i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 0 2 6 3 e3 0 2 8 1

Available online at www.sciencedirect.com

ScienceDirect

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

Hydrogen production from sorption enhanced

steam reforming of ethanol using bifunctional Ni
and Ca-based catalysts doped with Mg and Al

Cassiano Cunha Oliveira, Carla Eponina Hori*

^ ndia, School of Chemical Engineering, Av. Joa
Federal University of Uberla 
~o Naves de Avila, 2121, Bloco 1 K, 38400-
^ndia, MG, Brazil
902, Uberla

highlights graphical abstract

Steam reforming of ethanol with

simultaneous CO2 capture was
studied.

Ni/Ca, Ni/CaMgAl-79 and Ni/
CaMgAl-68 bifunctional catalysts
were tested.

CaO properties were enhanced by
the addition of Mg and Al in cyclic
process.

All catalysts achieved a high-
purity H2 (~90 mol%) during reac-
tion cycles.

Doping the catalysts with Mg and
Al decreased the coke formation.

article info abstract

Article history: This work evaluated the performance of nickel-based catalysts supported on CaO and CaO
Received 8 December 2022 eMgOeAl2O3 in the sorption enhanced steam reforming of ethanol (SESRE) aiming the
Received in revised form production of high purity H2. The catalysts were prepared by sol-gel method and charac-
5 February 2023 terized by different methods: Temperature programmed reduction (TPR), X-ray diffraction
Accepted 14 April 2023 (XRD), Scanning Electron Microscopy (SEM) with chemical element mapping, N2 phys-
Available online 4 May 2023 isorption and CO2 capture capacity determined by thermogravimetric analysis (TGA). XRD
analysis showed that the predominant phases were CaO, MgO, CaCO3, Ca(OH)2 and NiO in
Keywords: the calcined samples and Ni0 in the reduced and passivated samples. TPR profiles indicated
Nickel-based catalysts that all catalysts presented a high degree of reduction (Ni/CaMgAl-68 > Ni/CaMgAl-79 > Ni/
Ethanol Ca), although Ni/CaMgAl-X samples presented high reduction temperatures indicating the
Sol-gel method formation of NiAl2O4. The addition of MgO and Al2O3 to CaO was very beneficial since the
CO2 capture deactivation coefficients, calculated by the TGA data modeling, decreased by a factor of 3.8
Dopant for Ni/CaMgAl-79 and by a factor of 4.3 for Ni/CaMgAl-68 when compared to the Ni/Ca
catalyst. The catalytic tests in the SESRE showed that Ni/CaMgAl-79 catalyst had the best
performance since it had the longest high purity hydrogen production time. In the pre-

* Corresponding author.
E-mail address: cehori@ufu.br (C.E. Hori).
https://doi.org/10.1016/j.ijhydene.2023.04.178
0360-3199/© 2023 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
30264 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 0 2 6 3 e3 0 2 8 1

breakthrough period, the H2 mole fractions were close to 90% for all samples during all
reaction cycles. After the reaction-regeneration cycles, the average crystallite size of CaO
estimated by XRD increased around 38, 6 and 35% for Ni/Ca, Ni/CaMgAl-79 and Ni/CaMgAl-
68, respectively. Thus, adding a dopant to the sorbent material proved to be an effective
strategy to obtain a more stable catalyst capable to produce hydrogen of high purity.
© 2023 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.

acetaldehyde selectivity. The other catalysts had 100% of

Introduction
The global population growth, urbanization and industrial
development are leading to a continuous increase of energy
demand. Nowadays, energy supply is heavily dependent on
fossil fuels, which are non-renewable. This type of energy is
gradually depleted and has been causing serious environ-
mental problems, especially global warming effects [1].
Hydrogen has been considered to be one of the most prom-
ising future energy vectors due to its high power density and
clean combustion [2].
Catalytic reforming is an effective and classic method to
produce hydrogen, which has been used for several years in
commercial chemical production and fuel upgrading. Using
biomass as a renewable resource is very advantageous
because it is highly available in many forms, reduces the
consumption of fossil fuels and it has a low dependence on
location and climate. The expansion of feedstock usage from
hydrocarbons to oxygenated compounds has brought new
challenges and opportunities for the catalytic reforming pro-
cess. Steam reforming of ethanol (SRE) is a sustainable option
because ethanol can be produced from renewable biomass by
fermentation processes. Ethanol is a promising hydrogen
storage chemical for large-scale transportation since its global
annual output exceeds 900 million tons [3e7].
The oil obtained by pyrolysis, known as bio-oil, is a com-
plex mixture of alcohols, acids, aldehydes, and other oxygen-
containing components. Bio-oil, obtained from the rapid py-
rolysis of biomass, is one of the most promising and
economically viable sources of hydrogen due to its higher
energy density than biomass, easy storage and transportation,
and flexible use. The literature has been highlighted due to the
complexity of bio-oils for model compounds including
ethanol, acetic acid, methanol, acetone, etc. Ethanol is widely
studied as a model compound for the aqueous phase of bio-
oils because it is a representative molecule. Considering that
bio-oil production is rich in raw materials and the production
process is carbon-neutral, a good use of bio-oil is hydrogen
production through catalytic steam reforming [8,9]. Di Michele
et al. [10] performed steam reforming of ethanol (SRE) at 625  C
using MgAl2O4 as a stable support and varying nickel
composition as an active phase (1.5, 5 and 10 wt%). For 1.5 wt%
Ni, ethanol conversion was 86% and ethylene was a byproduct
due the catalyst low activity to reform this intermediate. This
metal loading is insufficient for CeC bond cleavage, keeping
unreacted ethanol and ethylene. This sample also had low
stability due to the decrease of ethanol conversion after
450 min of reaction and the simultaneous increase of
ethanol conversion and increasing the Ni content improved
H2 production because of the lower CO/CO2 ratio promoted by
the water-gas shift reaction and diminished CH4 selectivity.
Bepari et al. [11] investigated a series of Ce-promoted
NieMgeAl hydrotalcite catalysts, prepared by co-
precipitation followed by wet impregnation, during SRE. The
cerium content and reaction conditions were varied (tem-
perature, steam/ethanol molar ratio and space-time) in order
to observe the products distribution and catalytic activity. The
catalyst with 10% Ce loading had a good stability and activity
during a long time of reaction at atmospheric pressure. The
best operational conditions were with a space-time of 22.04
kgcat h kmol1 of ethanol fed and a steam/ethanol molar ratio
of 9 at 540  C. At these conditions, ethanol conversion was
about 97% and hydrogen yield of 4.13 mol of hydrogen per
mole of ethanol reacted.
Although it is a suitable alternative, SRE process produces
carbon dioxide as a byproduct, which is a main concern due to
environmental problems linked to global warming. To miti-
gate this effect, sorption enhanced steam reforming of
ethanol (SESRE) can be used. This process consists in the
addition of a solid sorbent into the reactor in order to remove
the CO2 produced during the reforming. As CO2 is removed
from the gas phase by the sorbent in situ, the reaction shifts to
the production of H2, which increases the production of
hydrogen, as predicted by Le Chatelier's principle [12e14]. The
main reactions of this process are the steam reforming of
ethanol (Eq. (1)), water gas shift reaction (Eq. (2)) and the
capture reaction performed by the sorbent (Eq. (3)):


C2 H5 OHðvÞ þ H2 OðvÞ %2COðgÞ þ 4H2ðgÞ DH0298 K ¼ 255:9 kJ mol
(1)


COðgÞ þ H2 OðvÞ %H2ðgÞ þ CO2ðgÞ DH0298 K ¼ 41:2 kJ mol (2)


CO2ðgÞ þ CaOðsÞ %CaCO3ðsÞ DH0298 K ¼ 178:2 kJ mol (3)
Olivas et al. [15] investigated SESRE using Ni/Al2O3 and CaO
as a sorbent. The authors obtained 97% of H2 concentration
during 120 min of reaction, similar value achieved in the
thermodynamics analysis (96.2%). They also concluded that
Na2ZrO3 can replace the use of CaO as a sorbent to capture the
CO2 from SRE due the limiting proprieties of CaO. Cortazar
et al. [16] studied the stability of NieCaO catalysts in SESRE in
fluidized bed reactor. The catalytic test was during 22 cycles,
and they reported that the CO2 uptake capacity had a decay
from 6 to 5.6 mmolCO2/gcat after 12 cycles. In addition, the H2
concentration had a slight decrease from 22.2 wt% to 18.5 wt%
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 0 2 6 3 e3 0 2 8 1
after 12 cycles of reaction and the following cycles remained
approximately constant. They also observed that performing
regeneration in the reforming temperature mitigates the
deactivation of the catalyst and enhances its stability. Xu et al.
[17] combined the traditional Ni/Al2O3 catalyst with the new
CaO-based porous spherical sorbent (prepared by a new
molding method) in the sorption-enhanced steam reforming
of ethanol process (SESRE) to improve H2 purity. The Ni/Al2O3
catalyst coupled with the modeled sorbent presented good
stability in the long-term reaction in repeated cycles and
hydrogen purity above 90%. Zeng et el [18]. also investigated
the steam reforming of ethanol with simultaneous CO2 cap-
ture. However, they used a chemical looping steam reforming
(CLSR) process as an alternative strategy to improve hydrogen
production using oxygen carriers. The authors used a CaO as a
sorbent and Ni-based catalysts varying NiO content and
reactional conditions. The results showed that the high ac-
tivity of the catalyst could be associated with the increase of
NiO content. According to these authors, more active sites on
the catalyst surface favor the formation of intermediates to H2
and CO. They also achieved 90% of ethanol conversion and
93% of H2 selectivity at 650  C using 20% of NiO content.
Although several works have been published previously in
the literature using CaO as a sorbent, not all have been suc-
cessful. This is because some authors reported the loss of CO2
capture capacity of CaO [19,20]. The strong decrease in sorp-
tion capacity after multiple carbonation/decarbonation cycles
is the main problem of CaO and sintering process is believed
to be the cause of the decay. Another reason may be related to
the formation of a non-porous calcium carbonate layer on the
surface of the CaO particles, limiting the CO2 diffusion to the
bulk of the grain [21].
One of the strategies to overcome this problem is the use of
additives to prevent the sintering process, by adding a re-
fractory phase to separate the sorbent particles [22]. Several
studies have shown that this refractory phase can be metal
oxides (MgO, Al2O3, ZrO2, TiO2, Y2O3, etc.) or mixture oxides
(Ca12Al14O33, Ca2SiO4, etc.) during sorption enhanced steam
reforming [23e26]. For example, Sang et al. compared the
performance of Ni/CaO and Ni/CaOeMgO catalysts during the
SESRE [24]. For both catalysts, 100% of ethanol conversion was
reached over the 10 cycles. However, the purity of hydrogen
decreased significantly for Ni/CaO sample, since the authors
observed a steady increase in CH4 concentration during the
cycles. On the other hand, Ni/CaOeMgO catalyst maintained
high hydrogen purity in 10 cycles, varying from 97.2 to 96 vol%
in the 1st and 10th cycles, respectively. The pre-breakthrough
duration of both catalysts was shortened, from 50 to 35 min in
the first and tenth cycles, respectively, which was attributed
to the decrease in the CO2 sorption capacity of the samples.
This work confirmed that adding MgO as a dopant improved
the catalyst cyclic stability, producing high-purity hydrogen
and reduced the coke deposition. Nimmas et al. [27] also
prepared a bifunctional material (Ni/CaOeCa12Al14O33) for
SESRE at 700  C. They used calcium acetate and urea as cal-
cium and carbonate precursors, respectively, via precipitation
method. The pre-breakthrough period lasted 60 min and
achieved 87% of hydrogen purity and this behavior was
maintained during 10 cycles of reaction. They observed the
30265
use of urea and calcium acetate favors high porosity of the
sorbent material and promotes mass transfer. This propriety
facilitates the contact between CO2 and CaO molecules that
are available in the surface of the material.
Chen et al. [28] studied different regeneration conditions
during SESRE in a fixed-bed reactor. The authors obtained
approximately 98% of H2 during catalytic tests for temperatures
between 600 and 850  C. They also revealed that there was a
formation of Ca12Al14O33 inert after regenerating at tempera-
tures above 800  C, which enhanced the duration of the sorp-
tion capacity of CaO in multiple cycles. They showed that using
air as a purge gas favors carbon decomposition over Ni/Ca
catalyst surface during regeneration process at 850  C. To
summarize, these conditions help to remove CO2 efficiently
and guarantee CO2 sorbent stability during the cycles of SESRE.
The addition of a refractory phase to CaO may enhance the
sorbent anti-sintering properties. There are few studies that
evaluate the stability over the cyclic reaction/regeneration
process for ethanol reforming since it is more common to find
sorption enhanced steam reforming of methane. Thus, the
main goal of this work is to study nickel-based catalysts
supported on CaO and CaOeMgOeAl2O3 for the sorption
enhanced steam reforming of ethanol (SESRE) over 10 cycles
aiming H2 production with high purity.
Experimental
Catalysts preparation
The experimental procedures to prepare the catalysts by sol-
gel method were based on Tian et al. [29]. This method was
chosen due to the formation of a porous microstructure that
favors the diffusion of CO2 into the CaO particle, improving
cyclic stability and reactivity [30]. Anhydrous citric acid
(99.5%, Dina ^ mica) and the precursors salts Ca(NO3)2$4H2O
(99%, Dina ^ mica), Mg(NO3)2$6H2O (98%, CAAL), Al(NO3)3$9H2O
(98%, Nuclear) and Ni(NO3)2$6H2O (97%, Vetec) were dissolved
into water considering a molar ratio water: anhydrous citric
acid of 25 and magnetically stirred at 90  C for 1 h. Then,
ethylene glycol (99.5%, Dina ^ mica) was added making the
molar ratio of metal: anhydrous citric acid: ethylene glycol at
1:2:2, maintaining the continuous stirring at 90  C until the
gelation promoted by the removal of excess solvent. The ob-
tained gel was dried overnight at 140  C in an oven to remove
water and causing the volume to increase considerably. The
as-burnt product was calcined up to 450  C at a heating rate of
1  C/min, temperature was held for 4 h and then heated until
800  C at 5  C/min and was kept constant for 2 h under airflow
to remove the residual organic materials. As the results re-
ported by Tian et al. [29] pointed that the best weight MgO/
Al2O3 ratio was 0.41, we used this same ratio of MgO/Al2O3 for
both doped samples. Therefore, the three samples used in this
work were synthesized with (i) 10% Ni þ 90%CaO; (ii) 10%
Ni þ 3.24% MgO þ 7.92% Al2O3 þ 78.84% CaO; and (iii) 10%
Ni þ 6.48% MgO þ 15.84% Al2O3 þ 67.68% CaO. The prepared
catalysts were named as Ni/Ca and Ni/CaMgAl-X, where X
represents the theoretical calcium content in each material
(X ¼ 79 and 68).
30266 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 0 2 6 3 e3 0 2 8 1
Characterization
X-ray diffraction (XRD) measurements were carried out using
a Shimadzu Lab X-XRD-6000 diffractometer operated at 40 kV
and 30 mA with CuK a radiation (¼0.154 nm). Data were
collected using the scan range (2 q) from 10 to 90 , steps of
0.02 and scan speed of 2 /min.
The N2 physisorption experiments were performed in
surface area and porosity analyzer, model NOVAtouch Gas
Sorption Analyzer LX1. Before the analysis, the samples were
previously heated under vacuum at 723 K overnight. The N2
physisorption on the solid was performed at liquid nitrogen
temperature (77 K). Finally, determination of specific surface
areas (Brunauer, Emmett and Teller Model e BET), volume
(Barrett, Joyner and Halenda Model e BJH) and pore sizes were
provided by the software installed in the equipment. This
analysis was carried out for the catalysts before (fresh) and
after (spent) the catalytic test.
H2-Temperature Programmed Reduction (TPR) was per-
formed in a quartz reactor which was heated by a furnace
connected to a temperature controller. Approximately 100 mg
of each sample was used. The samples were pretreated under
Ar flow at 500  C for 1 h and then cooled down to room tem-
perature. The tests were conducted using a thermal conduc-
tivity detector with a mixture of 1.96% H2 in Ar (30 mL/min),
increasing the temperature from room temperature to 1000  C
at a heating rate of 10  C/min. The experimental H2 con-
sumption was determined comparing the area below the
peaks obtained during TPR with a calibration curve using the
quantitative reduction of CuO to Cu0. The degree of reduction
was obtained by the ratio between experimental H2 con-
sumption and the theoretical one.
Scanning Electron Microscopy (SEM) analysis was used to
verify the changes in the structure and surface of the catalysts
before the sorption enhanced steam reforming of ethanol
(fresh) and after 10 reactional cycles (spent). The samples
were fixed in aluminum stubs with a double-sided carbon
tape, sputter-coated with one thin layer of gold (approxi-
mately 56 nm) using a metallizer (SCD050, Leica). The
morphology was investigated using scanning electron micro-
scopy, EVO-MA10, ZEISS with a voltage setting of 20 kV and
EHT (extra-high tension) mode. The SEM was also equipped
with an Energy-Dispersive X-ray Spectroscopy (EDS) detector
for the analysis of the individual chemical elements contained
in the samples.
The CO2 capture capacity was evaluated using a thermog-
ravimetric analyzer (TGA-50 from Shimadzu). A thermal pre-
treatment was performed before the carbonation/
decarbonation cycles at 800  C in order to remove impurities
and moisture from the samples. The carbonation cycle
occurred at 600  C during 30 min under 10.6 mL/min of diluted
CO2 in 60 mL/min of He. After that, CO2 feed was stopped,
remaining just the He flow rate and temperature was
increased to 800  C at 10  C/min, maintaining constant for
10 min aiming to regenerate the material by converting CaCO3
to CaO and releasing CO2. After the decarbonation process,
temperature was cooled down to 600  C, starting a new cycle.
A total of 20 carbonation/decarbonation cycles were per-
formed for each sample.
The thermogravimetric analysis (TGA) of the catalysts after
10 cycles of sorption enhanced steam reforming of ethanol,
also known as spent catalysts, were performed in the TGA-50
from Shimadzu. Approximately 10 mg of the regenerated
samples were inserted into the equipment and temperature
was increase from room temperature to 900  C at a heating
rate of 10  C/min with He atmosphere containing 15% O2.
Catalytic tests
The catalytic tests were divided into three steps: reduction,
reaction and decarbonation. The reduction occurred with 1.5 g
of sample in the quartz reactor with a mixture of 30 mL/min of
H2 and 100 mL/min of Ar. The furnace was heated up to 800  C
at a heat rate of 10  C/min and the temperature was held for
1 h. After the reduction step, the reactor was purged with Ar
and the temperature was cooled down until 600  C. A liquid
solution with a molar ratio water: ethanol of 6:1 was pumped,
evaporated forming a vapor phase with the following
composition: 27 mL/min of water, 4.5 mL/min of ethanol and
100 mL/min of argon. When the reactants stream was stabi-
lized, the reactants mixture was fed to the reactor. After 5 min,
the quantification of the products was done by a Gas Chro-
matograph GC-2014 (Shimadzu) every 7 min. The sorption
enhanced steam reforming of ethanol was performed at
600  C, 1 atm, for 117 min. After the reaction step, the furnace
was heated until 800  C at 10  C/min for 1 h under Ar atmo-
sphere aiming to decompose CaCO3, also known as the
decarbonation (or regeneration) step. The reaction and
decarbonation steps were repeated 10 times in a cyclic process
to monitor the performance of the catalysts.
After each cycle of reaction and decarbonation, the prod-
ucts distribution and ethanol conversion were calculated. Eq.
(4) was used to determine the molar fraction of the products
on dry basis, where yi represents the molar fraction of each
product in gas phase, Fi and Fj are the molar flow rate on dry
basis, in gas phase, from the species i and j at the exit of the
reactor and N represents the total number of species identified
by the GC.
Fi
yi ¼ (4)
P
N
Fj
j¼1
For the ethanol conversion, the pumped solution was
injected into the gas chromatograph before the beginning of
the cycle (feed, with the reactor in the bypass position) and
after each cycle (condensed solution), in order to obtain the
number of moles of ethanol in the feed and at the exit of the
reactor. Eq. (5) was used to calculate ethanol conversion
(XC2 H5 OH ), where FC2 H5 OH represents the molar flow rate of
ethanol.
   
FC2 H5 OH feed  FC2 H5 OH exit
XC2 H5 OH ð%Þ ¼   ,100 (5)
FC2 H5 OH feed
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 0 2 6 3 e3 0 2 8 1
Results and discussion
Characterization
Fig. 1 shows the diffractograms obtained at room temperature
for the calcined samples. Ni/Ca catalyst presented CaO,
CaCO3, Ca(OH)2 and NiO phases. For CaO (PDF 82e1691), the
characteristic peaks are represented in 2q positions equal to
37.4, 53.9 and 64.2 , with the corresponding planes (200), (220)
and (311), respectively. CaCO3 peak (PDF 4e636) was also
detected in 43.0 with the plane (202) suggesting that CaCO3
was incompletely decomposed to CaO and CO2 during calci-
nation at 800  C. Due the high hygroscopicity of CaO, most of
the peaks observed correspond to Ca(OH)2 (PDF 81e2041)
which diffraction angles are at 18.0, 28.7, 34.1, 47.1, 50.8, 71.8,
79.1 and 84.8 related to planes (001), (100), (101), (102), (110),
(022), (113) and (014), respectively. Besides the previously
mentioned phases, Ni/CaMgAl-79 and Ni/CaMgAl-68 catalysts
showed diffraction peaks related to the MgO phase. The three
samples also present NiO peaks (PDF 65e2901) at 37.1, 43.1,
62.6, 75.0 and 79.0 with the corresponding phases (111), (200),
(220), (311), (222). Unfortunately, these diffraction peaks have
coincidentally, the same 2q values for MgO according to PDF
65e476. Probably there are NiO and MgO peaks overlapping in
the Ni/CaMgAl-79 and Ni/CaMgAl-68 catalysts. Therefore, it is
not possible to affirm that the segregated MgO phase is pre-
sent. Sang et al. [24] prepared Ni/CaO catalysts modified with
the addition of MgO and they found similar phases to this
work after performing XRD analysis: CaO, CaCO3, Ca(OH)2,
NiO. Dang et al. [23] investigated Ni/CaMg catalysts prepared
by hydrothermal method, calcined at 800  C. The authors re-
ported the presence of CaO, MgO, and Ca(OH)2, which is in
agreement with the results presented in this work.
It is worth to point out that diffraction peaks related to
Al2O3 phases were not detected. This is an indication that the
aluminum added may have been converted to small amounts
of other phases such as nickel aluminate spinel (NiAl2O4).
Although the presence of small quantities of this kind of
spinel as a surface species is difficult to be detected by XRD, it
Fig. 1 e XRD patterns for the Ni/Ca, Ni/CaMgAl-79 and Ni/
CaMgAl-68 calcined catalysts.
30267
is considered to be the cause of a peak at high temperature in
TPR of nickel catalysts [31].
Fig. 2 displays XRD patterns obtained at room temperature
for the passivated catalysts. After passivation process, the
peaks that corresponds to reduced nickel appeared at 44.5,
51.8 and 76.4 (PDF 65e2865) which can be correlated to the
respective planes (111), (200) and (220). It is also important to
notice that Ca(OH)2 phase was not presented in the passivated
catalysts. This was due to the calcination process favoring the
dehydration reaction, decomposing into CaO, before passiv-
ation stage was completed. CaO phase (PDF 82e1690) was
present at 32.2, 37.4, 53.9, 64.2, 67.5, 79.8 and 88.7 with the
planes (111), (200), (220), (311), (222), (400) and (3 3 1), respec-
tively. CaCO3 (PDF 4e636) had a single peak at 43.0 with the
plane (202). The diffraction peaks detected at 37.1, 43.1 and
62.6 could be related to NiO which would suggest that nickel
was not completely reduced. The presence of this oxidized
nickel could be related to the passivation procedure. Unfor-
tunately, the peaks observed at 2q positions equal to 37.1, 43.1
and 62.6 are also characteristic of CaO, CaCO3 and MgO.
Therefore, we could not rule out the presence of NiO phase,
nor CaO, CaCO3, and MgO.
Fig. 3 displays TPR profiles obtained for Ni/Ca, Ni/CaMgAl-
79 and Ni/CaMgAl-68 catalysts after calcination process. Ni/Ca
catalyst presented two peaks at 489 and 628  C. According to
the literature, the lower temperature peak can be attributed to
NiO that had weak interaction with CaO. The higher temper-
ature peak can be associated to NiO that presented a stronger
interaction with CaO. This result was previously reported by
Sang et al. [24], since they studied Ni/CaO catalysts doped with
MgO using coprecipitation method, with a calcination process
of 2 h at 800  C. They also found reduction temperature values
between 477 and 601  C, which is similar to our work. For Ni/
CaMgAl-79 and Ni/CaMgAl-68 catalysts, the TPR profiles are
very similar. The only differences are the peak intensities and
the change of the reduction temperatures. Both profiles
exhibit two distinct peaks, a smaller one between 500 and
730  C and a larger one between 750 and 1000  C. The first half
of the lower temperature peak can be associated to NiO
interacting with CaO, as observed for Ni/Ca. As reported by
Fig. 2 e XRD patterns for the Ni/Ca, Ni/CaMgAl-79 and Ni/
CaMgAl-68 passivated catalysts.
30268 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 0 2 6 3 e3 0 2 8 1
Fig. 3 e TPR of the calcined catalysts: Ni/Ca, Ni/CaMgAl-79
and Ni/CaMgAl-68.
Sang et al. [24], the second half of the low temperature peak
could be related to the reduction of NixMg1-xO. Since many
MgO and NiO diffraction peaks have the same 2q positions, it
was not possible to identify the formation of small amounts of
this phase by X-ray diffraction. However, the formation of this
phase has been previously reported in the literature for ma-
terials prepared in a similar way used in this present work. For
the Ni/CaMgAl-79 catalyst, the formation of this mixed oxide
phase would be less notable probably because this sample has
a low theoretical MgO content (3.24 wt%). The larger H2 con-
sumption peak, above 880  C, could be associated to the
reduction of NiAl2O4 spinel [17,32,33]. Liu et al. [33], for
instance, prepared a NieCaOeAl2O3 catalyst by sol-gel
method and they found NiAl2O4 phase after performing XRD.
They also reported that the catalyst would hardly be reduced
completely at 850  C due to this spinel. Furthermore, they also
concluded that higher amount of calcium in the material in-
hibits the spinel formation, generating a lower reduction
temperature, which corroborates with the results found by
our work.
TPR was also performed under the same temperature
conditions used in the catalyst activation process preceding
the SESRE reaction: the furnace was heated until 800  C at a
heating rate of 10  C/min and kept at this temperature for 1 h.
This procedure was conducted in order to confirm that the
catalyst activation conditions were appropriate for the cata-
lytic tests and to estimate the degree of reduction.
TPR profiles obtained under activation conditions were
similar to the standard ones presented in Fig. 3. The main
difference is that there was an increase of H2 consumption,
since in these experiments, the catalysts remained for a
longer time under hydrogen atmosphere. Table 1 shows
experimental H2 consumption and degree of reduction ob-
tained. The theoretical H2 consumption for total NiO reduc-
tion in these samples is 13,388 mmol/gcat.
All the samples presented a degree of reduction above 90%,
indicating that this condition is favorable to activate the cat-
alysts. Most of the nickel is in the reduced form, which is the
active phase for steam reforming of ethanol promoting the
conversion of ethanol and water into high purity H2. It is also
Table 1 e Experimental consumption of H2 and degree of
reduction of the catalysts during the TPR under the same
temperature conditions used in activation process.
Catalyst H2 consumed Degree of
(mmol/gcat) Reduction (%)
Ni/Ca 12,234 91
Ni/CaMgAl-79 14,411 100
Ni/CaMgAl-68 12,054 90
worth noting that during catalytic tests, the hydrogen pro-
duced can promote reactivation of the sites that have been
oxidized throughout the reaction cycles, which may help to
maintain the catalytic activity. Passos et al. [34] also reported
this phenomenon using Ni and NiCu catalysts during oxida-
tive steam reforming of ethanol. In their work, the hydrogen
formed through the steam reforming of ethanol self-
reactivated the catalyst and ethanol conversion increased
from 89 to 98%. In addition, the TPR profiles using the same
temperature conditions as the one used in the reduction
process before catalytic tests showed that Ni/CaMgAl-79
catalyst achieved a degree of reduction of 100%.
The samples were submitted to carbonation and decar-
bonation cycles using thermogravimetric analysis. Fig. 4
shows the results of carbonation and decarbonation for all
the catalysts during 20 cycles. Ni/Ca catalyst (Fig. 4a) pre-
sented a weight increase of 47.7% in the 1st cycle and 19.7% in
the 20th. For Ni/CaMgAl-79 (Fig. 4b), the weight increases were
31.7 and 24.2% in the 1st and 20th cycles, respectively. The Ni/
CaMgAl-68 catalyst (Fig. 4c) presented an increase of 22.3% in
the 1st cycle and 17.7% in the 20th. These weight increases are
attributed to the conversion from CaO to CaCO3 during the
carbonation cycle, when the sample is exposed to a diluted
CO2 flow. The decrease of CO2 uptake capacity can be probably
attributed to the incomplete conversion of CaCO3 to CaO due
to CaO sintering and the multiple carbonation-decarbonation
cycles that caused the alteration in the physical properties of
the material [32]. Wu et al. [32] studied Ni/CaOeAl2O3 bifunc-
tional materials in cyclic steam reforming of ethanol with
simultaneous CO2 capture. They showed that depending on
the Ca/Al molar ratio of the material (above 3), the addition of
inert materials possibly delays the sintering of CaO during
calcination and contributes considerably to the stability and
CO2 uptake capacity.
In order to have a suitable comparison, the CO2 sorption
capacity was normalized by mass of CaO, considering the
theoretical CaO content of Ni/Ca, Ni/CaMgAl-79 and Ni/
CaMgAl-68 catalysts as 90, 78.84 and 67.68%, respectively.
Fig. 5 shows that in the first 3 cycles Ni/Ca catalyst has the
highest CO2 sorption capacity. As the number of cycles
increased, the ability to capture CO2 of Ni/Ca catalyst
decreased. This is probably because this sample had no spacer
material in its composition. Therefore, the sintering process
could not be avoided, and this sample was the worst one in
terms of CO2 capture capacity at the end of 20 carbonation/
decarbonation cycles. Similar results were previously re-
ported by Liu et al. [33], since they prepared Ni/CaO catalyst
using sol-gel method and after 20 cycles the sorption capacity
was 0.21 gCO2/gCaO while in our work we achieved 0.19 gCO2/
gCaO after the same number of cycles. The authors also
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 0 2 6 3 e3 0 2 8 1
Fig. 4 e 20 carbonation and decarbonation cycles from: a)
Ni/Ca, b) Ni/CaMgAl-79 and c) Ni/CaMgAl-68 catalysts.
Tcarbonation ¼ 600  C (15% of CO2 in He atmosphere),
Tdecarbnation ¼ 800  C (pure He atmosphere).
observed that their bifunctional material formed by NieCaeAl
had a capacity of 0.44 gCO2/gCaO at the end of the 20 cycles,
while our best catalyst, Ni/CaMgAl-79, was 0.31 gCO2/gCaO (Ni/
CaMgAl-79). Ni/CaMgAl-79 and Ni/CaMgAl-68 samples had the
addition of spacers, which play an important role in the
30269
Fig. 5 e Comparison among CO2 uptake capacity over 20
cycles from Ni/Ca, Ni/CaMgAl-79 and Ni/CaMgAl-68
catalysts.
stability of the catalyst, reducing its sintering and facilitating
the molecular contact between CO2 and CaO [15]. In their
work, Aceves Olivas et al. [15] studied steam reforming of
ethanol with CO2 capture using Ni/Al2O3 catalyst prepared by
incipient impregnation and three different sorbents: CaO,
dolomite (CaOeMgO) and Na2ZrO3. They used SEM technique
before and after reaction and showed that the addition of MgO
from dolomite inhibited the sintering process by mitigating
the agglomeration of particles.
Some models are considered to describe CO2 capture ki-
netics. One of them is the double-exponential model and it is
feasible to explain kinetic mechanisms that involves two
steps: a rapid and a slow phase. The rapid phase is controlled
by strong surface reactions (chemisorption of CO2 on CaO)
while the slow one is controlled by intra-particle CO2 diffusion
through the CaCO3 external layer formed [35]. The TGA data
acquired were adjusted by the double-exponential model
given in Eq (6). The two exponential terms represent the two
steps of the adsorption kinetics for an adsorbent: the rapid
and the slow carbonation. In Eq. (6), y represents the per-
centage of CO2 adsorbed at a given time, t is time, A and B are
pre-exponential factors, k1 and k2 represent the kinetic rate
constant of the two adsorption steps and C is the percentage
of CO2 adsorbed at the equilibrium [35,36].
y ¼ A,ek1 ,t þ B,ek2 ,t þ C (6)
Table 2 summarizes the kinetic parameters from all the
three catalysts in the 1st and 20th cycles and the ratio k1 /k2
and coefficient of determination (R2). According to Guo et al.
[37] if k1 /k2 value is more than 1, the CO2 diffusion through an
external layer of CaCO3 to react with CaO is the controlling
step. On the other hand, if k1 /k2 is lower than 1, the surface
reaction is the rate-determining step. Fig. 6 shows the exper-
imental data from TGA and the double-exponential model
fitting for all the three catalysts. The correlation coefficients
for the fitting (R2) are approximately 1, which indicates a good
agreement between the double-exponential model and the
experimental data This fitting was performed by Vieira et al.
30270 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 0 2 6 3 e3 0 2 8 1
Table 2 e Estimated parameters by the double-exponential model using TGA data for Ni/Ca, Ni/CaMgAl-79 and Ni/CaMgAl-
68 catalysts, comparing the 1st and 20th cycles.
Catalyst cycle A B
Ni/Ca 1st 34.075 32.378
20th 9.692 11.936
Ni/CaMgAl-79 1st 15.199 19.698
20th 16.355 9.970
Ni/CaMgAl-68 1st 14.391 14.640
20th 12.810 7.045
[38], which studied SESRE using NieCaO bifunctional catalysts
with Ca11Al14O33 and CeO2 as refractory materials. After 18
cycles of carbonation/regeneration in TGA they found k1 /k2
values higher than 1. This means that k2 value is much smaller
than k1 in the 1st and 18th cycles, indicating that CO2 diffusion
is the controlling step of the sorption process. In this present
work, both 1st and 20th cycles also indicated a k1 /k2 value over
1 for all the three catalysts, which suggests that the CO2
diffusion is the rate-limiting process. The smallest values for
k1 and k2 were calculated for Ni/Ca catalyst, which agrees with
the highest loss of CO2 uptake capacity of this catalyst.
Yoon, Mun and Lee [39] used a deactivation model to fit the
cyclic sorption data, which describes the sorbent activity in
each cycle. This model allows to calculate the deactivation
coefficients (kd ) of the materials based on Eq. (7) as follows:
a ¼ K,ekd ,N
The parameter N represents the number of carbonation/
decarbonation cycles, a indicates the sorbent activity at the
Nth cycle (amount of CO2 sorption uptake at the Nth cycle
divided by 78.5 wt%, that is the maximum CO2 sorption ca-
pacity of CaO) and the model pre-exponential factor is K.
According to these researchers, the deactivation coefficient
can indicate the degree of loss of the CO2 sorption perfor-
mance of the materials per cycle, and sorbents with lower
values tend to have a higher cyclic stability. The calculated
values of the deactivation model are presented in Table 3 for
the three CaO-based sorbents. Fig. 7 shows the experimental
data from the sorption activity over the cycles coupled with
the activity found by the deactivation model using the esti-
mated parameters. It is clear that Ni/CaMgAl-68, and Ni/
CaMgAl-79 presented kd values around 70% lower than Ni/
Ca. It is notable that adding spacers to Ni/Ca lead to the
decrease of the deactivation coefficient value, which means
that the dopped catalysts are more stable, as expected. Vieira
et al. [38] also used the deactivation model and they found
similar values to the parameter -kd comparing to our work
(0.048, 0.016 and 0.022). The first value reported by the au-
thors was higher because the sorption material was pure
CaO, so it has a more pronounced deactivation. The last two
were lower, since the CaO was improved by the addition of
mayenite and mayenite þ CeO2.
Catalytic tests
The catalytic tests were carried out in 10 cycles of reaction and
decarbonation for the SESRE of Ni/Ca, Ni/CaMgAl-79 and Ni/
CaMgAl-68 catalysts to compare the enhancement of the
C k1 (s1) k2 (s1) k1/k2 R2
60.502 0.0078 0.0005 15.600 0.992
20.278 0.0215 0.0007 30.714 0.998
33.503 0.0221 0.0012 18.417 0.998
25.088 0.0337 0.0012 28.083 0.996
25.687 0.0109 0.0008 13.625 0.988
18.020 0.0329 0.0011 29.909 0.994
materials. The ethanol conversions for all samples were
approximately 100% for all cycles. The main product was H2
and the other products observed were CO2, CO and CH4. In
Fig. 8 - Fig. 10, regions I, II and III represent the pre-
breakthrough, breakthrough and post-breakthrough periods,
respectively. The dashed lines refer to the products distribu-
tion predicted by the thermodynamics analysis in the equi-
librium using the software CEA (Chemical Equilibrium with
Applications) from NASA.
For Ni/Ca catalyst, Fig. 8 shows that during the pre-
breakthrough stage of the 1st cycle, the H2 concentration
reached values between 83.6 and 87.6%. The increase in
hydrogen purity may be an indication that the catalytic sys-
tem was being previously self-reactivated by the H2 formed by
SESRE and then it was stabilized in the subsequent cycles.
After 80 min of reaction (post-breakthrough), H2 decreased to
(7) 72.2% and the concentration of CO, CH4 and CO2 were 15.0, 2.0
and 10.8%, respectively. In the 10th cycle, the H2 and CO2
mean molar fractions were 88.3 and 1.5% during the pre-
breakthrough period. After 75 min of reaction (post-break-
through), H2 and CO2 molar fraction decreased whereas CO
increased, indicating that the reverse water-gas shift (RWGS)
reaction is happening, since high temperatures (700e1000 K)
favor this reaction, as affirmed by El-Hafiz et al. [40].
Comparing the first and 10th cycle, the pre-breakthrough
period decreased from 50 to 28 min. Furthermore, post-
breakthrough period increased over the cycles and started at
45 min in the last one, due to the loss of CO2 uptake capacity.
Liu et al. [33] performed SESRE at 600  C and S/E ¼ 4 using
50 mg of Ni/CaO catalyst. They observed that in the 1st cycle
the production of high purity H2 lasted for 12.5 min and after
10 cycles this time decreased to 10 min. The present work had
a longer pre-breakthrough period in the 1st and 10th cycles
when compared to the authors. This is possibly due to the
larger amount of catalyst in the reactor and the differences in
the reaction systems.
For Ni/CaMgAl-79, Fig. 9 shows that the pre-breakthrough
duration was 45 min in the 1st and 2nd cycles and was
reduced to 30 min at the end of the 10th cycle. At the same
time, the post-breakthrough started at approximately 80 min
in the 1st cycle and maintained the same value up to the 10th
cycle. Similar to Ni/Ca catalyst, the decrease of the pre-
breakthrough period is attributed to the reduction in the
capture capacity of the catalyst over the reaction/decarbon-
ation cycles. Liu et al. [33] also prepared a bifunctional catalyst
to perform SESRE using Al as a stabilizer material. They re-
ported that their material achieved 100% longer pre-
breakthrough time when compared to Ni/CaO. Furthermore,
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 0 2 6 3 e3 0 2 8 1
Fig. 6 e Experimental and double-exponential model fitting
in the 1st and 20th cycles for a) Ni/Ca, b) Ni/CaMgAl-79 and
c) Ni/CaMgAl-68 catalysts.
30271
Table 3 e Estimated parameters by the deactivation
model using TGA data for Ni/Ca, Ni/CaMgAl-79 and Ni/
CaMgAl-68 catalysts.
Catalyst K -kd R2
Ni/Ca 0.5777 0.0563 0.9566
Ni/CaMgAl-79 0.4065 0.0149 0.9794
Ni/CaMgAl-68 0.2791 0.0130 0.9680
Fig. 7 e Experimental and deactivation model fitting over
the cycles for Ni/Ca, Ni/CaMgAl-79 and Ni/CaMgAl-68
catalysts.
the addition of Al exhibited suitable CO2 removal performance
and stability for SESRE process. In addition, the purity of the
hydrogen produced by this catalyst gradually increased during
several SESRE cycles. This also can be seen in this work when
comparing the 1st and 2nd cycles.
Fig. 10 shows the performance of a material similar to the
previous one, changing only its weight composition. The pre-
breakthrough duration in the 1st cycle was 30 min and
maintained the same until the 10th cycle, indicating that the
addition of stabilizer material promoted an improvement in
the cyclic stability of CaO. Besides, the post-breakthrough in
the 1st cycle started at nearly 79 min and was reduced to
59 min in the 2nd cycle and kept the same value until the 10th
cycle. Although the pre-breakthrough time remains constant
throughout all cycles, this catalyst cannot be considered the
best. This is precisely because it has a low capacity to capture
CO2 right in the first cycle, and the material quickly becomes
saturated. Dewoolkar and Vaidya [41] studied SESRE using
hydrotalcite-like materials (HTlc). The catalytic test was per-
formed using several materials, including NiMgHTlc and
NiCaHTlc catalysts at 673 K and S/C ¼ 9. The pre-breakthrough
period in the 1st cycle was approximately 15 and 10 min,
respectively. NiMgHTlc achieved 80% of hydrogen purity
whereas NiCaHTlc had nearly 70%. Ca2þ probably needs to be
30272 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 0 2 6 3 e3 0 2 8 1

Fig. 8 e Product distribution from the SESRE for the Ni/Ca
catalyst in the 1st, 2nd and 10th cycles. Reactional
conditions: Treaction ¼ 600  C, molar ratio steam/
ethanol ¼ 6, W/F ¼ 11.41 min·mgcat·mL¡1. Legend: : H2; :
CO; : CH4 and : CO2.
Fig. 9 e Product distribution from the SESRE for the Ni/
CaMgAl-79 catalyst in the 1st, 2nd and 10th cycles.
Reactional conditions: Treaction ¼ 600  C, molar ratio steam/
ethanol ¼ 6, W/F ¼ 11.41 min·mgcat·mL¡1. Legend: : H2; :
CO; : CH4 and : CO2.
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 0 2 6 3 e3 0 2 8 1 30273

available in the sample and not inserted into the structure to

guarantee better CO2 capture capacity. This explains why the
authors obtained lower concentrations of hydrogen when
compared to this work. In addition, NiMgHTlc catalyst was
stable during 18 cycles at 573 K and S/C ¼ 5. Performing SESRE
at 773 K and S/C ¼ 10, NiCaHTlc remained stable during 6
multi-cycles. Dang et al. [23] also prepared a bifunction cata-
lyst using MgO as a refractory material and they observed that
this incorporation enhanced the sorption performance by
minimizing the particles of CaO. This characteristic of cyclic
stability over reaction cycles is desired in sorption enhanced
steam reforming reactions and can be seen in the perfor-
mance of the catalysts doped with MgO and Al2O3 in this work.
The catalyst containing only Ni and CaO showed the
longest pre-breakthrough period in the first cycle

Fig. 11 e XRD patterns for the Ni/Ca, Ni/CaMgAl-79 and Ni/

CaMgAl-68 spent and passivated catalysts.

Table 4 e Average crystallite size (CaO and Ni0) for the

passivated catalysts.
Catalysts Average crystallite size (nm)
CaO Ni0
passivated spent and passivated spent and
passivated passivated
Ni/Ca 32.3 44.4 19.0 27.7
Ni/CaMgAl-79 29.5 31.3 19.3 25.3
Ni/CaMgAl-68 26.0 35.0 19.0 37.0

Table 5 e Textural properties of fresh and spent Ni/Ca, Ni/

CaMgAl-79 and Ni/CaMgAl-68 catalysts.
Ni/Ca Ni/CaMgAl-79 Ni/CaMgAl-68
fresh spent fresh spent fresh spent
Fig. 10 e Product distribution from the SESRE for the Ni/ Surface area 46 58 79 66 37 48
CaMgAl-68 catalyst in the 1st, 2nd and 10th cycles. (m2/g)
Reactional conditions: Treaction ¼ 600  C, molar ratio steam/ Pore volume 0.147 0.205 0.177 0.215 0.115 0.158
ethanol ¼ 6, W/F ¼ 11.41 min·mgcat·mL¡1. Legend: : H2; : (cm3 /g)
Average pore 6.4 7.0 4.5 6.6 6.3 6.5
CO; : CH4 and : CO2.
size (nm)
30274 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 0 2 6 3 e3 0 2 8 1

(approximately 50 min), when compared to the other cata-
lysts. This fact is justified by the larger amount of CaO
available for CO2 capture. However, this excellent perfor-
mance was not repeated in the next cycles. This decline in
the capture capacity can be related to the low Tammann
temperature of CaCO3, also indicating the need to add sta-
bilizing material to the catalyst. These results corroborate
with Vieira et al. [38], since they used 2.2 g of bifunctional
material (NieCaO), as this work, and obtained 55 min in pre-
breakthrough of the 1st cycle and 35 min in the 10th cycle
over sorption enhanced steam reforming of ethanol
reaction.
For all catalysts it was possible to observe that the methane
formation remained low, which would imply that the CO
methanation reaction was mitigated. This result was also re-
ported by Liu et al. [33]. The distribution of products of
the reaction was not exactly the same values predicted by
the thermodynamic analysis. One can explain this by the
experimental setup that is in continuous flow, in other words,
the formed species are continuously exiting the reactor.

Fig. 12 e N2 adsorption and desorption isotherms for Ni/Ca, Ni/CaMgAl-79 and Ni/CaMgAl-68 spent and passivated
catalysts.
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 0 2 6 3 e3 0 2 8 1 30275

Thermodynamic analysis, on the other hand, considers a
batch system, where all species stays in the reactor until
equilibrium is reached [15].
The role of Ca in Ni/Al2O3 catalyst during the steam
reforming of ethanol was investigated by Choong et al. [42].
The presence of Ca can enhance water adsorption and
dissociation, providing high concentration of OH groups
around Ni particles. This also improves ethanol adsorption,
which facilitates the rapid formation of CH3COO from CH3CO
and CH3CHO intermediates. The OH groups can also facilitate
CeC bond cleavage from CH3COO, which lead to the formation
of the same products detected during the reaction in the
present work: H2, CH4, CO and CO2.
Characterizations after catalytic tests
To identify the existing phases in the catalyst after 10 reaction
cycles, XRD of the spent and passivated catalysts were per-
formed, as shown in Fig. 11. XRD profiles of the used catalyst
are quite similar to the passivated catalysts, although the
CaCO3 phase was not present. Another important fact was the
absence of peaks referring to the Ca(OH)2, since the material
was activated and passivated right before the DRX analysis. As
previously observed, we did not detect Al2O3 and NiAl2O4
phases nor the presence of NixMg1-xO phase, although the
formation of NiAl2O4 and NixMg1-xO would be coherent with
TPR results.

Fig. 13 e Pore size distribution for Ni/Ca, Ni/CaMgAl-79 and Ni/CaMgAl-68 spent and passivated catalysts.
30276 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 0 2 6 3 e3 0 2 8 1
The average crystallite sizes of CaO and Ni0 were calcu-
lated from the XRD data obtained for the passivated catalysts
and passivated after SESRE cycles, using the Scherrer Equa-
tion. The results can be found in Table 4. It was not possible
to calculate the average crystallite diameter for NiO, because
the peaks for this phase are interfered by other phases that
may be present (MgO, CaO). The peak at 2q ¼ 37.2 was used
for the calculation of CaO average crystallite size and the one
at 2q ¼ 44.5 was used for Ni0. The three passivated catalysts
showed CaO average crystallite sizes between 26 and
32.3 nm. Gao et al. [43] reported that their Ni/CaO bifunc-
tional material achieved a crystal size of 35.7 nm for CaO
after reduction. After the 10 SESRE cycles, in this present
work, CaO increased their size for all catalysts. Furthermore,
the addition of stabilizing material caused the average
crystallite size to be smaller at the end of the reaction cycles
when compared to Ni/Ca. The three passivated catalysts
presented the same Ni0 average crystallite size, around
19 nm, which increased their values after 10 cycles of reac-
tion. Hu et al. [9] prepared a Ni/CexZr1-xO2-CaO catalyst and
the crystallite size of Ni achieved 23.2 nm in the first cycle of
Fig. 14 e Scanning Electronic Microscopy of fresh and used catalysts. Magnification 10,000 times.
SESR of acetic acid and 30.8 nm in the 9th cycle, which are
quite similar to this work.
After the reaction cycles, N2 physisorption technique
(adsorption and desorption) was carried out for the textural
characterization of the catalysts. Table 5 presents the results
obtained by N2 physisorption for the fresh and spent catalysts.
For Ni/Ca and Ni/CaMgAl-68 catalysts, the surface area
increased after the 10 reaction-regeneration cycles, which is
an unexpected but plausible finding. For example, Ohenoja
et al. [44] used peat-wood fly ash as a sorbent under CO2 at-
mosphere and they found that the carbonation process
increased surface area between 40 and 57%. They assumed
that this process favors the filling of large pores with CaCO3,
which increases pore volume and corroborates with the re-
sults of the present work. The addition of Mg and Al in Ni/
CaMgAl-79 catalyst provided a higher surface area when
compared to Ni/Ca. After 10 reaction-regeneration cycles, the
surface area decreased but this sample still presented the
highest value among the spent samples. These results concur
well with the behavior observed during the TGA and reaction-
regeneration cycles.
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 0 2 6 3 e3 0 2 8 1 30277

Fig. 12 shows the adsorption and desorption isotherms for
both fresh and spent samples.
All catalysts obtained type IV isotherms, with different
paths in N2 adsorption and desorption, which characterizes a
hysteresis. In addition, this type of isotherm helps to classify
the material as mesoporous, indicating that the pore size is in
the range between 2 and 50 nm [45], values that are in
accordance with Table 5.
Regarding the average pore size, the data in Table 5 showed
that the average pore sizes increased for the three catalyst
after the reaction-regeneration cycles. Fig. 13 presents pore
size distributions for these materials.
All the samples presented a bimodal pore size distribution.
However, after 10 reaction-regeneration cycles, the area of the
first mode decreased and the other one increased. This was
also observed by Alvarez et al. [46] using CaO for several
calcination and carbonation cycles. They also revealed that
the material had an internal sintering during calcination,
which constantly increases the average pore size as the
number of cycles increases.
The Scanning Electron Microscopy (SEM) analysis was
carried out for the calcined and spent catalysts and the results
can be seen in Fig. 14. After the 10 reaction cycles Ni/Ca
catalyst presented greater contact between the CaO particles,

Fig. 15 e EDS mapping for the fresh Ni/CaMgAl-79 catalyst.

30278 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 0 2 6 3 e3 0 2 8 1

showing the sintering process of the material. This caused the
increase of the particles and consequently the reduction of the
specific surface area. The catalysts doped with metal oxides,
on the other hand, showed a well-preserved structure after
the reaction when compared to the Ni/Ca catalyst. This is an
expected behavior, because the inert ones acted as a physical
barrier and minimized the sintering of CaO particles during
the cycles [30]. From the images, it was not possible to observe
the apparent formation of filamentous carbon on the surface
of the spent catalysts, one of the main causes of deactivation
of nickel-based catalysts in the reforming reactions by
blocking the active sites. The absence of filamentous carbon in
the SEM analysis could have occurred due to the high amount
of catalyst mass, since all the ethanol was converted before
even reaching the end of the catalytic bed, resulting in the
formation of coke in the sample in a non-homogenous way.
Besides the SEM, the mapping of the elements present in
the catalysts studied in this work was performed. Figs. 15 and
16 show the mapping for the fresh and spent catalyst that
proved to be more stable throughout the catalytic tests. From
the EDS mapping, it was possible to notice that the dopant
particles, MgO and Al2O3, are uniformly supported on the CaO

Fig. 16 e EDS mapping for the spent Ni/CaMgAl-79 catalyst.

 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 0 2 6 3 e3 0 2 8 1
particles and the nickel element, responsible for the reforming
reaction, is well distributed throughout the material. It can
also be seen that Ni particles for the spent catalyst are larger
than the fresh one, which is coherent with the results of the
average crystallite size. According to Wang et al. [47], the
nickel present in the structure of NiAl2O4 can lead to high
stability and good dispersion for the metal particles, which
can be observed by EDS. The other EDS mapping for the fresh
and spent catalysts show a similar behavior when compared
to the Ni/CaMgAl-79 catalyst.
Conclusions
This work showed the effect of adding Mg and Al on the per-
formance of Ni/CaO catalyst during SESRE process. XRD re-
sults showed the presence of CaO, NiO, CaCO3, Ca(OH)2 and
NiO for calcined samples and Ni0 for reduced ones. TPR pro-
files showed the occurrence of weak and strong interactions
between NiO and CaO. The samples doped with Mg and Al
presented a high temperature of reduction, which could
indicate NiAl2O4 formation even though no characteristic
peaks of this spinel were shown by XRD. TPR results also
showed that the temperature used for catalyst activation was
enough to ensure a high degree of Ni reduction. The TGA
evaluated the effect of the addition of inert materials into CaO
during 20 carbonation and decarbonation cycles. Comparing
the Ni/Ca catalyst to the doped ones, the deactivation co-
efficients were 3.8 times for Ni/CaMgAl-79 and 4.3 for Ni/
CaMgAl-68, when compared with Ni/Ca. The addition of
MgO and Al2O3 to CaO improved the performance of the ma-
terials during CO2 capture-regeneration cycles. For instance,
the deactivation coefficients, calculated by the TGA data
modeling, were around 4 times smaller for the doped mate-
rials than the value determined for Ni/Ca catalyst. Although
Ni/CaMgAl-68 showed higher stability, in terms of CO2 capture
capacity, Ni/CaMgAl-79 showed the best result after the SESRE
cycles. The results showed that the addition of these dopants
increased cyclic stability and reduced the loss of CO2 uptake
capacity. For all the catalysts during SESRE, the H2 mean molar
fraction, in dry basis, was almost the same in the 1st cycle
during pre-breakthrough stage. Furthermore, Ni/Ca and Ni/
CaMgAl-79 catalysts had a slight increase in the H2 molar
fraction during pre-breakthrough of the 10th cycle. The results
showed that doping Ni/Ca catalyst with Al and Mg may lead to
materials with higher stability, which can be advantageous for
SESRE processes. In addition, this work shows that there is an
optimum amount to be doped, however, a more detailed study
is needed to ensure better characteristics, such as higher CO2
capture capacity and stability of the material.
Declaration of competing interest
The authors declare that they have no known competing
financial interests or personal relationships that could have
appeared to influence the work reported in this paper.
30279
Acknowledgements
The authors are grateful to CAPES, CNPq and FAPEMIG for the
financial support and scholarships. We also acknowledge the
support of FEQUI/UFU and the Grupo de Materiais Inorga^nicos
^ ngulo (GMIT), a research group supported by FAPEMIG
do Tria
(APQ-00330-14), for N2 physisorption analyses.
references
[1] Chen F, Tao Y, Ling H, Zhou C, Liu Z, Huang J, et al. Ni-Cu
bimetallic catalysts on Yttria-stabilized zirconia for
hydrogen production from ethanol steam reforming. Fuel
2020;280:118612. https://doi.org/10.1016/j.fuel.2020.118612.
[2] Oni AO, Anaya K, Giwa T, Di Lullo G, Kumar A. Comparative
assessment of blue hydrogen from steam methane
reforming, autothermal reforming, and natural gas
decomposition technologies for natural gas-producing
regions. Energy Convers Manag 2022;254:115245. https://
doi.org/10.1016/J.ENCONMAN.2022.115245.
[3] Ashok J, Dewangan N, Das S, Hongmanorom P, Wai MH,
Tomishige K, et al. Recent progress in the development of
catalysts for steam reforming of biomass tar model reaction.
Fuel Process Technol 2020;199:106252. https://doi.org/
10.1016/j.fuproc.2019.106252.
[4] Chen D, Liu C, Mao Y, Wang W, Li T. Efficient hydrogen
production from ethanol steam reforming over layer-
controlled graphene-encapsulated Ni catalysts. J Clean Prod
2020;252. https://doi.org/10.1016/j.jclepro.2019.119907.
[5] Chen S, Pei C, Gong J. Insights into interface engineering in
steam reforming reactions for hydrogen production. Energy
Environ Sci 2019;12:3473e95. https://doi.org/10.1039/
c9ee02808k.
[6] Chen M, Li X, Wang Y, Wang C, Liang T, Zhang H, et al.
Hydrogen generation by steam reforming of tar model
compounds using lanthanum modified Ni/sepiolite catalysts.
Energy Convers Manag 2019;184:315e26. https://doi.org/
10.1016/J.ENCONMAN.2019.01.066.
[7] Khademi MH, Lotfi-Varnoosfaderani M, Moghadasin MH.
Catalytic conversion of bio-renewable glycerol to pure
hydrogen and syngas: energy management and mitigation of
environmental pollution. Energy Convers Manag
2021;247:114719. https://doi.org/10.1016/
J.ENCONMAN.2021.114719.
[8] yong Zhao X, ping Xue Y, feng Yan C, da Wang Z, qing Guo C,
lin Huang S. Sorbent assisted catalyst of Ni-CaO-La2O3 for
sorption enhanced steam reforming of bio-oil with acetic
acid as the model compound. Chem. Eng. Process. Process
Intensif. 2017;119:106e12. https://doi.org/10.1016/
j.cep.2017.05.012.
[9] Hu R, Li D, Xue H, Zhang N, Liu Z, Liu Z. Hydrogen production
by sorption-enhanced steam reforming of acetic acid over
Ni/CexZr1xO2-CaO catalysts. Int J Hydrogen Energy
2017;42:7786e97. https://doi.org/10.1016/
j.ijhydene.2017.01.167.
[10] Di Michele A, Dell'Angelo A, Tripodi A, Bahadori E, Sa nchez F,
Motta D, et al. Steam reforming of ethanol over Ni/MgAl2O4
catalysts. Int J Hydrogen Energy 2019;44:952e64. https://
doi.org/10.1016/j.ijhydene.2018.11.048.
[11] Bepari S, Basu S, Pradhan NC, Dalai AK. Steam reforming of
ethanol over cerium-promoted Ni-Mg-Al hydrotalcite
30280 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 0 2 6 3 e3 0 2 8 1
catalysts. Catal Today 2017;291:47e57. https://doi.org/
10.1016/j.cattod.2017.01.027.
[12] Vanga G, Gattia DM, Stendardo S, Scaccia S. Novel synthesis
of combined CaO-Ca 12 Al 14 O 33 -Ni sorbent-catalyst
material for sorption enhanced steam reforming processes.
Ceram Int 2019;45:7594e605. https://doi.org/10.1016/
j.ceramint.2019.01.054.
[13] Capa A, Garcı́a R, Chen D, Rubiera F, Pevida C, Gil MV. On the
effect of biogas composition on the H2 production by
sorption enhanced steam reforming (SESR). Renew Energy
2020;160:575e83. https://doi.org/10.1016/
j.renene.2020.06.122.
[14] Huang WJ, Yu CT, Sheu WJ, Chen YC. The effect of non-
uniform temperature on the sorption-enhanced steam
methane reforming in a tubular fixed-bed reactor. Int J
Hydrogen Energy 2021;46:16522e33. https://doi.org/10.1016/
J.IJHYDENE.2020.07.182.
[15] Aceves Olivas DY, Baray Guerrero MR, Escobedo Bretado MA,
Marques Da Silva Paula M, Salinas Gutie rrez J, Guzman
Velderrain V, et al. Enhanced ethanol steam reforming by
CO2 absorption using CaO, CaO*MgO or Na2ZrO3. Int J
Hydrogen Energy 2014;39:16595e607. https://doi.org/10.1016/
j.ijhydene.2014.04.156.
[16] Cortazar M, Sun S, Wu C, Santamaria L, Olazar L,
Fernandez E, et al. Sorption enhanced ethanol steam
reforming on a bifunctional Ni/CaO catalyst for H2
production. J Environ Chem Eng 2021;9:106725. https://
doi.org/10.1016/j.jece.2021.106725.
[17] Xu Y, Lu B, Luo C, Chen J, Zhang Z, Zhang L. Sorption
enhanced steam reforming of ethanol over Ni-based catalyst
coupling with high-performance CaO pellets. Chem Eng J
2021;406:126903. https://doi.org/10.1016/j.cej.2020.126903.
[18] Zeng P, Dou B, Zhang H, Wu K, Zhao L, Luo C, et al. Chemical
looping steam reforming of ethanol without and with in-situ
CO2 capture. Int J Hydrogen Energy 2022;47:6552e68. https://
doi.org/10.1016/j.ijhydene.2021.12.066.
[19] Pecharaumporn P, Wongsakulphasatch S, Glinrun T,
Maneedaeng A, Hassan Z, Assabumrungrat S. Synthetic CaO-
based sorbent for high-temperature CO2 capture in sorption-
enhanced hydrogen production. Int J Hydrogen Energy
2019;44:20663e77. https://doi.org/10.1016/
j.ijhydene.2018.06.153.
[20] Azimi B, Tahmasebpoor M, Sanchez-Jimenez PE, Perejon A,
Valverde JM. Multicycle CO2 capture activity and
fluidizability of Al-based synthesized CaO sorbents. Chem
Eng J 2019;358:679e90. https://doi.org/10.1016/
j.cej.2018.10.061.
[21] Cesa  rio MR, Barros BS, Courson C, Melo DMA,
Kiennemann A. Catalytic performances of Ni-CaO-mayenite
in CO2 sorption enhanced steam methane reforming. Fuel
Process Technol 2015;131:247e53. https://doi.org/10.1016/
j.fuproc.2014.11.028.
[22] Dang C, Yu H, Wang H, Peng F, Yang Y. A bi-functional Co-
CaO-Ca12Al14O33 catalyst for sorption-enhanced steam
reforming of glycerol to high-purity hydrogen. Chem Eng J
2016;286:329e38. https://doi.org/10.1016/j.cej.2015.10.073.
[23] Dang C, Liu L, Yang G, Cai W, Long J, Yu H. Mg-promoted Ni-
CaO microsphere as bi-functional catalyst for hydrogen
production from sorption-enhanced steam reforming of
glycerol. Chem Eng J 2020;383:123204. https://doi.org/10.1016/
j.cej.2019.123204.
[24] Sang S, Zhao ZJ, Tian H, Sun Z, Li H, Assabumrungrat S, et al.
Promotional role of MgO on sorption-enhanced steam
reforming of ethanol over Ni/CaO catalysts. AIChE J 2019.
https://doi.org/10.1002/aic.16877.
[25] Wang X, He Y, Xu T, Xiao B, Liu S, Hu Z, et al. CO2 sorption-
enhanced steam reforming of phenol using NieM/
CaOeCa12Al14O33 (M ¼ Cu, Co, and Ce) as catalytic sorbents.
Chem Eng J 2020;393:124769. https://doi.org/10.1016/
j.cej.2020.124769.
[26] Zhao M, Shi J, Zhong X, Tian S, Blamey J, Jiang J, et al. A novel
calcium looping absorbent incorporated with polymorphic
spacers for hydrogen production and CO2 capture. Energy
Environ Sci 2014;7:3291e5. https://doi.org/10.1039/
c4ee01281j.
[27] Nimmas T, Jamrunroj P, Wongsakulphasatch S,
Kiatkittipong W, Laosiripojana N, Gong J, et al. Influence of
CaO precursor on CO2 capture performance and sorption-
enhanced steam ethanol reforming. Int J Hydrogen Energy
2019;44:20649e62. https://doi.org/10.1016/
j.ijhydene.2018.07.095.
[28] Chen CH, Yu CT. Influence of regeneration conditions on
cyclic sorption-enhanced steam reforming of ethanol in
fixed-bed reactor. Int J Hydrogen Energy 2022;47:2412e22.
https://doi.org/10.1016/j.ijhydene.2021.10.192.
[29] Tian S, Yang X, Chen X, Li G, Aikelaimu A, Meng Y, et al.
Catalytic calcium-looping reforming of biogas: a novel
strategy to produce syngas with improved H2/CO molar
ratios. J Clean Prod 2020;270. https://doi.org/10.1016/
j.jclepro.2020.122504.
[30] Xu Y, Luo C, Zheng Y, Ding H, Wang Q, Shen Q, et al.
Characteristics and performance of CaO-based high
temperature CO2 sorbents derived from a sol-gel process
with different supports. RSC Adv 2016;6:79285e96. https://
doi.org/10.1039/c6ra15785h.
[31] Di Giuliano A, Gallucci K, Foscolo PU, Courson C. Effect of Ni
precursor salts on Ni-mayenite catalysts for steam methane
reforming and on Ni-CaO-mayenite materials for sorption
enhanced steam methane reforming. Int J Hydrogen Energy
2019;44:6461e80. https://doi.org/10.1016/
j.ijhydene.2019.01.131.
[32] Wu G, Zhang C, Li S, Huang Z, Yan S, Wang S, et al. Sorption
enhanced steam reforming of ethanol on Ni-CaO-Al2O 3
multifunctional catalysts derived from hydrotalcite-like
compounds. Energy Environ Sci 2012;5:8942e9. https://
doi.org/10.1039/c2ee21995f.
[33] Liu L, Hong D, Wang N, Guo X. High purity H2 production
from sorption enhanced bio-ethanol reforming via sol-gel-
derived NieCaOeAl2O3 bi-functional materials. Int J
Hydrogen Energy 2020;45:34449e60. https://doi.org/10.1016/
j.ijhydene.2020.02.158.
[34] Passos AR, Pulcinelli SH, Santilli CV, Briois V. Operando
monitoring of metal sites and coke evolution during non-
oxidative and oxidative ethanol steam reforming over Ni and
NiCu ex-hydrotalcite catalysts. Catal Today 2019;336:122e30.
https://doi.org/10.1016/j.cattod.2018.12.054.
[35] Loganathan S, Tikmani M, Mishra A, Ghoshal AK. Amine
tethered pore-expanded MCM-41 for CO2 capture:
experimental, isotherm and kinetic modeling studies. Chem
Eng J 2016;303:89e99. https://doi.org/10.1016/
j.cej.2016.05.106.
[36] Qi Z, Daying H, Yang L, Qian Y, Zibin Z. Analysis of CO2
sorption/desorption kinetic behaviors and reaction
mechanisms on Li4SiO4. AIChE J 2013;59:901e11. https://
doi.org/10.1002/aic.13861.
[37] Guo H, Feng J, Zhao Y, Wang S, Ma X. Effect of micro-
structure and oxygen vacancy on the stability of (Zr-Ce)-
additive CaO-based sorbent in CO2 adsorption. J CO2 Util
2017;19:165e76. https://doi.org/10.1016/j.jcou.2017.03.015.
[38] Vieira LM, Lima Santos DB, Hori CE. Bifunctional NieCaO
catalysts modified with CeO2 and inert Ca12Al14O33 for
sorption-enhanced steam reforming of ethanol. Int J
Hydrogen Energy 2022;47:38556e70. https://doi.org/10.1016/
j.ijhydene.2022.09.037.
[39] Yoon HJ, Mun S, Lee KB. Facile reactivation of used CaO-
based CO2 sorbent via physical treatment: critical
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 0 2 6 3 e3 0 2 8 1
relationship between particle size and CO2 sorption
performance. Chem Eng J 2021;408:127234. https://doi.org/
10.1016/j.cej.2020.127234.
[40] Abd El-Hafiz DR, Ebiad MA, Elsalamony RA, Mohamed LS.
Highly stable nano Ce-La catalyst for hydrogen production
from bio-ethanol. RSC Adv 2015;5:4292e303. https://doi.org/
10.1039/c4ra12679c.
[41] Dewoolkar KD, Vaidya PD. Tailored hydrotalcite-based
hybrid materials for hydrogen production via sorption-
enhanced steam reforming of ethanol. Int J Hydrogen Energy
2016;41:6094e106. https://doi.org/10.1016/
j.ijhydene.2015.10.034.
[42] Choong CKS, Huang L, Zhong Z, Lin J, Hong L, Chen L. Effect
of calcium addition on catalytic ethanol steam reforming of
Ni/Al2O3: II. Acidity/basicity, water adsorption and catalytic
activity. Appl Catal Gen 2011;407:155e62. https://doi.org/
10.1016/j.apcata.2011.08.038.
[43] Gao K, Shokrollahi Yancheshmeh M, Duchesne J, Iliuta MC.
Valorization of coal fly ash as a stabilizer for the
30281
development of Ni/CaO-based bifunctional material. ACS
Sustainable Chem Eng 2020;8:3885e95. https://doi.org/
10.1021/acssuschemeng.9b07360.
[44] Ohenoja K, Rissanen J, Kinnunen P, Illikainen M. Direct
carbonation of peat-wood fly ash for carbon capture and
utilization in construction application. J CO2 Util 2020;40.
https://doi.org/10.1016/j.jcou.2020.101203. 0e7.
[45] Li W, Zhao D. An overview of the synthesis of ordered
mesoporous materials. Chem Commun 2013;49:943e6.
https://doi.org/10.1039/c2cc36964h.
[46] Alvarez D, Abanades JC. Pore-size and shape effects on the
recarbonation performance of calcium oxide submitted to
repeated calcination/recarbonation cycles. Energy Fuel
2005;19:270e8. https://doi.org/10.1021/ef049864m.
[47] Wang C, Dou B, Jiang B, Song Y, Du B, Zhang C, et al.
Sorption-enhanced steam reforming of glycerol on Ni-based
multifunctional catalysts. Int J Hydrogen Energy
2015;40:7037e44. https://doi.org/10.1016/
j.ijhydene.2015.04.023.