Coupled reforming of methane to syngas (2H2-CO) over Mg-Al oxide supported Ni catalyst

Article abstract of DOI:10.1016/j.apcata.2017.11.004

We report bi-reforming and coupled reforming of methane with carbon dioxide, steam and/or oxygen to produce syngas over Ni supported on Mg-Al mixed oxide. The catalyst has been characterised in terms of specific surface area, TPR, XRD, TGA-DTG, SEM and TPO-MS analysis. Ni/Mg-Al mixed oxide exhibited Ni particle size range (11–30 nm) with a mean of 20.7 nm. Syngas H₂/CO = 2.0, suitable for methanol/Fischer-Tropsch fuel synthesis, has been achieved for both reactions (T = 1048 K, P = 1 atm). The impact of process parameters including temperature, feeding concentration and GHSV on conversion and H₂/CO ratio has been demonstrated. The Ni catalyst suffered temporal activity decline in bi-reforming that can be linked to formation of carbon whiskers encapsulated Ni particles resulting in a loss of active sites. Coupled reforming delivered higher CH₄ conversion and enhanced stability, but lower CO₂ conversion than bi-reforming under similar conditions. The enhanced stability in coupled reforming can be attributed to lower carbon deposition on Ni particles due to combustion of carbon by oxygen.

Full text of DOI:10.1016/j.apcata.2017.11.004

AppliedCatalysisA,General550(2018)176–183
Contents lists available at ScienceDirect
Applied Catalysis A, General
journal homepage: www.elsevier.com/locate/apcata
Research Paper
Coupled reforming of methane to syngas (2H2-CO) over Mg-Al oxide
supported Ni catalyst
Maoshuai Li, André C. van Veen⁎
School of Engineering, The University of Warwick, Coventry, CV4 7AL, United Kingdom
A R T I C L E I N F O
A B S T R A C T
Keywords:
Syngas
Methane
Carbon dioxide
Coupled reforming
Supported Ni
Carbon deposition
We report bi-reforming and coupled reforming of methane with carbon dioxide, steam and/or oxygen to produce
syngas over Ni supported on Mg-Al mixed oxide. The catalyst has been characterised in terms of specific surface
area, TPR, XRD, TGA-DTG, SEM and TPO-MS analysis. Ni/Mg-Al mixed oxide exhibited Ni particle size range
(11–30 nm) with a mean of 20.7 nm. Syngas H2/CO = 2.0, suitable for methanol/Fischer-Tropsch fuel synthesis,
has been achieved for both reactions (T = 1048 K, P = 1 atm). The impact of process parameters including
temperature, feeding concentration and GHSV on conversion and H2/CO ratio has been demonstrated. The Ni
catalyst suffered temporal activity decline in bi-reforming that can be linked to formation of carbon whiskers
encapsulated Ni particles resulting in a loss of active sites. Coupled reforming delivered higher CH4 conversion
and enhanced stability, but lower CO2 conversion than bi-reforming under similar conditions. The enhanced
stability in coupled reforming can be attributed to lower carbon deposition on Ni particles due to combustion of
carbon by oxygen.
−1
1. Introduction
based pyrochlore catalyst (1023 K, 9.8 × 104 cm3 h
g
−1) [9] and
cat
H2/CO = 2.0 over (Al2O3 [10], MgO-Al2O3 [11] and CeO2 [12]) sup-
−1
Conversion of methane and carbon dioxide, two major greenhouse
gases, to valuable chemicals has great significance in terms of global
energy security and climate change [1,2]. Reforming of these gases to
syngas, a critical intermediate in the manufacture of hydrogen, am-
monia, methanol and Fischer-Tropsch products represents a promising
route [3]. A variety of reforming reactions including
ported Ni (1023–1073 K, 5.3 × 105 cm3 h
g
cat
−1). In addition to Ni
based catalyst, bimetallic Ru-Ni (supported on MgO) [13] and Ru/
ZnLaAlO4 [14] have been examined in bi-reforming reaction to enhance
catalyst resistance to carbon deposition.
4CH4 + CO2 + H2O + O2 → 9H2 + 5CO
(4)
Coupled reforming (Eq. (4)), combined exothermic methane oxi-
dation with bi-reforming, represents an alternative route for syngas
production. Relative to bi-reforming, coupled reforming is more energy
efficient due to heat release from methane oxidation [1,15]. Use of
oxygen provides higher level of oxidant that can address carbon de-
position, a critical cause of catalyst deactivation in methane reforming
reaction [16]. Coupled reforming of methane to syngas, target at H2/
CO = 2 has not been studied to any significant extent. A search through
literature found reported studies on coupled reforming of methane over
Ni/MgO [17] and tri-reforming over (MgO [18], SiO2 [19] and CeO2-
ZrO2 [20]) supported Ni catalysts, producing syngas H2/CO = 1.5-3.0
(973–1123 K).
CH4 + CO2 → 2H2 + 2CO
(1)
(2)
(3)
3CH4 + CO2 + O2 → 6H2 + 4CO
3CH4 + CO2 + 2H2O → 8H2 + 4CO
dry reforming (Eq. (1)), oxy-CO2 reforming (Eq. (2)) and bi-re-
forming (Eq. (3)) have been studied in the existing literature [4–6]. Bi-
reforming to produce syngas offer significant advantages over dry re-
forming and oxy-CO2 reforming with respect to (I) flexibility in H2/CO
ratio adjustment for downstream chemical (e.g., methanol, Fischer-
Tropsch fuel) synthesis and (II) decreased carbon deposition due to
presence of steam [1]. Olah et al. have demonstrated achievement of
syngas (H2/CO = 2.0), suitable for methanol synthesis, and stable ac-
For methane reforming reactions, catalysts based on noble metals
exhibit high catalytic stability and great resistance to coke formation
[21]. However, high cost restricts industrial scale application. In-
expensive Ni based catalysts have been used in reforming reaction due
tivity with 320 h on-stream for high pressure (7–42 atm) bi-reforming
−1
reaction over Ni/MgO (1103 K, 6.0 × 104 cm3 h
g
cat
−1) [7,8]. Re-
action at atmospheric pressure delivered syngas H2/CO = 2.1 over Ni-
⁎
Corresponding author.
E-mail address: Andre.vanVeen@warwick.ac.uk (A.C. van Veen).
https://doi.org/10.1016/j.apcata.2017.11.004
Received 21 June 2017; Received in revised form 13 October 2017; Accepted 4 November 2017
Availableonline06November2017
0926-860X/©2017PublishedbyElsevierB.V.

M. Li, A.C. van Veen
AppliedCatalysisA,General550(2018)176–183
Fig. 1. Schematic diagram of experimental setup.
to high activity comparable to noble metals, but suffered rapid deac-
tivation because of Ni particle agglomeration and carbon deposition
[22]. MgO as basic support can enhance CO2 chemisorption [23]. In-
clusion of Al2O3 can increase surface area. Moreover, Ni supported on
Mg-Al mixed oxide have been shown to bear strong metal-support in-
teraction to suppress Ni sintering and exhibits enhanced catalytic sta-
bility in methane reforming reaction [24–26]. In this study, we for the
first time examine coupled reforming of methane with steam, carbon
dioxide and oxygen to produce syngas H2/CO = 2 at atmospheric
pressure over Mg-Al mixed oxide supported Ni catalyst and provide a
comparison of catalytic performance to bi-reforming. We also evaluate
carbon deposition in both reactions as one critical consideration for
catalyst stability.
plasma optical emission spectroscopy (ICP-OES, PerkinElmer 5300DV)
from the diluted extract in HNO3. Temperature programmed reduction
(TPR) was conducted in a quartz tube cell. The sample was heated in
84 cm3 min−1 5% v/v H2/Ar at 6 K min−1 to 1073 K and held for 1 h.
Hydrogen consumption was monitored by a thermal conductivity de-
tector (TCD). X-ray diffractograms (XRD) were recorded on a Bruker
D5005 X-ray diffractometer using Cu Kα radiation. Samples were
scanned at 0.02° step−1 over the range 30° ≤ 2θ ≤ 80° at ambient
temperature and the diffractograms identified against the JCPDS-ICDD
reference standards, i.e. Ni (04-0850), MgO (89-7746), γ-Al2O3 (10-
0425) and MgAl2O4 (77-1193). Nickel particle morphology (size and
shape) and carbon deposition was examined by Zeiss Supra 55VP field
emission scanning electron microscopy (SEM). Mean metal size (d) was
based on a count of up to 500 particles. Thermogravimetric-derivative
thermogravimetric analysis (TGA-DTG) of the samples post-reaction
was performed on a simultaneous thermal analyser (NETZSCH STA449)
by monitoring temporal mass with temperature. The samples (ca.
30 mg) were heated in 50 cm3 min−1 air to 1073 K (at 10 K min−1).
TPO-MS analysis of the spent catalysts was conducted in a quartz tube
by recording CO2 signal with time and temperature on a Pfeiffer
OMNIStar mass spectrometer. The samples (ca. 15 mg) were heated in
20 cm3 min−1 10% O2/Ar to 1073 K (at 5 K min−1).
2. Experimental
2.1. Materials and catalyst preparation
Mg-Al mixed oxide was prepared by co-precipitation of metal ni-
trates (Sigma Aldrich, > 98%) with aqueous ammonia (10% w/w,
Fisher) and ammonium carbonate (VWR Chemicals, 31.4% Assay NH3)
using flow synthesis. Aqueous nitrate salts (Mg2+ = 1.5 M, Mg/
Al = 3/1, 100 cm3) and a mixture (100 cm3) of ammonia (5 M) with
ammonium carbonate (0.25 M) was delivered separately via teflon line
2.3. Catalytic procedures
using
a
peristaltic pump (ISMATEC) at
a
fixed flow rate
(1.5 cm3 min−1), mixed in a tee (bore = 1.8 mm) and transported
continuously to 50 cm3 water. The suspension was stirred (500 rpm) at
353 K for 2 h. The solid obtained was separated by filtration, washed
with distilled water and dried at 393 K overnight. The dried sample was
calcined in air at 873 K (10 K min−1) for 4 h. Nickel (15% w/w) on Mg-
Al mixed oxide was prepared by deposition-precipitation using aqueous
ammonia. An aqueous solution of nickel nitrate (0.25 M, 100 cm3) and
ammonia (0.6 M, 100 cm3) was added to the support (8 g). The sus-
pension was stirred and heated to 353 K. The solid obtained was se-
parated by filtration, washed with distilled water and dried at 393 K
overnight. The catalyst precursor was sieved (ATM fine test sieves) to
60–200 mesh and activated at 10 K min−1 to 1073 K in
10 cm3 min−1 H2 (BOC, 99.99%) for characterisation.
Catalyst testing was conducted at atmospheric pressure
(973–1073 K), in situ after activation, in a continuous flow fixed bed
(alumina) tubular reactor (i.d. = 8 mm). The schematic diagram of the
reactor and gas analysis system is shown in Fig. 1. The catalyst
(5–30 mg) was mixed with ground quartz (60–200 mesh) and sand-
wiched between quartz wool. A layer of quartz particles was placed on
the top of quartz wool before the catalyst bed. Reaction temperature
was monitored by a thermocouple inserted in the catalyst bed. Water
was delivered to the reactor using a Shimadzu HPLC (LC-20AD) pump
and vaporized to steam in the upper part of the reactor. Reactant gases
(CH4, CO2 and/or O2, BOC, 99.99%), N2 (BOC, 99.99%) as internal
standard and Ar (BOC, 99.99%) as balance gas were introduced to re-
actor by Brooks mass flow controller (SLA5800 series) at (reactant) gas
hourly space velocity (GHSV) = 2 × 104–2 × 105 h−1. For all reac-
tions, the flow rate of methane was fixed at 9 cm3 min−1. The reactor
effluent was condensed in a gas sample cooler (Bühler) for subsequent
analysis using online gas chromatography (Shimadzu 2014) equipped
with a 0.5 cm3 sampling loop, thermal conductive detector (TCD) and
2.2. Catalyst characterisation
Nitrogen physisorption was performed on the Micromeritics ASAP
2020 system and total specific surface area (SSA) was calculated using
the standard BET method with pore volume obtained from BJH deso-
rption. Prior to analysis, samples were vacuumed and outgassed at
573 K for 1 h. Nickel content was measured by inductive coupled
flame ionization detector (FID), employing serial Hayesep
Q
(3.0 m × 2.1 mm i.d.) and Molecular Sieve 5A packed columns
(2.0 m × 2.1 mm i.d.). Data acquisition and manipulation were
177

M. Li, A.C. van Veen
AppliedCatalysisA,General550(2018)176–183
Table 1
Ni loading, specific surface area (SSA), total pore volume, TPR H2 consumption and Ni particle size (from SEM analysis) of the fresh and spent catalysts.
SSA (m2 g−1
)
Pore volume (cm3 g−1
)
H2 consumption (mol molNi
)
Size range (nm)
d (nm)
−1
Ni loading (% w/w)
Fresh
15.4
–
–
127
136
130
0.29
0.27
0.25
0.93
–
–
11–30
11–90
11–90
20.7
33.9
26.5
Spent Ia
Spent IIb
a
Spent catalyst post bi-reforming.
Spent catalyst post coupled reforming.
b
performed using GCsolution Lite (Version 2.4) chromatography data
system. Reactant i (i = CH4 or CO2) conversion (Xi) is calculated by the
change of volumetric flow rate
required for the Ni2+ → Ni0 step. Düdder et al. reported a single TPR
peak (Tmax = 943–1188 K) during activation (6 K min−1
of Ni/
)
MgAlOx with a shift to higher temperature with lowering Ni content
(50 → 1 mol%) [29]. Oemar et al. [30] have ascribed TPR signal in the
temperature range (673–773 K) to reduction of NiO species bearing
weak interaction with oxide support and high temperature (> 873 K)
peak to NiO species that have strong metal-support interaction in the
TPR analysis (10 K min−1) of Ni/SBA-15. XRD analysis (Fig. 2(II)) re-
vealed three diffraction signals at 2θ = 44.5°, 51.9° and 76.6° corre-
sponding to Ni (111), (200) and (220) planes. Additional peaks at
2θ = 36.9°, 43.0°, 62.5° and 78.8° can be attributed to cubic MgO [31].
There was no clearly discernible peak for magnesium aluminate spinel
(2θ = 31.3°, 36.9°, 44.8°, 59.4 and 65.2°) and γ-Al2O3 (2θ = 39.5°,
45.9° and 67.0°), which may be due to masking by stronger signals of
MgO and Ni and/or below detection limit. Nickel particle size is critical
in determining carbon formation/accumulation and catalyst stability in
methane reforming [32]. Representative SEM image and associated Ni
particle size distribution histogram (Fig. 3) revealed that the sample
exhibited pseudo-spherical Ni particles with a size range (11–30 nm)
and mean (20.7 nm, Table 1). This value is smaller than that (54.2 nm)
reported by Hadian et al. for (15% w/w) Ni impregnated on MgAl2O4
[24].
Q[reactant]i,in − Q[reactant]i,out
Xi (%) =
× 100
Q[reactant]i,in
(5)
where subscripts “in” and “out” refer to inlet and outlet gas streams. In
blank tests, passage of reactant gases to empty reactor did not result in
any detectable conversion. Repeated reactions delivered data re-
producibility and carbon balance within 7%.
3. Results and discussion
3.1. Catalyst characterisation
Sample physicochemical properties are presented in Table 1. The
specific surface area (SSA) and pore volume of activated Ni/Mg-Al
mixed oxide (127 m2 g−1, 0.29 cm3 g−1) is lower than the support
(172 m2 g−1, 0.51 cm3 g−1) and can be attributed to a partial pore
filling during catalyst preparation [27]. The values are in good agree-
ment with those reported for Ni supported on Mg-Al mixed oxide in the
literature [26,28]. TPR profile of Ni/Mg-Al mixed oxide (Fig. 2(I)) ex-
hibited a reduction peak at the extended time, where hydrogen con-
−1
3.2. Catalysis of bi-reforming
sumption (Table 1) is in good agreement with that (1.0 mol molNi
)
Thermodynamic analysis of bi-reforming demonstrates that CH4/
(H2O + CO2) = 0.9-1.2 and high temperature (≥973 K) facilitates
conversion of CH4 and CO2 [33]. The reaction was initially carried out
at 1048 K using a feeding ratio of CH4/H2O/CO2 = 3/2.2/1.2. Re-
presentative time on-stream conversion and H2/CO ratio profiles for bi-
reforming are shown in Fig. 4. An initial decrease in activity was ob-
served with steady conversion (CH4 = 73%, CO2 = 64%) attained after
8 h on-stream. Although activities cannot be directly compared due to
different conditions, our conversions are higher than that
(CH4 = 52–60%, CO2 = 48–52%) reported elsewhere [9], but lower
than the corresponding equilibrium conversions (CH4 = 94%,
CO2 = 77%) [33]. Methane and carbon dioxide were solely converted
to syngas with no detectable by-products (e.g., alkane and/or alkene),
indicative of 100% selectivity to CO (based on CH4). H2/CO ratio
fluctuated within the 1.99–2.03 range with a mean value of 2.01
achieved under the reaction condition employed. This meets the re-
quirement for methanol/Fischer-Tropsch fuel synthesis and circum-
vents extra adjustment of H2/CO ratio when compared to methane dry
reforming (H2/CO = 1) and steam reforming (H2/CO = 3).
To understand bi-reforming, the impact of process parameters in-
cluding temperature, CH4/H2O feeding ratio and GHSV on catalytic
activity and H2/CO ratio were examined and presented in Fig. 5. An
increase in temperature from 973 K to 1073 K (AI) resulted in higher
conversion of CH4 (52% → 78%) and CO2 (39% → 68%), consistent
with the nature of endothermic reaction. The conversions were lower
than the equilibrium values. H2/CO ratio decreased (2.22 → 1.95) with
increasing temperature (AII). Higher temperature favors methane dry
reforming that generates lower amounts of H2 relative to steam re-
forming. This can account for decreased H2/CO ratio observed at higher
temperature. Water serves as hydrogen donors in steam reforming,
Fig. 2. (I) TPR profile and (II) XRD pattern for Ni/Mg-Al mixed oxide (Δ MgO, ♦ Ni, □
MgAl2O4 spinel).
178

M. Li, A.C. van Veen
AppliedCatalysisA,General550(2018)176–183
Fig. 3. Representative SEM image and Ni size distribution
histogram for Ni/Mg-Al mixed oxide.
which can impact on hydrogen production [34]. We considered eva-
luation of varying water concentration (presented as molar CH4/H2O
ratio) on activity and H2/CO adjustment (Fig. 5(B)). A three-fold in-
crease in CH4 conversion with a consequently decreased CO2 conver-
sion (to 50%) was observed with increasing water feeding (CH4/H2O:
3/0 → 3/3, BI). This suggests a competition of dry reforming with
steam reforming, where steam reforming predominates in the presence
of excess water. Increasing water concentration served to enhance H2/
CO ratio (1.00 → 2.25, BII), confirming the role of water in adjusting
H2/CO. We can note that H2/CO = 1 obtained at CH4/H2O = 3/0
equals to the reaction stoichiometry of methane dry reforming (Eq. (1)).
This demonstrates an increase in water feeding switched the reaction
from exclusive dry reforming to predominant steam reforming. To this
point, bi-reforming of methane to produce syngas H2/CO = 2 has been
established. However, conversion was still low. The possibility of de-
creasing GHSV by increasing catalyst amount was considered as a
means of enhancing conversions of methane and carbon dioxide. The
Fig. 4. Conversions of methane (■), carbon dioxide (●) and H2/CO ratio (♦) with time
on-stream in the bi-reforming of methane (T = 1048 K, CH4/H2O/CO2 = 3/2.2/1.2,
GHSV = 8.6 × 104 h−1, dash line: equilibrium value).
Fig. 5. Variation of (I) conversions (square: CH4, circle: CO2) and (II) H2/
CO ratio
with (A)
temperature
(CH4/H2O/CO2 = 3/2.2/1.2,
GHSV = 8.6 × 104 h−1); (B) CH4/H2O feeding ratio (T = 1048 K, CH4/
CO2 = 3/1.2) and (C) GHSV (T = 1048 K, CH4/H2O/CO2 = 3/2.2/1.2) in
the bi-reforming of methane (solid: experimental value, open: equilibrium
value).
179

M. Li, A.C. van Veen
AppliedCatalysisA,General550(2018)176–183
reforming, coupled reforming delivered higher CH4 conversion, but
lower CO2 conversion under similar conditions. This can be to some
extent attributed to complete combustion of methane to carbon dioxide.
Syngas was the sole product. H2/CO ratio (mean = 1.99), close to 2.0
was maintained for 24 h on-stream with no obvious fluctuation. This
value was lower than that (mean = 2.01) obtained in bi-reforming. Our
results have established coupled reforming of methane to syngas with
H2/CO = 2.
In common to bi-reforming, higher conversion of CH4 (up to 94%)
and CO2 (up to 74%) were observed at elevated temperature
(Fig. 8(AI)). The conversions at 1073 K were close to the equilibrium
values. Oxygen was completely converted at any temperature, con-
sistent with the equilibrium conversions. It should be noted that CO2
conversion was down to 3% at 973 K, suggesting low temperature does
not facilitate CO2 activation and conversion. H2/CO ratio decreased
dramatically (2.27 → 1.98) from 973 K to 1073 K (Fig. 8(AII)), con-
sistent with that observed in bi-reforming reaction (Fig. 5(A)). In
common to water effect on bi-reforming (Fig. 5(B)), an increase in
water feeding (CH4/H2O: 4/0 → 4/3) served to enhance CH4 conver-
sion (70% → 93%) with steady conversion attained at higher feeding
(CH4/H2O: 4/3-4/4). This suggests use of excess water (three times
higher than the reaction stoichiometry) does not further promote me-
thane steam reforming. A consequential decrease (95% → −12%) in
CO2 conversion was observed with increasing water feeding
(Fig. 8(BI)). Negative CO2 conversion at CH4/H2O = 4/4, also reported
in Choudhary’s study [17], can be attributed to that dry reforming was
suppressed with lower CO2 conversion and/or methane steam re-
forming with CO2 generation was promoted, resulting in net formation
of CO2. Full conversion of oxygen was obtained at any CH4/H2O. In
common to bi-reforming, lower water feeding generated less amount of
hydrogen, resulting in decreased H2/CO ratio (Fig. 8(BII)). For reaction
conducted in absence of steam (CH4/H2O = 4/0), H2/CO equaled to
the reaction stoichiometry (1.5) of oxy-CO2 reforming (Eq. (2)). Oxygen
can participate in CH4 oxidation, CO oxidation and/or carbon com-
bustion in the coupled reforming [4], which impact on activity and H2/
CO ratio. The influence of oxygen feeding content (presented as molar
CH4/O2 ratio) was examined and presented in Fig. 8(C). A decrease in
oxygen feeding (CH4/O2: 4/2 → 4/0) lowered CH4 conversion (95% →
59%) and enhanced CO2 conversion (-17% → 81%). Negative CO2
conversion observed at CH4/O2 = 4/2 can be attributed to excess
oxygen promoted methane combustion resulting in net formation of
CO2. H2/CO ratio was decreased (2.07 → 1.90) with decreasing O2
feeding, consistent with Choudhary’s observation [17]. Choudhary
et al. [17] concluded that H2/CO ratio was determined by steam re-
forming, partial oxidation and dry reforming, occurring to different
extents depending upon the process conditions. We can attribute de-
creased H2/CO to dry reforming predominated over steam reforming
with less H2 generation. A decrease in GHSV from 8.7 × 104 h−1 to
4.4 × 104 h−1 resulted in enhanced conversions of CH4 and CO2 (up to
96% for CH4 and 63% for CO2, Fig. 8(DI)). CH4 and CO2 conversions
Fig. 6. Analysis of net heat change (ΔHr) as a function of temperature and CH4/O2 (4/2
(▲), 4/1 (■) and 4/0 (●); CH4/H2O/CO2 = 4/1.6/1).
profiles shown in Fig. 5(C) revealed that both conversions increased (up
to 87% for CH4 and 81% for CO2) with lowering GHSV (13.6 × 104 →
2.8 × 104 h−1). The values at GHSV = 2.8 × 104 h−1 were close to the
equilibrium conversions (CH4 = 94%, CO2 = 77%) [33]. H2/CO did
not vary significantly, maintaining within 1.99-2.01. This suggests
variation in GHSV did not change the equilibrium between steam and
dry reforming, and lower GHSV promoted both reactions.
3.3. Catalysis of coupled reforming
To probe autothermal reaction condition for the coupled reforming,
net heat change (ΔHr) was analysed as a function of temperature and
CH4/O2; the result is shown in Fig. 6. Regardless of CH4/O2, the net
heat of the reaction increased with increasing temperature (973 →
1073 K), indicating an increase in the reaction endothermicity and/or a
decrease in the exothermicity. This can be attributed to higher tem-
perature favors endothermic (dry and steam) reforming of methane.
With an increase in oxygen feeding (CH4/O2: 4/0 → 4/2), the value of
net heat went from positive to negative, suggesting a move from en-
dothermal to exothermal zone. This can be linked to increased oxygen
feeding promoting exothermic methane oxidation. Net heat balance
(ΔHr = 0) was observed at CH4/O2 = 4/1 and 1023 K.
Representative time on-stream conversion and H2/CO ratio profiles
for the coupled reforming are shown in Fig. 7. In contrast to bi-re-
forming, stable conversions, e.g., CH4 = 81%, CO2 = 52% and
O2 = 100%, were maintained within 24 h on-stream, indicative of no
apparent catalyst deactivation. There is consensus that presence of
oxygen serves to limit carbon deposition with enhanced catalytic sta-
bility for reforming reactions [35]. Coupled reforming of methane has
not been examined in detail with few studies available for direct
comparison. But we can note 96% and 40% conversion for CH4 and
CO2, reported by Choudhary et al. for reaction at higher temperature
(1123 K) and lower GHSV (4.8 × 104 h−1) [17]. Compared to bi-
tended to be constant with
a further decrease in GHSV (to
2.9 × 104 h−1) due to thermodynamic equilibrium control. In contrast
to bi-reforming, H2/CO decreased (1.99 → 1.86) with decreasing GHSV
(Fig. 8(DII)). Olah et al. [8] have observed decreased H2/CO (< 2) at
lower GHSV in high pressure bi-reforming reaction, but without spe-
cifying a reason. We can tentatively attribute the decreased H2/CO to
predominance of methane combustion and dry reforming in the reac-
tion network.
3.4. Catalyst characterisation post-reaction
Catalyst deactivation due to carbon deposition and Ni sintering was
a feature of methane reforming reaction [16,36]. Catalysts post bi-re-
forming (Fig. 4) and coupled reforming (Fig. 7) were subjected to
characterisation measurements to study catalyst deactivation. A slight
increase in SSA and decrease in pore volume was recorded over the
Fig. 7. Conversions of methane (■), carbon dioxide (●), oxygen (▲) and H2/CO ratio
(♦) with time on-stream in the coupled reforming of methane (T = 1048 K, CH4/H2O/
CO2/O2 = 4/1.6/1/1, GHSV = 8.7 × 104 −1, dash line: equilibrium value).
h
180

M. Li, A.C. van Veen
AppliedCatalysisA,General550(2018)176–183
Fig. 8. Variation of (I) conversions (square: CH4, circle: CO2, triangle: O2)
and (II) H2/CO ratio with (A) temperature (CH4/H2O/CO2/O2 = 4/1.6/1/
1, GHSV = 8.7 × 104 h−1); (B) CH4/H2O feeding ratio (T = 1048 K,
CH4/CO2/O2 = 4/1/1); (C) CH4/O2 feeding ratio (T = 1048 K, CH4/H2O/
CO2 = 4/1.6/1) and (D) GHSV (T = 1048 K, CH4/H2O/CO2/O2 = 4/1.6/
1/1) in the coupled reforming of methane (solid: experimental value,
open: equilibrium value).
spent samples exhibited a similar (2-2.3% w/w) mass loss at T ≤ 393 K
due to water removal. Mass loss at higher temperature (ca. 473 K) can
be attributed to desorption of adsorbed reactants (e.g., H2O, CO2 and
CH4) and removal of easily oxidisable carbonaceous species as reported
elsewhere [37]. Amorphous carbon and/or graphic type carbon con-
tribute to the mass loss at 660 K [38]. Additional weight loss at 783 K
for the sample post bi-reforming and at 873 K for the spent catalyst post
coupled reforming can be attributed to combustion of different types of
whisker carbon [39]. Carbon deposition on the catalyst surface was
further confirmed by SEM analysis (Fig. 10(A)). Significant amount of
carbon whiskers were detected on the catalyst surface post bi-reforming
(IA). A number of Ni particles were encapsulated on the end of carbon
whiskers.
In
methane
reforming,
methane
dissociation
Fig. 9. TGA and DTG analysis for catalysts post bi-reforming (solid line) and coupled
(CH4 → C + 2H2) and Boudouard reaction (2CO → CO2 + C) are main
carbon formation reactions, where the former is more significant at
temperature > 873 K [40,41]. Djinović et al. [42] studying carbon
formation in dry reforming of methane over Ni/Al2O3, reported that
encapsulating carbon leading to loss of Ni active surface was re-
sponsible for catalyst deactivation. Moreover, Ni particle size distribu-
tion histogram for the catalyst post bi-reforming (Fig. 10 (IB)) revealed
a wider Ni size range (11–90 nm) and larger mean (33.9 nm) relative to
the fresh catalyst (11–30 nm, mean = 20.7 nm), suggesting severe Ni
particle agglomeration and sintering during reaction. The spent catalyst
reforming (dash line).
spent catalysts (Table 1), which can be a consequence of carbon accu-
mulation. TGA result (Fig. 9) revealed that the spent catalyst after the
coupled reforming exhibited a lower (by 3%) mass loss relative to bi-
reforming in the whole investigated temperature range (298–1073 K).
A calculation demonstrated the amount of carbon deposition equaled to
0.23-0.24% of total carbon feeding. Derivative plots for the spent
samples demonstrated four steps occurred during weight loss. Both
181

M. Li, A.C. van Veen
AppliedCatalysisA,General550(2018)176–183
Fig. 10. (A) Representative SEM images and (B) Ni size distribution his-
togram for catalysts post (I) bi-reforming and (II) coupled reforming.
4. Conclusion
We have established bi-reforming and coupled reforming of me-
thane with carbon dioxide, steam and/or oxygen to produce syngas
(H2/CO = 2.0) over Ni particles (mean = 20.7 nm) supported on Mg-
Al mixed oxide. Higher conversions were observed at elevated tem-
perature and lower GHSV for both reactions. Increased water feeding
served to generate greater hydrogen with enhanced H2/CO. Coupled
reforming delivered higher CH4 conversion and enhanced stability than
bi-reforming under similar conditions, but with lower CO2 conversion.
The temporal activity loss in bi-reforming can be linked to carbon de-
position (on the basis of TGA-DTG, TPO-MS and SEM analysis).
Acknowledgement
Fig. 11. TPO-MS profile for catalysts post bi-reforming (solid line) and coupled reforming
(dash line).
This work was financially supported by European Commission 7th
Framework program: BIOGO-for-Production project (Grant No.
604296). We are grateful to Mr Lukas Tillmann, Dr. Nicolay Cherkasov,
Prof. Shanwen Tao and Mr Steve J. York for the contribution to TPR,
TPO-MS, TGA-DTG and SEM analysis, respectively.
for coupled reforming (IIB) exhibited the same Ni size range
(11–90 nm), but a smaller mean (26.5 nm) than that for the catalyst
post bi-reforming, suggesting slower sintering. TPO-MS measurement
(Fig. 11) used to analyse carbon deposition revealed one low-tem-
perature exothermic peak due to the removal of amorphous carbon
and/or graphic carbon at 659 K for the spent catalysts post both reac-
tions. Additional high-temperature exothermic signal due to combus-
tion of carbon whisker was observed at 746 K for the catalyst post bi-
reforming and at 844 K for the catalyst post coupled reforming. TPO-MS
analysis again confirmed formation of amorphous/graphic carbon and
carbon whisker. In this study, carbon generated from methane dis-
sociation on Ni surface, subsequent growth to whiskers and en-
capsulation of Ni particles, and severe Ni sintering can be linked to the
initial decrease of activity in bi-reforming reaction. Less carbon for-
mation and slower Ni sintering contributes to enhanced stability in the
coupled reforming.
References
[1] N. Kumar, M. Shojaee, J.J. Spivey, Curr. Opin. Chem. Eng. 9 (2015) 8–15.
[2] J.M. Lavoie, Front. Chem. 2 (2014).
[3] N. Abatzoglou, C. Fauteux-Lefebvre, Wiley Interdiscip. Rev. Energy Environ. 5
(2016) 169–187.
[4] T.V. Choudhary, V.R. Choudhary, Angew. Chem. Int. Ed. 47 (2008) 1828–1847.
[5] M. Usman, W. Daud, H.F. Abbas, Renew. Sust. Energ. Rev. 45 (2015) 710–744.
[6] G.A. Olah, A. Goeppert, M. Czaun, G.K.S. Prakash, J. Am. Chem. Soc. 135 (2013)
648–650.
[7] G.A. Olah, G.K.S. Prakash, A. Goeppert, M. Czaun, T. Mathew, J. Am. Chem. Soc.
135 (2013) 10030–10031.
[8] G.A. Olah, A. Goeppert, M. Czaun, T. Mathew, R.B. May, G.K.S. Prakash, J. Am.
Chem. Soc. 137 (2015) 8720–8729.
[9] N. Kumar, A. Roy, Z. Wang, E.M. L'Abbate, D. Haynes, D. Shekhawat, J.J. Spivey,
Appl. Catal. A: Gen. 517 (2016) 211–216.
[10] K.Y. Koo, H.S. Roh, U.H. Jung, W.L. Yoon, Catal. Today 185 (2012) 126–130.
[11] K.Y. Koo, H.S. Roh, U.H. Jung, D.J. Seo, Y.S. Seo, W.L. Yoon, Catal. Today 146
(2009) 166–171.
[12] H.S. Roh, K.Y. Koo, W.L. Yoon, Catal. Today 146 (2009) 71–75.
[13] M.A. Alvarez, M.A. Centeno, J.A. Odriozola, Top. Catal. 59 (2016) 303–313.
182

M. Li, A.C. van Veen
AppliedCatalysisA,General550(2018)176–183
[14] Y. Khani, Z. Shariatinia, F. Bahadoran, Chem. Eng. J. 299 (2016) 353–366.
[15] V.R. Choudhary, K.C. Mondal, T.V. Choudhary, Appl. Catal. A: Gen. 306 (2006)
45–50.
[30] U. Oemar, Y. Kathiraser, L. Mo, X.K. Ho, S. Kawi, Catal. Sci. Technol. 6 (2016)
1173–1186.
[31] M. Li, Y. Hao, F. Cárdenas-Lizana, M.A. Keane, Catal. Commun. 69 (2015) 119–122.
[16] O. Muraza, A. Galadima, Int. J. Energy Res. 39 (2015) 1196–1216.
[17] V.R. Choudhary, B.S. Uphade, A.S. Mamman, Appl. Catal. A: Gen. 168 (1998)
33–46.
[32] K.O. Christensen, D. Chen, R. Lodeng, A. Holmen, Appl. Catal. A: Gen. 314 (2006)
9–22.
[33] W.J. Jang, D.W. Jeong, J.O. Shim, H.M. Kim, H.S. Roh, I.H. Son, S.J. Lee, Appl.
[18] C.S. Song, P. Wei, Catal. Today 98 (2004) 463–484.
[19] A.J. Majewski, J. Wood, Int. J. Hydrogen Energy 39 (2014) 12578–12585.
[20] R.K. Singha, S. Das, M. Pandey, S. Kumar, R. Bal, A. Bordoloi, Catal. Sci. Technol. 6
(2016) 7122–7136.
Energy 173 (2016) 80–91.
[34] T.L. LeValley, A.R. Richard, M.H. Fan, Int. J. Hydrogen Energy 39 (2014)
16983–17000.
[35] H. Harju, J. Lehtonen, L. Lefferts, Catal. Today 244 (2015) 47–57.
[21] D. Pakhare, J. Spivey, Chem. Soc. Rev. 43 (2014) 7813–7837.
[22] S. Kawi, Y. Kathiraser, J. Ni, U. Oemar, Z.W. Li, E.T. Saw, ChemSusChem 8 (2015)
3556–3575.
[36] H.J. Wu, V. LaParola, G. Pantaleo, F. Puleo, A.M. Venezia, L.F. Liotta, Catalysts 3
(2013) 563–583.
[37] J.J. Guo, H. Lou, H. Zhao, D.F. Chai, X.M. Zheng, Appl. Catal. A: Gen. 273 (2004)
[23] A. Serrano-Lotina, L. Daza, Appl. Catal. A: Gen. 474 (2014) 107–113.
[24] N. Hadian, M. Rezaei, Z. Mosayebi, F. Meshkani, J. Nat. Gas Chem. 21 (2012)
200–206.
75–82.
[38] A. Kambolis, H. Matralis, A. Trovarelli, C. Papadopoulou, Appl. Catal. A: Gen. 377
(2010) 16–26.
[25] C.J. Liu, J.Y. Ye, J.J. Jiang, Y.X. Pan, ChemCatChem 3 (2011) 529–541.
[26] Y.J. Zhu, S.H. Zhang, B.B. Chen, Z.S. Zhang, C. Shi, Catal. Today 264 (2016)
163–170.
[39] H. Arbag, S. Yasyerli, N. Yasyerli, G. Dogu, T. Dogu, Appl. Catal. B: Environ. 198
(2016) 254–265.
[40] M.A. Ermakova, D.Y. Ermakov, A.L. Chuvilin, G.G. Kuvshinov, J. Catal. 201 (2001)
[27] S. Jujjuri, E. Ding, S.G. Shore, M.A. Keane, J. Mol. Catal. A: Chem. 272 (2007)
96–107.
183–197.
[41] A.F. Cunha, J.J.M. Orfao, J.L. Figueiredo, Int. J. Hydrogen Energy 34 (2009)
[28] M. Li, X. Wang, S.R. Li, S.P. Wang, X.B. Ma, Int. J. Hydrogen Energy 35 (2010)
6699–6708.
[29] H. Düdder, K. Kähler, B. Krause, K. Mette, S. Kühl, M. Behrens, V. Scherer,
M. Muhler, Catal. Sci. Technol. 4 (2014) 3317–3328.
4763–4772.
[42] P. Djinović, I.G.O. Ĉrnivec, B. Erjavec, A. Pintar, Appl. Catal. B: Environ. 125 (2012)
259–270.
183