-
re-oxidation during the reaction was analyzed. The smaller Ni particles, vestigation on their stability in tri-reforming of methane, Fuel Process. Technol. 88
higher Ni dispersion, stronger metal-support interaction, higher degree (2007) 988–995.
of reducibility and higher concentration of surface basic sites led to [15] D.M. Walker, S.L. Pettit, J.T. Wolan, J.N. Kuhn, Synthesis gas production to desired
hydrogen to carbon monoxide ratios by tri-reforming of methane using
higher activity of Ni-based catalysts. Al2O3 supported Ni catalyst was Ni–MgO–(Ce,Zr)O2 catalysts, Appl. Catal. A Gen. 445–446 (2012) 61–68.
found to be the best with 8.72 × 10−2 mmol/gcat·s, 2.31 × 10−2 mmol/ [16] J.S. Kang, D.H. Kim, S.D. Lee, S.I. Hong, D.J. Moon, Nickel-based tri-reforming
gcat·s and 4.2 × 10−2 mmol/gcat·s CH4, CO2, and H2O conversion rate catalyst for the production of synthesis gas, Appl. Catal. A Gen. 332 (2007)
153–158.
respectively in the reaction conditions studied, while, Ni/SBA-15 and [17] G. Du, S. Lim, M. Pinault, C. Wang, F. Fang, L. Pfefferle, G.L. Haller, Synthesis,
Ni/ZrO2's performance was satisfactory. Ni/ZrO2 and Ni/CeO2-ZrO2 characterization, and catalytic performance of highly dispersed vanadium grafted
were the only catalysts that remained stable against Ni re-oxidation SBA-15 catalyst, J. Catal. 253 (2008) 74–90.
[18] T. Mondal, K.K. Pant, A.K. Dalai, Catalytic oxidative steam reforming of bio-ethanol
owing to oxophilic and redox property respectively. Ni/TiO2 and Ni/ for hydrogen production over Rh promoted Ni/CeO2-ZrO2 catalyst, Int. J. Hydrog.
MgO were found to be the least effective of all the tested catalysts. The Energy 40 (2015) 2529–2544.
low degree of reducibility of these catalysts was responsible for their [19] L. Chen, X. Sun, Y. Liu, Y. Li, Preparation and characterization of porous MgO and
NiO/MgO nanocomposites, Appl. Catal. A Gen. 265 (2004) 123–128.
poor catalytic performance. Weak metal-support interaction with fewer [20] A.P.F. Arena, A. Licciardello, The role of Ni2+ diffusion on the reducibility of NiO/
numbers of basic sites of Ni/CeO2-ZrO2 led to its average activity. The MgO system: a combined TRP-XPS study, Catal. Lett. 6 (1990) 139–150.
Ni confinement within pores of SBA-15 helped Ni/SBA-15 catalyst to [21] Z. Wang, X. Hu, D. Dong, G. Parkinson, C.Z. Li, Effects of calcination temperature of
electrospun fibrous Ni/Al2O3 catalysts on the dry reforming of methane, Fuel
attain remarkable stability against sintering and coking in the range of
Process. Technol. 155 (2017) 246–251.
experiments. Overall, the present experimental investigation identifies [22] G. Sankar, T. Raymentb, Promotion of the metal-oxide support interaction in the
the suitability of potential support materials to formulate Ni-based Ni/TiO2 catalyst, J. Mater. Chem. 1 (1991) 299–300.
catalyst for complex tri-reforming of methane process for CO2 utiliza- [23] K. Nagaoka, K. Takanabe, K. Aika, Influence of the phase composition of titania on
catalytic behavior of Co/TiO2 for the dry reforming of methane, Chem. Commun.
tion. It could also serve as the basis for selection of promoter for a given (2002) 1006–1007.
catalyst support and could be extended to other CH4/CO2 conversion [24] S. Wang, Y. Wang, C. Hu, The effect of NH3·H2O addition in Ni/SBA-15 catalyst
processes. preparation on its performance for carbon dioxide reforming of methane to produce
H2, Int. J. Hydrog. Energy 43 (2018) 13921–13930.

51
R. Kumar et al. Fuel Processing Technology 186 (2019) 40–52

[25] A. Chirieac, B. Dragoi, A. Ungureanu, C. Ciotonea, I. Mazilu, S. Royer, A.S. Mamede, [39] Y. Li, L. Wang, R. Yan, J. Han, S. Zhang, Promoting effects of MgO, (NH4)2SO4 or
E. Rombi, I. Ferino, E. Dumitriu, Facile synthesis of highly dispersed and thermally MoO3 modification in oxidative esterification of methacrolein over Au/
stable copper-based nanoparticles supported on SBA-15 occluded with P123 sur- Ce0.6Zr0.4O2-based catalysts, Catal. Sci. Technol. 6 (2016) 5453–5463.
factant for catalytic applications, J. Catal. 339 (2016) 270–283. [40] M.A. Ardagh, Z. Bo, S.L. Nauert, J.M. Notestein, Depositing SiO2 on Al2O3: a route
[26] C. Wang, N. Sun, M. Kang, X. Wen, N. Zhao, F. Xiao, W. Wei, T. Zhao, Y. Sun, The bi- to tunable Brønsted acid catalysts, ACS Catal. 6 (2016) 6156–6164.
functional mechanism of CH4 dry reforming over a Ni–CaO–ZrO2 catalyst: further [41] T. Mondal, K.K. Pant, A.K. Dalai, Oxidative and non-oxidative steam reforming of
evidence via the identification of the active sites and kinetic studies, Catal. Sci. crude bio-ethanol for hydrogen production over Rh promoted Ni/CeO2-ZrO2 cata-
Technol. 3 (2013) 2435–2443. lyst, Appl. Catal. A Gen. 499 (2015) 19–31.
[27] R.L.-F. Zouhair Boukha, Cristina Jiménez-González, Miryam Gil-Calvo, Beatriz de [42] B. Li, X. Lin, Y. Luo, X. Yuan, X. Wang, Design of active and stable bimodal nickel
Rivas, Juan Ramón González-Velasco, Jose Ignacio Gutiérrez-Ortiz, MgO/NiAl2O4 catalysts for methane reforming with CO2, Fuel Process. Technol. 176 (2018)
as a new formulation of reforming catalysts: tuning the surface properties for the 153–166.
enhanced partial oxidation of methane, Appl. Catal. B Environ. 199 (2016) [43] G. Słowik, M. Greluk, M. Rotko, A. Machocki, Evolution of the structure of un-
372–383. promoted and potassium-promoted ceria-supported nickel catalysts in the steam
[28] A. Ungureanu, B. Dragoi, A. Chirieac, C. Ciotonea, D. Duprez, A.S. Mamede, reforming of ethanol, Appl. Catal. B Environ. 221 (2018) 490–509.
E. Dumitriu, Composition-dependent morphostructural properties of Ni–Cu oxide [44] J.M. García-Vargas, J.L. Valverde, F. Dorado, P. Sánchez, Influence of the support
nanoparticles confined within the channels of ordered mesoporous SBA-15 silica, on the catalytic behaviour of Ni catalysts for the dry reforming reaction and the tri-
ACS Appl. Mater. Interfaces 5 (2013) 3010–3025. reforming process, J. Mol. Catal. A Chem. 395 (2014) 108–116.
[29] J. Zhu, X. Peng, L. Yao, D. Tong, C. Hu, CO2 reforming of methane over Mg-pro- [45] D. Pakhare, J. Spivey, A review of dry (CO2) reforming of methane over noble metal
moted Ni/SiO2 catalysts: the influence of Mg precursors and impregnation se- catalysts, Chem. Soc. Rev. 43 (2014) 7813–7837.
quences, Catal. Sci. Technol. 2 (2012) 529–537. [46] M.C. Silaghi, A. Comas-Vives, C. Copéret, CO2 activation on Ni/γ-Al2O3 catalysts by
[30] Y. Wang, L. Yao, S. Wang, D. Mao, C. Hu, Low-temperature catalytic CO2 dry re- first-principles calculations: from ideal surfaces to supported nanoparticles, ACS
forming of methane on Ni-based catalysts: a review, Fuel Process. Technol. 169 Catal. 6 (2016) 4501–4505.
(2018) 199–206. [47] I. Rossetti, A. Gallo, V. DalSanto, C.L. Bianchi, V. Nichele, M. Signoretto,
[31] K. Shimizu, K. Kon, W. Onodera, H. Yamazaki, J.N. Kondo, Heterogeneous Ni cat- E. Finocchio, G. Ramis, A. Di Michele, Nickel catalysts supported over TiO2, SiO2
alyst for direct synthesis of primary amines from alcohols and ammonia, ACS Catal. and ZrO2 for the steam reforming of glycerol, ChemCatChem 5 (2013) 294–306.
3 (2013) 112–117. [48] L. Kundakovic, D.R. Mullins, S.H. Overbury, Adsorption and reaction of H2O and
[32] A.H. Khoja, M. Tahir, N.A.S. Amin, Cold plasma dielectric barrier discharge reactor CO on oxidized and reduced Rh/CeOx (111) surfaces, Surf. Sci. 457 (2000) 51–62.
for dry reforming of methane over Ni/ɤ-Al2O3-MgO nanocomposite, Fuel Process. [49] R.K. Singha, S. Das, M. Pandey, S. Kumar, R. Bal, A. Bordoloi, Ni nanocluster on
Technol. 178 (2018) 166–179. modified CeO2–ZrO2 nanoporous composite for tri-reforming of methane, Catal. Sci.
[33] R.K. Singha, A. Yadav, A. Agrawal, A. Shukla, S. Adak, T. Sasaki, R. Bal, Synthesis of Technol. 6 (2016) 7122–7136.
highly coke resistant Ni nanoparticles supported MgO/ZnO catalyst for reforming of [50] S.N. Bukhari, C.Y. Chin, H.D. Setiabudi, D.V.N. Vo, Tailoring the properties and
methane with carbon dioxide, Appl. Catal. B Environ. 191 (2016) 165–178. catalytic activities of Ni/SBA-15 via different TEOS/P123 mass ratios for CO2 re-
[34] J. Spivey, K.M. Dooley, Y.-F. Han, Catalysis, The Royal Society of Chemistry, 2014. forming of CH4, J. Environ. Chem. Eng. 5 (2017) 3122–3128.
[35] J.I. Di Cosimo, V.K. Dı́ez, M. Xu, E. Iglesia, C.R. Apesteguı́a, Structure and surface [51] P.M. De Souza, R.C. Rabelo-Neto, L.E.P. Borges, G. Jacobs, B.H. Davis, T. Sooknoi,
and catalytic properties of Mg-Al basic oxides, J. Catal. 178 (1998) 499–510. D.E. Resasco, F.B. Noronha, Role of keto intermediates in the hydrodeoxygenation
[36] D. Wang, X. Zhang, J. Ma, H. Yu, J. Shen, W. Wei, La-modified mesoporous Mg–Al of phenol over Pd on oxophilic supports, ACS Catal. 5 (2015) 1318–1329.
mixed oxides: effective and stable base catalysts for the synthesis of dimethyl car- [52] I. Rossetti, C. Biffi, C.L. Bianchi, V. Nichele, M. Signoretto, F. Menegazzo,
bonate from methyl carbamate and methanol, Catal. Sci. Technol. 6 (2016) E. Finocchio, G. Ramis, A. Di Michele, Ni/SiO2 and Ni/ZrO2 catalysts for the steam
1530–1545. reforming of ethanol, Appl. Catal. B Environ. 117–118 (2012) 384–396.
[37] J.W. Han, J.S. Park, M.S. Choi, H. Lee, Uncoupling the size and support effects of Ni [53] Q. Ma, J. Sun, X. Gao, J. Zhang, T. Zhao, Y. Yoneyama, N. Tsubaki, Ordered me-
catalysts for dry reforming of methane, Appl. Catal. B Environ. 203 (2017) soporous alumina-supported bimetallic Pd–Ni catalysts for methane dry reforming
625–632. reaction, Catal. Sci. Technol. 6 (2016) 6542–6550.
[38] E. Farfan-Arribas, R.J. Madix, Characterization of the acid−base properties of the [54] J.H. Bitter, K. Seshan, J.A. Lercher, Deactivation and coke accumulation during
TiO2 (110) surface by adsorption of amines, J. Phys. Chem. B 107 (2003) CO2/CH4 reforming over Pt catalysts, J. Catal. 183 (1999) 336–343.
3225–3233.

52