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 1 ( 2 0 1 6 ) 2 2 7 3 2 e2 2 7 4 2

Available online at www.sciencedirect.com

ScienceDirect

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

Steam reforming of glycerol to hydrogen over ceria

promoted nickelealumina catalysts

Hundessa Dessalegn Demsash, Ratan Mohan*

Department of Chemical Engineering, Indian Institute of Technology Delhi, Hauz Khas, New Delhi 110016, India

article info abstract

Article history: Increased production of biodiesel in recent years has resulted in a surfeit of crude glycerol
Received 26 February 2016 as by-product. This oversupply has triggered efforts to valorize glycerol to clean fuels such
Received in revised form as hydrogen. In the present work, five kinds of nickelealumina and ceria promoted nickel
28 August 2016 ealumina catalysts of different nickel and ceria loadings were prepared by wet impreg-
Accepted 15 October 2016 nation method and tested for conversion of 30 wt.% glycerol in water. Characterization of
Available online 10 November 2016 fresh and spent catalysts were carried out with BET surface area, X-ray diffraction and
scanning electron microscopy techniques. The reaction was conducted at temperatures
Keywords: from 550
C to 800
C, atmospheric pressure and weight hourly space velocity of 6.8e17 h1.
Glycerol steam reforming Thermodynamic simulations were also made for theoretical estimates of conversion and
Hydrogen selectivity and for deciding suitable reaction conditions for experiments. Among the tested
Ceriaenickelealumina catalyst catalysts, 10Ni/Al2O3/5CeO2 was found to be the most efficient and stable for 16 h time on
Kinetic model stream. Complete conversion of glycerol and hydrogen selectivity of 85.7% was obtained
over this catalyst at 650
C, against 95.9% thermodynamic value. Kinetic study was also
done for this catalyst to get the reaction rate equation. Using power law model, the order of
reaction and activation energy were determined to be 0.45 and 62.91 kJ/mol respectively.
© 2016 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.

sustainable [4,5]. Hydrogen is a clean and efficient fuel.

Introduction
Energy supply security, pollution reduction, and global
warming mitigation are strong arguments for reduced reli-
ance on fossil fuels use. With CO2 levels in air in excess
400 ppm [1], substitute fuels are also needed for sustainability
and climatic balance. In this context, biodiesel is an environ-
mentally friendly, biodegradable fuel [2]. However, for
100 tons of biodiesel produced, 10 tons of glycerol is produced
as byproduct. The current glycerol production has already
exceeded market demand and given rise to disposal issues [3].
Hence glycerol valorization, to fuels like hydrogen, can lead to
reduction in production costs of biodiesel and make it
Industrially, it is produced nearly 90% from methane and 5%
from other hydrocarbons [6]. Therefore, its production from
renewable sources such as glycerol is a prudent option.
Glycerol can be converted to hydrogen through a variety of
reforming processes such as steam reforming, aqueous phase
reforming, partial oxidation, autothermal reforming and dry
reforming. It can also be converted to hydrogen by supercrit-
ical water gasification and pyrolysis. Hydrogen production
from glycerol has been reviewed by Adhikari et al. [7], Lin [5],
Dou et al. [8], Silva et al. [9] and Schwengber et al. [10]. Among
the reforming techniques, glycerol steam reforming has the
merit that it is a mature technique, it can be conveniently
conducted at atmospheric pressure and its operation does not

* Corresponding author.
E-mail address: ratan@chemical.iitd.ac.in (R. Mohan).
http://dx.doi.org/10.1016/j.ijhydene.2016.10.082
0360-3199/© 2016 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
 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 1 ( 2 0 1 6 ) 2 2 7 3 2 e2 2 7 4 2
involve sophisticated control. However, the need for an effi-
cient commercial catalyst for this process is still there [11].
In this regard, several catalysts with different support and
modifiers have been studied under different reaction condi-
tions. Both noble and transition metal catalysts have been
used in glycerol steam reforming studies. Hirai et al. [12],
Adhikari et al. [6], Zhang et al. [13], Doukkali et al. [14] and Wei
et al. [15] have studied noble metal catalysts like ruthenium,
rhodium, palladium, platinum and iridium in their works,
amongst others. On the other hand, Adhikari et al. [16], Iriondo
et al. [17], Cheng et al. [18], Dave and Pant [19], Sanchez and
Comelli [20,21] and Lin et al. [22] have used transition metal
catalysts such as nickel, copper and doped nickel with copper
in their studies. In recent studies, use of Niealumina in
combination with other metals or oxides such as Cu, Al, Co,
Sn, Pd, K, Ca, Sr, Mg(II), MgO, B2O3, La2O3, as well as use of
CeO2, La2O3, ZrO2, SiO2 or their mix as support modifiers have
been investigated by Ebsish et al. [23], Dieuzeide et al. [24],
Seung-hoon et al. [25], Dou et al. [26], Yurdakul et al. [27],
Bobadilla et al. [28], Papageridis et al. [29] and Zamzuri et al.
[30].
In most of the studies, the preferred choice of catalyst
support is g-alumina due to its large surface area and thermal
stability. In some cases, ceria is also used in addition to
alumina. The noble metal catalysts demonstrate excellent
catalytic performance, including suppression of carbon for-
mation on the catalyst surface that results in higher catalyst
life and activity. However, the main difficulty with these cat-
alysts is the cost. As such, cheaper metallic catalysts such as
nickel (in some cases copper) that show good catalytic activity
are the subject of much research for glycerol steam reforming.
Among the transition metal catalysts, nickel is the most
studied active metal. Nickel has good activity for CeC, OeH,
and CeH bond cleavage. It also catalyzes the wateregas-shift
reaction involved in the steam reforming. Also, it is less
expensive and more abundant than noble metals. However, Ni
catalysts get deactivated during the steam reforming of
oxygenated hydrocarbons due to coke deposition. How to
subdue coke deposition to augment the catalytic stability is
still a major concern. One of the ways to improve Niealumina
catalyst properties and resistance to carbon/coke deposition is
by modifying the support by rare earth oxides e.g. CeO2. Ceria
improves the catalytic activity of Ni/Al2O3 catalyst. Moreover,
its basic character gives better redox delivery and storage of
lattice oxygen, by which water dissociation, methane
reforming, coke reduction, and water gas shift reaction can be
promoted.
Various authors have investigated different aspects of
ceria promoted nickelealumina catalysts. Iriondo et al. [17]
prepared and utilized Ni/Al2O3/CeO2 catalyst with different
ceria loadings at temperatures of 500e600
C and a WHSV of
7.7 h1. They reported complete glycerol conversion and the
maximum hydrogen yield of 80.3%. However, stability was
observed up to 8 h on stream. Lin et al. [22] developed and
tested 15Ni/10CeO2/Al2O3 catalyst in a packed bed and PdAg/
PSS membrane reactor for a range of steam to glycerol molar
ratios (3:1e9:1), WHSV (1.5e5 h1) and temperatures
(450e800
C). Glycerol conversion of 96.7% and highest
hydrogen yield of 5.82H2 moles per mole of glycerol were re-
ported. Sanchez and Comelli [24] studied the promoter effect
22733
of Co on Ni/Al2O3 in the temperature range 300e500
C, steam
to glycerol molar ratio of 6:1 and WHSV 10 h1. They reported
hydrogen yield of 83.6% and stability of the catalyst for 8 h on
stream. Profeti et al. [31] also developed and evaluated Ni/
CeO2eAl2O3 catalysts doped with the noble metals (platinum,
palladium, ruthenium, iridium) at glycerol to water molar
ratio of 1:6, a flow rate of 2.5 cm3/h and temperature of 700
C.
The gaseous product distribution was reported to be similar
for all catalysts, 55e60% H2, 20% CO2, 15e25% CO and 25% CH4.
Thus it is seen that several studies have been made on the
Niealumina/ceria catalysts and their performance. However,
in view of its potential as an effective and less costly catalyst,
there is need for further investigation on its activity and sta-
bility over different Ni, ceria content and extended range of
reaction conditions. In the present work therefore experi-
mental measurement of catalyst activity has been made for
different Ni, ceria loadings. A thermodynamic simulation
analysis has been made to indicate suitable conditions for
study. Stability is also examined over a longer time period.
Moreover, from engineering point of view kinetic model and
rate equation are needed for process calculations. A higher
weight hourly space velocity (WHSV) range has been studied
to keep mass transfer effects insignificant and get the kinetic
rate equation, not often reported in literature.
Materials and methods
Experimental
Catalyst preparation
The nickelealumina catalysts used in this study are of 5, 10
and 15 percent nickel loading by weight over g-alumina and
are represented as 5Ni/Al2O3, 10Ni/Al2O3, and 15Ni/Al2O3.
These were prepared by wet impregnation techniques [32]. An
aqueous solution of nickel nitrate hexahydrate
(Ni(NO3)2$6H2O, 98% from SigmaeAldrich) was used as pre-
cursor of nickel for impregnation onto commercial g-Al2O3
support (SigmaeAldrich). Likewise, CeO2 promoted nickel-
ealumina catalysts with 5 and 10 percent loading of ceria by
weight (10Ni/Al2O3/5CeO2 and 10Ni/Al2O3/10CeO2) were pre-
pared by a two-step procedure. First same as above for Ni/
Al2O3, then impregnation of aqueous solution of Ce(N-
O3)3$6H2O (Jonhson Mathew Chemicals, India). After impreg-
nation, the catalysts were dried for 8 h at 125
C followed by
calcination at 550
C for 5 h. Then the catalysts were crushed
and sieved to 95e130 mm size, for use in the reaction evalua-
tion study.
Catalyst characterization
The surface areas and pore volumes of the prepared catalysts
were determined by ASAP 2010 Micromeritics surface area
analyzer using adsorption with nitrogen at 196.15
C. X-Ray
diffraction (XRD) analysis of the calcined and powdered
samples were carried out to examine crystal structure of
materials with Rigaku Miniflex 600 diffractometer. Measure-
ments were made using Cu-Ka radiation operating at 30 kV
voltage, 40 mA current, 2
/min scanning rate and 1.54 Å. The
2q diffraction angle ranged from 20
to 80
. For surface
morphological study of catalysts, Scanning Electron
22734 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 1 ( 2 0 1 6 ) 2 2 7 3 2 e2 2 7 4 2
Microscope (ZEISS EVO 50 SEM) was employed. Temperature
programmed oxidation (TPO) was carried out with TGA/DTA/
DSC (TA-Q600) instrument.
Experimental set-up
The catalytic steam reforming experiments were conducted in
a fixed-bed quartz tubular reactor. Fig. 1 shows the schematic
diagram of the experimental set-up. The quartz reactor tube is
10 mm i.d. and 500 mm long, but initial 150 mm is the pre-
heating zone while the next 250 mm is the main heating
zone with the catalyst bed in it. The actual catalyst bed is
much smaller, 8e16 mm high, depending the loading (1 or 2 g
catalyst, 95 mme130 mm pellet size). It is supported by ceramic
wool from above and below to minimize sintering. An electric
furnace encloses the tube and supplies heat to both the zones.
It has a PID controller and an electronic temperature indicator
for accurate temperature measurement. 30 wt.% glycerol
aqueous solution was prepared by mixing deionized distilled
water (ultrapure and reagent grade, Organo Biotech Labs,
India) with glycerol (99.5%, SigmaeAldrich). The solution is
fed to the reactor by a syringe pump (AKAS Terumax, India).
Argon (GC grade, Sigma Gases, India) is used as the carrier gas
at a flow rate of 30 ml/min. Prior to reaction, the packed
catalyst was pre-reduced insitu in 10% H2/argon at 550
C for
2 h and then cooled down in argon gas flow. After steady state
Fig. 1 e Laboratory set-up for glycerol steam reforming.
is obtained, reaction products are analyzed to evaluate cata-
lytic activity. Products leaving the reactor are passed through
a condenser which is supplied with ice-cold water as cooling
medium. Liquid and gaseous products are separated by a
separating funnel. Non-condensable products and conden-
sates were analyzed by a gas chromatograph (NUCON Series
5700), equipped with a thermal conductivity detector (TCD)
and a flame ionization detector (FID). The GC has two col-
umns, a 2 m SS column packed with PORAPAKQ 80/100 mesh
and a 30 m fused silica capillary column.
Catalytic evaluations
Glycerol steam reforming experiments were carried out using
the above experimental set-up. Catalyst screening tests were
done by using the optimal conditions indicated by thermo-
dynamic analysis. The thermodynamic analysis is described
in more detail later. Reaction conditions studied were tem-
peratures 550
Ce800
C, atmospheric pressure and WHSV of
6.8 h1 to 17 h1. Subsequently the best performing catalyst
was employed for characterization, stability and kinetic
studies as discussed in Sections 'Results and discussion' and
'Reaction kinetic modelling'.
Glycerol conversion calculation for the experiments is
based on carbon molar outflow of the gas products on a dry
basis and the glycerol molar flow rate of feed to the reactor.
 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 1 ( 2 0 1 6 ) 2 2 7 3 2 e2 2 7 4 2 22735

During GC analysis and thermodynamic simulation, the only

70
significant dry gases observed are H2, CO2, CO and CH4 evolved
from the reactor along with argon carrier gas. Hence, reaction 60

Molar Composition ( % )
evaluation is based on these dry gas compositions as shown CH4
below. Molar flow rate of the dry gas was obtained from 50 CO
component mass balance. Conversion and selectivity calcu- CO2
lation formulae are as given below (Adhikari et al. [16], Dave 40 H2
and Pant [19]).
30
Fglycerol; in  Fglycerol; out
Glycerol conversion ð%Þ ¼  100 (1) 20
Fglycerol; in

10
FCH4; out
SCH4 ð%Þ ¼  100 (2)
FCH4; out þ FCO2; out þ FCO; out 0

FCO2; out 400 600 800 1000 1200

SCO2 ð%Þ ¼  100 (3)
FCH4; out þ FCO2; out þ FCO; out Temperature ( oC )

FCO; out Fig. 2 e Product composition versus temperature (Other

SCO ð%Þ ¼  100
FCH4; out þ FCO2; out þ FCO; out
H2 moles produced
SH2 ð%Þ ¼
Carbon atoms in gaseous products
1 concentration. The concentrations of methane and carbon
  100
Reforming ratio
Results and discussion
Thermodynamic simulation
It is well-known that steam reforming of glycerol is strongly
endothermic and ideally it must be carried out at high tem-
peratures, low pressure, and high steam to glycerol ratio to
achieve high conversion. However, what set of reaction con-
ditions favor maximum yield can be indicated by thermody-
namic analysis [33e36]. Then appropriate selection of reaction
conditions can be made for experiment. In this work, ther-
modynamic analysis of glycerol steam reforming was done
using Aspen plus™ simulator. The procedure available in the
subroutine RGIBBS was used to determine the thermody-
namic equilibrium compositions of each species present.
SoaveeRedlicheKwong model was chosen as the equation of
state to correct the ideal Gibbs energy values. The effect of
operating parameters such as temperature and steam-to-
(4) simulation parameters: atmospheric pressure, 30 wt.%
glycerol feed).
(5) dioxide were found to be decreasing with increasing temper-
ature. This is an implication of reverse water gas shift reaction
and carbon dioxide methanation which are favored at high
temperatures, consuming the hydrogen produced. Thus it is
seen that increasing temperature has positive effect on till
650
C but negative effect then onwards. The temperature
sensitivity analysis shows that carrying out the reforming
experiments between 550
C and 800
C is appropriate.
Glycerol to steam ratio
To see the effect of variation of steam to glycerol feed ratio on
the product compositions, feed ratio was varied from 10 wt.%
glycerol to 90 wt.%. The results are shown in Fig. 3. Lower
glycerol weight percent than 10% was not considered because
this would imply, in practice, high cost for water vaporization;
0.3
70
60 H2
Molar Composition (%)
Molar Composition (%)

glycerol feed ratio on hydrogen production and product CO

50 CO2 0.2
composition were evaluated.
Glycerol
40
CH4
Effect of temperature
30
The effect of temperature variation on product composition is
0.1
shown in Fig. 2. During this temperature sensitivity analysis, 20
other parameters were held constant (30 wt.% glycerol feed
10
ratio and 1 atm pressure). As seen from the figure, the amount
of hydrogen produced appreciably increases with the increase 0
0.0
of temperature from 300
C to 650
C. This is primarily 20 40 60 80
attributed to the thermal decomposition of glycerol and steam
Glycerol in the Feed (Wt.%)
reforming of methane which are the main reactions of glyc-
erol steam reforming producing hydrogen under strongly Fig. 3 e The effect of variation of glycerol to water feed ratio
endothermic condition. Further increase of temperature from on product composition (Other simulation parameters:
650
C to 1200
C results in a slight reduction of hydrogen atmospheric pressure, temperature 650
C).
22736 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 1 ( 2 0 1 6 ) 2 2 7 3 2 e2 2 7 4 2
otherwise for lower concentration of glycerol in the feed the
hydrogen produced is higher.
In the glycerol feed range from 10 to 30 wt.%, product mole
percent of hydrogen does not change much. Beyond 30 wt.%
hydrogen concentration decreases more. CO2 also shows a
near uniform decrease. On the other hand CO, CH4, and even
unreacted glycerol after 50 wt.%, increase. Hence, it was
decided to carry out the reforming experiments at 30 wt.%
glycerol in the feed.
Catalyst characterization
Surface area, pore size
Table 1 shows the results of the BET (Brunauer, Emmett and
Teller) surface area analysis of the support and the promoter.
Impregnation of Ni and CeO2 as a promoter on g-Al2O3 de-
creases the surface area and pore volume of the catalysts. This
is due to the blockage of the pores during impregnation by
active metal and modifier. Despite the reduction, the available
surface area is still sufficient for catalysis, in agreement with
literature [17,32] and showing the adequacy of the preparation
method.
The X-ray diffraction (XRD) pattern
The XRD patterns of fresh and spent nickelealumina and ceria
promoted nickelealumina (10Ni/Al2O3 and 10Ni/Al2O3/5CeO2)
catalysts are shown in Fig. 4. XRD characterization as well as
the SEM, TPO and stability tests, were done for these two
catalysts only as 10Ni/Al2O3/5CeO2 was found to be the best
and 10Ni/Al2O3 is its non-ceria version. Identification of the
best catalyst is discussed later in Section 'Performance
evaluation of catalysts'. Peak matching is carried out using
X'pert HighScore Plus software for each species. It is observed
that all XRD patterns show several sharp peaks, indicative of
the crystallite structure formations in the calcined catalysts.
In all samples, well defined reflections at 2q angles of 32.0
,
37.4
, 45.9
, 61.7
, and 66.7
were observed and can be ascribed
to the crystalline Al2O3 with a cubic structure (Powder
Diffraction File (PDF): 01-074-2206). The crystalline phases of
NiO were observed at 37.2
, 43.3
, 62.9
and 75.6
(PDF: 01-078-
0643) for both 10Ni/Al2O3 and 10Ni/Al2O3/5CeO2 fresh and
spent catalysts, as can be seen from Fig. 4aed [37]. There is
weak appearance of peaks of NiO in diffractograms of ceria
promoted catalyst suggesting fine dispersion. No CeO2 peak
observed in ceria promoted catalysts is attributed to better
dispersion of ceria over the alumina support due to the low
concentration of ceria. The stoichiometric nickel aluminate
crystallites (NiAl2O4) were observed at 32
, 37
, 45
and 66

[24,28]. Since it has spinel structure (similar to gamma
Table 1 e BET surface area, pore volume, and pore
diameter.
Catalysts BET surface Pore volume Average pore
area (m2/g) (cc/g) diameter (Å)
10Ni/Al2O3/5CeO2 112 0.50 64
10Ni/Al2O3 122 0.56 64
10Ni/Al2O3/10CeO2 104 0.49 66
g-Al2O3 200 0.6 70
alumina), the presence of these crystallites are good for
catalysis. In the spent 10Ni/Al2O3, coke peak was observed at
26
indicating that there is more coking tendency with the
nickel over alumina support compared to ceria promoted
nickelealumina. It may be noted that crystalline Ni with
characteristic reflections at 2q angles of 44.2
, 51.9
, and 76.1

(PDF: 010870712) and NiO were detected in the spent catalysts
but Ni species was not detected in the fresh catalysts [38]. This
was because characterizations of fresh catalysts were carried
out without reduction while the spent catalysts were reduced
insitu with hydrogen before testing, in which case NiO gets
converted to Ni metal.
Scanning electron microscopy (SEM) images
Further characterization of fresh and spent 10Ni/Al2O3/5CeO2
and 10Ni/Al2O3 catalysts were done by Scanning Electron Mi-
croscopy for morphological investigation at 10,000 magnifi-
cation. The results are shown in Fig. 5. Fig. 5a shows fresh
ceria promoted nickelealumina while Fig. 5b shows the spent
one of the same. Fig. 5c and d shows the un-promoted nick-
elealumina, fresh and spent catalysts respectively. Spent
10Ni/Al2O3 (Fig. 5d) shows notable morphological change. The
hair-like structure observed is due to strong coke/carbon
deposition on the surface of spent catalyst. But the ceria
promoted spent catalyst (10Ni/Al2O3/5CeO2) showed no sig-
nificant morphological change. This suggests that the ceria
modified nickelealumina has better anti-coking attribute
than the bare nickelealumina catalyst. The anti-coking attri-
bute of 10Ni/Al2O3/5CeO2 catalyst is further confirmed by
temperature programmed oxidation experiments.
Temperature programmed oxidation
Carbon deposited on the surface of the spent catalysts was
measured by temperature programmed oxidation experi-
ments on a thermogravimetric analysis (TGA) instrument by
derivative weight loss. The percentage weight loss between
400
C and 800
C is shown in Fig. 6. As can be seen in the
figure, the deposited carbon began to be oxidized at 510
C
onwards. The maximum carbon oxidation rate occurred at
around 630
C and 680
C for 10Ni/Al2O3 and 10Ni/Al2O3/5CeO2
respectively.
The TPO profiles were analysed by Gaussian method of
peak fitting and the fitted curves were de-convoluted to get
individual peaks as shown in Fig. 6. There is a relatively weak
band at lower temperatures (567, 586
C) in Fig. 6a for spent Ni/
Al2O3 catalyst and a major peak at 627
C. The peaks at low
temperature are attributed to whisker like coke deposited on
the catalyst, same also observed in SEM diagram, Fig. 5d for
spent 10Ni/Al2O3. High temperature peak at 627
C is attrib-
uted to pyrolytic coke, indicative of monatomic and polymeric
coke adsorbed on the Nickel metal. This kind of coke encap-
sulates the active nickel site and blocks the accessibility to
further reaction. These carbonaceous deposits decrease the
ability to convert glycerol to H2 and increase both CO and CH4
[20,39], seen in Fig. 9b. For CeO2 promoted catalyst, the low
temperature peaks are absent. Again same can be observed in
SEM image (Fig. 5b) showing no apparent morphological
changes. The relative intensity of the peak at 627
C in the
Fig. 6b is very small compared to that of 10Ni/Al2O3 catalyst.
The major peak observed in ceria doped catalyst is at 685
C,
 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 1 ( 2 0 1 6 ) 2 2 7 3 2 e2 2 7 4 2 22737

Fig. 4 e XRD patterns of fresh and spent catalysts: (a) fresh 10Ni/Al2O3; (b) 10Ni/Al2O3/5CeO2; (c) spent 10Ni/Al2O3; (d) spent
10Ni/Al2O3/5CeO2.

Fig. 5 e SEM image of catalysts: (a) fresh 10Ni/Al2O3/5CeO2 (b) spent 10Ni/Al2O3/5CeO2 (c) fresh 10Ni/Al2O3 (d) spent 10Ni/
Al2O3.

indicative of coke formed with different graphitization degree
[40], which does not block the active sites.
Carbon deposition on the two catalysts was determined.
The derivative percentage weight loss during TPO experiment
and the results of carbon balance are compared in Table 2.
Assuming steady state, material balance was done in terms of
carbon atom, using molar flow rates of glycerol feed, gylcerol
out, and the products carbon monoxide, carbon dioxide,
methane and coke in the form of carbon.
The result shows that ceria modified catalyst (10Ni/Al2O3/
5CeO2) suppresses carbon deposition to 0.141 mg C/g cat
against 0.237 mg C/g for Niealumina catalyst. This notable
difference in coking is a clear indication that incorporation of
ceria in nickelealumina catalyst results in reduced coking
tendency.
Performance evaluation of catalysts
In all the experiments, data were taken after one hour at
steady state. The data reported are averaged over three runs
for reproducibility. As already mentioned, 30 wt.% glycerol in
water was used as feed in the experiments. For the catalysts
developed and tested, selectivity of hydrogen and conversion
of glycerol were determined and are presented in Figs. 7 and 8.
Stability tests were done for only two of the catalysts, 10Ni/
Al2O3/5CeO2 (the best) and its without ceria version, 10Ni/
Al2O3. Catalyst activity and stability are discussed.
Effect of nickel loading and ceria as a promoter
Fig. 7 shows the selectivity of hydrogen vs temperature vari-
ation for the five catalysts as well the thermodynamic values.
22738 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 1 ( 2 0 1 6 ) 2 2 7 3 2 e2 2 7 4 2

Fig. 6 e Temperature programmed oxidation of spent catalysts.

to 10% ceria caused a decline. Overall, hydrogen selectivity

Table 2 e Experimental and theoretical coke varied from 49.5% for 5Ni/Al2O3 at 550
C to 85.7% for 10Ni/
determination. Al2O3/5CeO2 at 650
C. The experimental selectivity trend with
Catalysts Coke determination temperature is similar to the thermodynamic variation, as
(mg C/g cat.) also seen in the figure.
TPO Carbon balance Conversion results for the catalysts are shown in Fig. 8.
Conversion ranged from 74.1% (5Ni/Al2O3, 550
C) to 100%
10Ni/Al2O3 0.237 0.241
(10Ni/Al2O3/5CeO2, 550
C). In fact, for the latter catalyst
10Ni/Al2O3/5CeO2 0.141 0.153
complete conversion of glycerol was attained at all tempera-
tures. Among the tested catalysts thus, 10Ni/Al2O3/5CeO2
Selectivity is seen to increase with increase in nickel content showed the highest catalytic activity. Even at a lower tem-
from 5 to 10 percent in Ni/Al2O3. However, further increase to perature of 550
C, this catalyst gave complete conversion but
15%, results in a decline of selectivity. Similarly incorporation a lower selectivity, about 74%. Higher loading of both nickel
of 5% ceria to 10Ni/Al2O3 increased the selectivity, but increase (>10%) and ceria (>5%) affect the performance adversely. This
is an indication of aggregation of crystals that reduces the
dispersion and results in reduced catalytic activity [41].

90
Catalyst stability tests
Fig. 9a and b shows catalyst stability results for 10Ni/Al2O3/
80 5CeO2 and 10Ni/Al2O3 respectively. H2 selectivity and exit
H2 Selectivity (%)

molar compositions of CO, CH4 and CO2 are almost same for
16 h TOS (Time on Stream, i.e. total run-time) for 10Ni/Al2O3/
70
5CeO2. This implies that there is no significant deactivation of
the catalyst over this time. On the other hand, deactivation of
60 10Ni/Al2O3 catalyst was observed after 7e8 h TOS. Improved
resistance to deactivation of the ceria modified nickel-
50 ealumina compared to bare nickelealumina is also reported
5Ni/Al O 15Ni/Al O
10Ni/Al O 10Ni/Al O /10CeO by Iriondo et al. [17] (in their case for 13 wt.% Ni loading) up to
10Ni/Al O /5CeO Thermodynamic Simulation 8 h on stream; in the present case 10Ni/Al2O3/5CeO2 catalyst
40 was found to be stable for 16 h TOS.
550 600 650 700 750 800
Better stability of 10Ni/Al2O3/5CeO2 may be attributed to
Temperature(oC) enhancement of dissociation of steam by ceria
(CeO2x þ H2 O/CeO2 þ xH2 ) and acceleration of the reaction of
Fig. 7 e Thermodynamic and experimental selectivity of H2
steam with adsorbed species on nickel. Ceria is able to donate
for different catalysts (Reaction conditions: atmospheric
its oxygen for the removal of carbon monoxide
pressure, WHSV 12 h¡1, 30 wt.% glycerol).
 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 1 ( 2 0 1 6 ) 2 2 7 3 2 e2 2 7 4 2 22739

110
5Ni/Al O 10Ni/Al O
15Ni/Al O 10Ni/Al O /5CeO
105
Thermodynamic 10Ni/Al O /10CeO

100

Conversion (%)
95

90

85

80

75

70
550 600 650 700 750 800
Temperature(oC)

Fig. 8 e Glycerol conversion versus temperature for different catalysts (Reaction conditions: atmospheric pressure, WHSV
12 h¡1, 30 wt.% glycerol).

Fig. 9 e Product molar composition over different run times: a) 10Ni/Al2O3/5CeO2. b) 10Ni/Al2O3 (Reaction conditions:
Temperature 650
C; WHSV 12 h¡1, 30 wt.% glycerol feed).

(CeO2 þ xCO/CeO2x þ Co2 ) and hydrocarbons such as to acetate, finally to CO2, i.e. inhibiting the formation of CH4.
methane (CeO2 þ Hydrocarbons/CeO2x þ ðCO; CO2 ; H2 O; H2 Þ) in The inhibition effect of CeO2 to form carbonaceous deposits by
the oxygen-deficient atmosphere. Even after loss of oxygen enhancing the hydration of the support and its capacity to
from its lattice and the consequent formation of a large store and release of oxygen is well documented [20,42e45].
number of oxygen vacancies, ceria retains its fluorite crystal Reasons for better activity and stability of ceria promoted
structure (Lin et al. [22]). This also facilitates rapid and com- nickelealumina catalysts have also been discussed earlier
plete refilling of oxygen vacancies upon exposure to steam, with XRD and SEM characterizations and with TPO results.
with recovery of ceria.
The enhanced performance of ceria promoted catalysts is
also explained by steam reforming pathway, through the Table 3 e Conversion versus weight hourly space velocity
synergistic effect of NieCeO2 in ethanol steam reforming, Xu (WHSV).
et al. [42]. They showed that acetate or ethoxy groups form on WHSV (h1) Conversion (%)
the nickel surface, which are precursors to coke formation. On
95 mm pellets 130 mm pellets 130 mm pellets
nickel surface under waterless environment, these groups 1 g loading 1 g loading 2 g loading
favours the formation of carbon in a relative short time, where
6.8 100 99.8 99.9
as in water rich environment, carbon accumulation is very
10.2 88.2 88.0 88.0
slow. When ceria is added as a promoter, Ce(III) sites on ceria 13.6 82.1 82.2 81.8
dissociate the water forming the surface OH groups. In pres- 17.0 65.2 65.0 64.8
ence these OH group, ceria promotes the oxidation of ethoxy
Reaction conditions: Temperature 550
C; TOS ¼ 5 h.
22740 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 1 ( 2 0 1 6 ) 2 2 7 3 2 e2 2 7 4 2

Fig. 10 e a) Glycerol conversion versus Wcat/FAo. b) Reaction constant versus temperature.

rate, power law and the plug flow reactor design equation
Reaction kinetic modelling were used, given in Eqs. (8) and (9).
 
Mass transfer considerations E
rgly ¼ ko exp  Cn (8)
RT gly
To avoid falsified kinetics, the intrusion of mass transfer
resistance has to be excluded [46]. By varying the feed rates, dXgly
rgly ¼   (9)
external mass transfer resistance-free regime was deter-
d FWgly;in
cat:

mined with constant space-time Wcat/FAO (catalyst weight to

molar flow of the reactor feed) and constant catalyst size. The
results are shown in Table 3. The conversion of glycerol
remained constant at different space-times, which indicates
the absence of external diffusional resistance in the range of
the feed flow rates chosen. Additional runs were done to
examine the intraparticle diffusional limitations by varying
the catalyst pellet size.
From Table 3, it may be seen that no significant change in
conversion occurs either with pellet size or with loading
(catalyst bed height). This is implies negligible intraparticle
diffusional resistance, in the range investigated (pellet size
95 mm130 mm).
Reaction order and rate constants
The constants k0 , n and E are frequency factor, order of
reaction and the activation energy respectively. Since excess
steam was used, glycerol is assumed to be the limiting
reactant.
Using differential method, from the slope of the plot of
glycerol conversion versus Wcat/FAo, reaction rates were
computed at three different temperatures (Fig. 10a). Then
Arrhenius plot (Fig. 10b) is drawn and used to determine
activation energy. From regression analysis, reaction constant
and order of reaction were evaluated as shown in Table 4.
Order of reaction with respect to glycerol and activation en-
ergy were determined to be 0.45 and 62.91 kJ/mol respectively
with a coefficient of regression of 0.99.

It is not only important to avoid mass transfer resistance but Conclusions

also necessary to obtain kinetic data in the absence of fluid
channeling and back-mixing. To attain this condition (plug In this work, glycerol steam reforming over nickelealumina
flow), the ratio of the internal diameter of the tube to the and ceria promoted nickelealumina catalysts has been stud-
diameter of the pellet (dt/Dp) should be greater than 10 and ied. Five catalysts of different nickel and ceria loadings (5, 10,
that of bed height to diameter of the pellet (Lbed/Dp) should be 15 wt.% Ni and 5, 10 wt.% ceria) were developed and tested in a
>50 [47]. In the present study, these are 76.9 and 61.5 respec- continuous flow packed bed reactor. Incorporation of ceria as
tively. After absence of mass transfer intrusion and plug flow a promoter was found to give better results at lower loading of
conditions were ensured, a kinetic study was done. For kinetic 5 wt.%; for higher ceria loading of 10 wt.% the results were not
modeling, multiple runs at 550
C, 600
C and 650
C with as good. Likewise 10 wt.% loading of Ni was better. Amongst
different WHSV were conducted. For evaluation of reaction the five catalysts experimented upon, 10Ni/Al2O3/5CeO2 was
seen to give best results in terms of activity and stability.
Highest H2 selectivity of 85.7% at 650
C was obtained with this
Table 4 e Reaction constant and reaction order at catalyst, against 95.9% thermodynamic value. The catalyst
different temperatures.
was observed to be stable for 16 h time on stream. XRD, SEM
T (K) 1/T (1/K) k (kmol1n/ ln k n R2 and temperature programmed oxidation characterizations
(h1n Kg cat.)
were done for the fresh and spent 10Ni/Al2O3/5CeO2 and 10Ni/
923 0.0010834 0.1430 1.9445 0.450 0.99 Al2O3 catalysts, which revealed that 10Ni/Al2O3/5CeO2 showed
873 0.0011455 0.0914 2.3925 0.451 the lowest coking activity. Finally a kinetic study was also
823 0.0012151 0.0529 2.9403 0.449
made for 10Ni/Al2O3/5CeO2 using WHSV of 6.8e17 h1. From
 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 1 ( 2 0 1 6 ) 2 2 7 3 2 e2 2 7 4 2
the data, kinetic model was obtained using power law. Order
of reaction with respect to glycerol and activation energy were
determined to be 0.45 and 62.91 kJ/mol respectively.
Acknowledgements
This work was carried out as part of the Ph.D. research pro-
gram of Mr. Hundessa D. Demsash at the Indian Institute of
Technology Delhi, Department of Chemical Engineering. No
external funding was involved.
references
[1] NASA. Climate change: vital signs of the planet: carbon
dioxide. 2014. http://climate.nasa.gov/vital-signs/carbon-
dioxide/ [Accessed 26 September 2015].
[2] Talebian-Kiakalaieh A, Amin NAS, Hezaveh H. Glycerol for
renewable acrolein production by catalytic dehydration.
Renew Sustain Energy Rev 2014;40:28e59.
[3] Katryniok B, Paul S, Dumeignil F. Recent developments in the
field of catalytic dehydration of glycerol to acrolein. ACS
Catal 2013;3:1819e34.
[4] Ayoub M, Abdullah AZ. Critical review on the current
scenario and significance of crude glycerol resulting from
biodiesel industry towards more sustainable renewable
energy industry. Renew Sustain Energy Rev 2012;16:2671e86.
[5] Lin YC. Catalytic valorization of glycerol to hydrogen and
syngas. Int J Hydrogen Energy 2013;38:2678e700.
[6] Adhikari S, Fernando S, Haryanto A. Production of hydrogen
by steam reforming of glycerin over alumina-supported
metal catalysts. Catal Today 2007;129:355e64.
[7] Adhikari S, Fernando SD, Haryanto A. Hydrogen production
from glycerol: an update. Energy Convers Manag
2009;50:2600e4.
[8] Dou B, Song Y, Wang C, Chen H, Xu Y. Hydrogen production
from catalytic steam reforming of biodiesel byproduct
glycerol: issues and challenges. Renew Sustain Energy Rev
2014;30:950e60.
[9] Silva JM, Soria MA, Madeira LM. Challenges and strategies for
optimization of glycerol steam reforming process. Renew
Sustain Energy Rev 2015;42:1187e213.
[10] Schwengber A, Alves HJ, Schaffner RA, daSilva FA,
Sequinel R, Bach VR, et al. Overview of glycerol reforming for
hydrogen production. Renew Sustain Energy Rev
2016;58:259e66.
[11] Subramani V, Basile A, Veziroglu TN. Compendium of
hydrogen energy. In: Hydrogen production and purification,
vol. 1. Woodhead Publishing; 2015.
[12] Hirai T, Ikenaga NO, Miyake T, Suzuki T. Production of
hydrogen by steam reforming of glycerin on ruthenium
catalyst. Energy Fuels 2005;19:1761e2.
[13] Zhang B, Tang X, Li Y, Xu Y, Shen W. Hydrogen production
from steam reforming of ethanol and glycerol over ceria-
supported metal catalysts. Int J Hydrogen Energy
2007;32:2367e73.
[14] Doukkali ME, Iriondo A, Arias PL, Cambra JF, Gandarias I,
Barrio VL. Bioethanol/glycerol mixture steam reforming over
Pt and PtNi supported on lanthana ceria doped alumina
catalysts. Int J Hydrogen Energy 2012;37:8298e309.
[15] Wei Z, Karim AM, Li Y, King DL, Wang Y. Elucidation of the
roles of Re in steam reforming of glycerol over PteRe/C
catalysts. J Catal 2015;322:49e59.
22741
[16] Adhikari S, Fernando SD, Haryanto A. Kinetics and reactor
modeling of hydrogen production from glycerol via steam
reforming process over Ni/CeO2 catalysts. Chem Eng
Technol 2009;32:541e7.
[17] Iriondo A, Barrio VL, Cambra JF, Arias PL, Guemez MB,
Sanchez-Sanchez MC, et al. Glycerol steam reforming over Ni
catalysts supported on ceria and ceria-promoted alumina.
Int J Hydrogen Energy 2010;35:11622e33.
[18] Cheng CK, Foo SY, Adesina AA. Steam reforming of glycerol
over Ni/Al2O3 catalyst. Catal Today 2011;178:25e33.
[19] Dave CD, Pant KK. Renewable hydrogen generation by steam
reforming of glycerol over zirconia promoted ceria supported
catalyst. Renew Energy 2011;36:3195e202.
[20] Sanchez EA, Comelli RA. Hydrogen by glycerol steam
reforming on a nickelealumina catalyst: deactivation
processes and regeneration. Int J Hydrogen Energy
2012;37:14740e6.
[21] Sanchez EA, Comelli RA. Hydrogen production by glycerol
steam-reforming over nickel and nickelecobalt impregnated
on alumina. Int J Hydrogen Energy 2014;39:8650e5.
[22] Lin K-H, Lin W-H, Hsiao C-H, Chang H-F, Chang AC-C.
Hydrogen production in steam reforming of glycerol by
conventional and membrane reactors. Int J Hydrogen Energy
2012;37:13770e6.
[23] Ebshish A, Yaakob Z, Narayanan B, Bshish A, Wan Daud WR.
Steam reforming of glycerol over Ni supported alumina
xerogel for hydrogen production. Energy Procedia
2012;18:552e9.
[24] Dieuzeide ML, Jobbagy M, Amadeo N. Glycerol steam
reforming over Ni/g-Al2O3 catalysts, modified with Mg(II).
Effect of Mg (II) content. Catal Today 2013;213:50e7.
[25] Seung-hoon K, Jae-suna J, Eun-hyeoka Y, Kwan-Young L,
Ju MD. Hydrogen production by steam reforming of biomass-
derived glycerol over Ni-based catalysts. Catal Today
2014;228:145e51.
[26] Dou B, Wang C, Song Y, Chen H, Xu Y. Activity of NieCueAl
based catalyst for renewable hydrogen production from steam
reforming of glycerol. Energy Convers Manag 2014;78:253e9.
[27] Yurdakul M, Ayas N, Bizkarra K, Doukkali MEl, Cambra JF.
Preparation of Ni-based catalysts to produce hydrogen from
glycerol by steam reforming process. Int J Hydrogen Energy
2016;41:8084e91.

[28] Bobadilla LF, Penkova A, Alvarez A, Domı́nguez MI, Romero-
Sarria F, Centeno MA, et al. Glycerol steam reforming on
bimetallic NiSn/CeO2eMgOeAl2O3 catalysts: influence of
the support, reaction parameters and deactivation/
regeneration processes. Appl Catal A Gen 2015;492:38e47.
(a)Bobadilla LF, Romero-Sarria F, Centeno MA, Odriozola JA.
Promoting effect of Sn on supported Ni catalyst during steam
reforming of glycerol. Int J Hydrogen Energy 2016;41:9234e44.
[29] Papageridis KN, Siakavelas G, Charisiou ND, Avraam DG,
Tzounis L, Kousi K, et al. Comparative study of Ni, Co, Cu
supported on g-alumina catalysts for hydrogen production
via the glycerol steam reforming reaction. Fuel Process
Technol 2016;152:156e75.
[30] Zamzuri NH, Mat R, Amin NAS, Talebian-Kiakalaieh A.
Hydrogen production from catalytic steam reforming of
glycerol over various supported nickel catalysts. Int J
Hydrogen Energy 2016. http://dx.doi.org/10.1016/
j.ijhydene.2016.05.084. In press.
[31] Profeti LPR, Ticianelli EA, Assaf EM. Production of hydrogen
via steam reforming of biofuels on Ni/CeO2-Al2O3 catalysts
promoted by noble metals. Int J Hydrogen Energy
2009;34(12):5049e60.
[32] Regalbuto J. Catalyst preparation: science and engineering. Ed
by John R. Regalbuto. Taylor & Francis; 2007 (CRC Press, 2015).
[33] Adhikari S, Fernando S, Gwaltney SR, Filip To SD, Mark
Bricka R, Steele PH, et al. A thermodynamic analysis of
22742 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 1 ( 2 0 1 6 ) 2 2 7 3 2 e2 2 7 4 2
hydrogen production by steam reforming of glycerol. Int J
Hydrogen Energy 2007;32:2875e80.
[34] Wang H, Wang X, Li M, Li S, Wang S, Ma X. Thermodynamic
analysis of hydrogen production from glycerol autothermal
reforming. Int J Hydrogen Energy 2009;34:5683e90.
[35] Chen H, Ding Y, Cong NT, Dou B, Dupont V, Ghadiri M, et al. A
comparative study on hydrogen production from steam-
glycerol reforming: thermodynamics and experimental.
Renew Energy 2011;36:779e88.
[36] Demsash HD, Mohan R. The thermodynamic analysis of
hydrogen production from glycerol steam reforming. In:
Mishra GC, Yadav SK, editors. Energy Res. Environ. Manag.
An Innov. Approach. New Delhi: Krishi Sanskriti
Publications; 2015. p. 16e25.
[37] Calles JA, Carrero A, Vizcaı́no AJ, Garcı́a-Moreno L. Hydrogen
production by glycerol steam reforming over SBA-15-
supported nickel catalysts: effect of alkaline earth promoters
on activity and stability. Catal Today 2014;227:198e206.
[38] Gallegos-Sua rez E, Guerrero-Ruiz A, Ferna  ndez-Garcı́a M,
Rodrı́guez-Ramos I, Kubacka A. Efficient and stable NieCe
glycerol reforming catalysts: Chemical imaging using X-ray
electron and scanning transmission microscopy. Appl Catal
B Environ 2015;165:139e48.
[39] Silverwood IP, Hamilton NG, Staniforth JZ, Laycock CJ,
Parker SF, Ormerod RM, et al. Persistent species formed
during the carbon dioxide reforming of methane over a
nickelealumina catalyst. Catal Today 2010;155:319e25.
[40] He Z, Yang M, Wang X, Zhao Z, Duan A. Effect of the
transition metal oxide supports on hydrogen production
from bio-ethanol reforming. Catal Today 2012;194:2e8.
[41] Thyssen VV, Maia TA, Assaf EM. Ni supported on La2O3-SiO2
used to catalyze glycerol steam reforming. Fuel
2013;105:358e63.
[42] Xu W, Liu Z, Johnston-Peck AC, Senanayake SD, Zhou G,
Stacchiola D, et al. Steam reforming of ethanol on Ni/CeO2:
reaction pathway and interaction between Ni and the CeO2
support. ACS Catal 2013;3:975e84.
[43] Kim P, Kim Y, Kim H, Song IK, Yi J. Synthesis and
characterization of mesoporous alumina with nickel
incorporated for use in the partial oxidation of methane into
synthesis gas. Appl Catal A Gen 2004;272:157e66.
[44] Zhang B, Tang X, Li Y, Cai W, Xu Y, Shen W. Steam reforming
of bio-ethanol for the production of hydrogen over ceria-
supported Co, Ir and Ni catalysts. Catal Commun
2006;7(6):367e72.
[45] Srisiriwat N, Therdthianwong S, Therdthianwong A.
Oxidative steam reforming of ethanol over Ni/Al2O3
catalysts promoted by CeO2, ZrO2 and CeO2eZrO2. Int J
Hydrogen Energy 2009;34(5):2224e34.
[46] Fogler HS. Elements of chemical reaction engineering. 4th ed.
Prentice Hall; 2006.
[47] Froment GF, Bischoff KB, Wilde JD. . Chemical reactor
analysis and design. 3rd ed. John Wiley and Sons, Inc; 2011.