Applied Catalysis A: General 381 (2010) 1–7

Contents lists available at ScienceDirect

Applied Catalysis A: General

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

Review

Catalytic features of Rh and Ni supported catalysts in the steam reforming of

glycerol to produce hydrogen
V. Chiodo a,∗ , S. Freni a , A. Galvagno b , N. Mondello a , F. Frusteri a
a
CNR-ITAE “Nicola Giordano”, Via S. Lucia sopra Contesse, 5–98126 - Messina, Italy
b
DIIM, Università degli Studi di Catania, Viale A. Doria, 6–95125 - Catania, Italy

a r t i c l e i n f o a b s t r a c t

Article history: Glycerol steam reforming on Rh and Ni supported catalyst was investigate in view to produce syn-gas to
Received 9 December 2009 feed a high temperature fuel cell system (SOFC-MCFC). Results obtained revealed that glycerol is subjected
Received in revised form 17 March 2010 to pyrolysis phenomena and at temperature higher than 720 K drastically decomposes before to reach
Accepted 18 March 2010
the catalyst surface. Rh/Al2 O3 catalyst resulted to be more active and stable than Ni supported catalysts
Available online 25 March 2010
but independently from the catalyst used (Rh or Ni) and temperature investigated, the reaction is affected
by coke formation mainly promoted by the large presence of olefins formed by glycerol thermal decom-
Keywords:
position. Even if for thermodynamic reasons the hydrogen production should be favored by operating
Rh and Ni catalysts
H2 production
at high temperature, results obtained demonstrate that it is more convenient to operate at temperature
Glycerol steam reforming not higher than 923 K, since higher reaction temperature promotes the formation of encapsulated carbon
Coke formation which negatively reflects on catalyst stability.
© 2010 Elsevier B.V. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
2. Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2.1. Catalyst preparation and characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2.2. Catalytic measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
3. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
4. Discussions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
5. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

1. Introduction 17 wt% of the bio-diesel mass produced. The glycerol separated

from bio-fuel contain about 20 wt% of water.
The increasing commercial interest on diesel production by One promising way to use glycerol as carbon source is the pro-
transesterification of vegetable oil aimed several research groups to duction of hydrogen or syn-gas by steam reforming (SR) process
investigate innovative routes to improve the economic balance of in order to supply a solid oxide or molten carbonate fuel cells to
the overall process. For this purpose, most emphasis has been also produce electricity and heat [1,2].
addressed towards the development of processes that can convert Theoretical feasibility of the steam reforming process of glyc-
the glycerol, the main by-product of FAME process, into added value erol has been demonstrated by different authors [3,4] in terms of
chemicals like bio-fuels or hydrogen. hydrogen yield as a function of process temperature, pressure and
The amount of glycerol produced in the bio-fuel production pro- steam/carbon ratio.
cess, which use vegetable oil as renewable source, is about the Steam reforming process of hydrocarbons is, as well known, a
strongly endothermic reaction usually performed on Ni based cata-
lysts that allow to obtain a hydrogen rich gas mixture containing H2 ,
∗ Corresponding author. Tel.: +39 090 624219. CO2 , CH4 , CO and H2 O. However, the composition of reformed gas
E-mail address: [email protected] (V. Chiodo). change as a function of operative conditions (temperature, pressure

0926-860X/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2010.03.039
2 V. Chiodo et al. / Applied Catalysis A: General 381 (2010) 1–7

and steam/carbon ratio), while catalyst life is depending upon sev- Table 1
eral deactivation phenomena that usually affect steam reforming
reaction [5–7]. Metal loading (wt%) MSA (m2 Me/gcai ) BETSA (m2 /g) ds (nm)
Ni/based catalysts, typically used for methane SR have been Rh/Al2 O3 5 8.1 163 2.1
recently investigated to produce H2 by SR of glycerol. In partic- Ni/Al2 O3 30 12 44 7
ular, Ni based catalysts investigated in aqueous phase reforming Ni/MgO 21 6.3 45 8.1
(AFR) of glycerol resulted to be affected by significant deactivation Ni/CeO2 30 6.2 27.2 22.8

phenomena [5]. Ni appear more suitable for application on steam

reforming of glycerol if supported on bare alumina [6] or modi-
der”; SA (BET): 27 m2 g−1 ) as carriers. Before impregnation MgO
fied with promoters [5,7]. These studies indicate that Ni catalysts
was stabilized in steam atmosphere at 973 K for 12 h. Using a
are sufficiently active and selective to syn-gas production but coke
RhCl2 aqueous solution for the preparation of 5% Rh/Al2 O3 and
formation represents a serious problem to overcame since glycerol
Ni-acetate for the synthesis of 21% Ni/MgO. An aqueous solution
give rise to the formation of several intermediate products that
of Ni(NO3 )2 ·6H2 O was used to synthesize 30% Ni/CeO2 and 30%
promote coke formation. Pt/alumina catalysts were also investi-
Ni/Al2 O3 catalysts. All samples were dried at 353 K for 24 h and
gated in SR of glycerol [8] but, even if, higher reaction temperature
then air calcined at 673 K for 12 h. Catalysts were pressed at 400 bar,
enhances the overall reforming process, reaction pathways is char-
crushed and sieved and the 40–70 mesh fraction was used for the
acterized by several side reactions (like dehydration) that leads to
catalytic tests.
carbon formation that negatively affects catalyst stability. Carbon
The BET surface area and porosity of carriers have been deter-
formation has been also observed on La2 O3 supported Ru catalysts
mined in a conventional volumetric apparatus at 77 K with Carlo
[9].
Erba Sorptomatic 1990 series instrument.
On the basis of a preliminary experience acquired in the use of
The metal particle size distribution of catalyst was evaluated
Rh and Ni catalysts to produce syn-gas by ethanol steam reforming
by TEM analysis, using a Philips CM12 instrument provided with
[10,11], in this paper we have focused our attention to evaluate
a high resolution camera which allow acquisition and elaboration
the behavior of such catalysts in the glycerol steam reforming.
of images. The samples were ultrasonically dispersed in isopropyl
The study was focused to investigate the deactivation phenomena
alcohol and deposited on a carbon supported film.
occurring on such systems with the objective to find technolog-
Carbon deposed during reaction was evaluated using a CHNS
ical solutions to design an active and stable catalyst for glycerol
ThermoFlash EA 1112 Series instrument. Catalyst samples were
conversion into syn-gas.
treated at high temperature in air and the CO2 produced by oxida-
tion with oxygen was determined by a high sensitivity TC detector.
2. Experimental The main chemical–physical properties of the catalyst prepared
according to the above procedure are reported in Table 1.
2.1. Catalyst preparation and characterization
2.2. Catalytic measurements
Catalyst were prepared by impregnation (incipient wetness)
using ␥-Al2 O3 AKZO-NOBEL 001-3P (SA (BET) 260 m2 g−1 ), MgO The flow-sheet of glycerol steam reforming process is shown
Martin Marietta (SA (BET) 120 m2 g−1 ) and CeO2 (“smoke pow- Fig. 1. Catalytic experiments were performed at atmospheric

Fig. 1. Experimental set up used to perform the glycerol steam reforming reaction.
 V. Chiodo et al. / Applied Catalysis A: General 381 (2010) 1–7 3

Fig. 2. Blank test. Glycerol conversion and product selectivity as a function of reaction time: TR = 1073 K; S/C molar ratio = 3.

pressure in a fixed-bed linear quartz micro-reactor (i.d. = 4 mm; should be considered:

Hbed = 1.0–2.0 cm). 0.02–0.06 g of catalyst (grain size: 40–70 mesh) m
diluted with 200 mg of similar sized inert material (SiC) were used Cn Hm Ok + (n − k)H2 O → nCO + [(n + 2−k
)]H2
for the catalytic tests. Before each test run, catalysts were reduced Nevertheless, for an appropriated evaluation of catalytic phe-
“in situ” at T = 1073 K for 1 h under hydrogen flow. Water/glycerol nomena occurring during SR of such compounds, should take into
mixture (steam–carbon molar ratio; S/C = 3) was fed by isocratic HP account that, at high temperature they could decompose in pri-
1100 pump and vaporized in a N2 flow in a stainless steel reactor mary and secondary pyrolysis products before to reach the catalyst
held at 493 K. surface.
Experiments were performed a temperature raging from 723 Glycerol (C3 H8 O3 ) is an oxygenated organic compounds con-
to 1073 K and gas hourly space velocity (GHSV) raging from taining three alcoholic groups. On the base of theoretical evaluation
5000 to 30,000 mlC3H8O3 h−1 mlcat −1 . The total flow correspond- the following reactions should occur in the steam reforming of
ing at the GHSV of 30,000 h−1 was 130 ml min−1 and the reaction glycerol:
stream composition was: 23.08 vol% N2 ; 69.23 vol% H2 O; 7.69 vol%
C3 H8 O3 ). C3 H8 O3 + 3H2 O → 7H2 + 3CO2 (Glycerol steam reforming)
Before analysis, the gas reaction stream was cooled at 275 K and
separated into liquid and gas phases. The gas phase was analyzed on C3 H8 O3 → 4H2 + 3CO (Direct decomposition)
line by gas-chromatography (HP Agilent 6890) equipped with three
columns (Molecular Sieve 5A, Porapack QS and Hysep Q) system
CO + H2 O → H2 + CO2 (WGS)
and conductivity detector (TCD); GC data were acquired and elabo-
rated by HP-IB interfaced computer system (HP Chemistation). The
liquid phase was analyzed by GC-MS Agilent 5975C provided with CH4 + H2 O → CO + 3H2 (CH4 steam reforming)
capillary column HP Innowax. The following equations were used
However, taking into account that glycerol is not a thermally
to calculate the glycerol conversion, selectivity and H2 productivity
stable molecule its could decomposes, then a preliminary test with-
(mol of H2 produced per mol of glycerol converted):
out catalyst has been carried out to ascertain the extent of pyrolysis
C atoms in gas products phenomena that could contribute to negatively affects the catalyst
Glycerin conversion to gas =
total C atoms in the feedstock resistance and the evolution of steam reforming process.
× 100 In Fig. 2 a blank test performed at 1073 K using a reaction stream
containing 7% of glycerol, under steam reforming conditions is
C atoms in species ␣ shown in terms of glycerol conversion and products selectivity.
C selectivity = × 100 Results indicate that glycerol significantly decomposes (conver-
C atoms produced in gas phase
sion close to 65 mol%) giving rise to the production of a series of
where species ␣ is CO, CO2 , CH4 , C2 H4 , C2 H6 and C3 H6 , products like CO (upper 50 mol%), CO2 (1 mol%), olefins (ethylene
H2 moles produced 1 and propylene about 25 mol%), CH4 (∼13 mol%) and H2 (∼10 mol%).
H2 selectivity = × × 100 In addition the analysis of the condensed phase of reaction stream
C atoms produced in gas phase R
(30 mol%), performed by gas-chromatography GC/MS, revealed the
where R is H2 /CO2 ratio (7/3 in the case of glycerol steam reforming presence of several unsaturated compounds (not well quantified)
reaction). such as acetone, acetaldehyde, ethanol, propanol, acetic acid and
2,3-dyhydroxylpropanal.
3. Results Therefore, taking into account that the real composition of reac-
tion stream that reach the catalyst surface is strictly depending
The steam reforming is the most commonly used method for upon the extent of glycerol decomposition, a preliminary test a
producing hydrogen in the chemical industry. In case of oxygenated TR = 1073 K to compare the performance of different catalysts inves-
organic compounds use, the following stoichiometry of reaction tigated in this study was carried out feeding the reactor with a
4 V. Chiodo et al. / Applied Catalysis A: General 381 (2010) 1–7

Fig. 4. Coke formation rate observed using different metal supported catalysts:
TR = 1073 K; GHSV = 30,000 h−1 ; S/C molar ratio = 3.

CHNS analysis of Rh catalyst used at different reaction tempera-

ture (Fig. 7) revealed that coke forms at low rate (1.9 mgC/gcat h) at
Fig. 3. H2 yield (mol of hydrogen produced per mol of glycerol converted) vs time 723 K, while a comparable coke formation rate was observed at 823
on stream: TR = 1073 K; GHSV = 30,000 h−1 ; S/C = 3 mol/mol. and 1073 K (about 2.3 mgC/gcat h). The highest coke formation rate
was observed at 923 K (4.2 mgC/gcat h). By comparing the results
reported in Figs. 6 and 7 it can be deduced that the trend of coke
formation follow the same trend of H2 productivity.
reaction mixture containing 23.08 vol% N2 , 69.23 vol% H2 O and According to our previously experience [10,11,13] to shed light
7.69 vol% C3 H8 O3 . The results obtained are shown in Fig. 3 in terms on the behavior of catalysts used at different temperatures we have
of H2 yield (mol of H2 produced/mol of glycerol converted) as a focused our attention to investigate the metal sintering phenomena
function of reaction time. It can be seen that a significant difference eventually occurring during reaction and the nature of coke formed.
among catalysts exist both in terms of initial activity and deactiva- TEM analysis of used catalysts (see images shown in Fig. 8)
tion trend. Rh/Al2 O3 catalyst showed the best performance both highlights a substantial morphological difference between cata-
in terms of H2 yield and stability while, among nickel supported lysts used at different temperature. In terms of metal sintering it
catalysts, the Ni/Al2 O3 resulted to be the most active ones. What can be observed that up to 923 K the Rh particles size distribution
it is very important to underline is that all catalysts drastically does not change significantly, while at 1073 K the presence of sev-
deactivate during the first 2 h of reaction. For Ni catalysts activity eral Rh particles larger than 15 nm are diagnostic of the occurrence
was completely lost after 6 h of reaction, but a residual hydrogen of a drastic sintering. As regard the coke formation, it seems enough
production remain due to thermal decomposition of glycerol (see clear that up to 923 K (see Fig. 8A–C) the type of coke formed is of
results reported in Fig. 2). filamentous nature while at higher temperature (1073 K) encapsu-
To ascertain if the difference in the Rh and Ni catalysts behavior lating carbon forms prevalently.
could depend upon the extent of coke formation, CHNS elementary
analysis was carried out on spent catalysts. The results obtained (see
Fig. 4) clearly indicated that coke forms on all catalytic systems with
significant rate but on Rh/Al2 O3 catalyst the coke formation take
place with lower rate (about 2.2 mgC/gcat h) respect to the Ni cata-
lysts. The rate of carbon formation was calculated considering the
amount of coke formed in 20 h of reaction. On differently supported
Ni catalysts, coke formation can be considered to occurs with simi-
lar rate (4–6 mgC/gcat h) independently of carrier employed (Al2 O3 ,
MgO or CeO2 ). The small difference observed can be justified on the
basis of the different initial conversion level.
In order to investigate the influence of reaction temperature
on hydrogen yield a series of experiments have been performed
using Rh/Al2 O3 catalyst. Results show in Fig. 5 reveal a progres-
sive decreasing of hydrogen production during the first 5 h of the
reaction at all temperatures investigated. Furthermore, by compar-
ing the results in terms of initial H2 yield as a function of reaction
temperature a not linear trend was observed. This is more evident
from data show in Fig. 6. Hydrogen production increase from 1.8
to 5 mol% as the temperature increases from 723 to 923 K, but for
temperatures higher than 923 K, on the contrary of what foreseen
from thermodynamic equilibrium, a sudden decreasing of hydro- Fig. 5. H2 yield (mol of hydrogen produced per mol of glycerol converted) as a
gen yield is observable. function of time on stream on Rh/Al2 O3 catalyst. GHSV = 30,000 h−1 ; S/C = 3 mol/mol.
 V. Chiodo et al. / Applied Catalysis A: General 381 (2010) 1–7 5

Fig. 6. H2 yield (mol of hydrogen produced per mol of glycerol converted) vs

temperature on Rh/Al2 O3 catalyst. GHSV = 30,000 h−1 ; S/C = 3 mol/mol. Dashed line
represent H2 yield at different temperatures according to thermodynamic analysis
[3]. Fig. 7. Coke formation rate on Rh used catalysts: tests performed at
GHSV = 30,000 h−1 ; S/C molar ratio = 3 mol/mol.

The resistance of Rh towards sintering ad coke formation phe-

by amorphous coke (see Fig. 10A) and some particles (see the high
nomena was also investigated by an endurance test performed at
magnification image–Fig. 10B) result to be encapsulated by coke.
923 K and low GHSV (5000 h−1 ). As it can be seen from the Fig. 9,
the glycerol conversion remain almost constant up to 8 h of reac-
tion after that it drastically decreases with a progressive trend. A 4. Discussions
similar behavior is observable also for the H2 yield.
TEM images of catalyst used in the endurance test (see Fig. 10) Blank test shown in Fig. 2 indicate that at 1073 K, glycerol
revealed that even at 923 K the catalyst surface appear to be covered is not stable and decomposes giving rise to the formation of

Fig. 8. TEM images of Rh/Al2 O3 catalysts used at (A) 723 K; (B) 823 K; (C) 923 K; (D) 1073 K.
6 V. Chiodo et al. / Applied Catalysis A: General 381 (2010) 1–7

Fig. 9. Endurance test performed using Rh/Al2 O3 catalyst: TR = 923 K; GHSV = 5000 h−1 ; S/C molar ratio = 3.

several compounds. Ethylene is prevalent as a consequence of metal to activate C–O (typical of oxygenated compounds), C–C and
the initial C–C bond cleavage which leads to the formation of C C bonds.
CH2 (OH)–CH(OH) and CH2 –OH radicals that subsequently evolve Several papers concerning H2 production from glycerol mainly
towards the formation of C2 H4 ,H2 and CH4 [14–22]. Furthermore, by steam reforming process have been recently published
considering that the radical decomposition mechanism foresee the [1–9,12,15]. Performance of both rhodium and nickel supported
formation of acetaldehyde (dehydrogenation reaction) the large catalysts has been evaluated in terms of activity and hydrogen
presence of CO and CH4 can be associated to its decomposition productivity, but not clear evidences have been reported to estab-
(CH3 CHO → CH4 + CO). Instead, the low presence of CO2 is an index lish the most suitable catalyst system for glycerol steam reforming
of the fact that water gas shift reaction (WGS) does not occur and [7,12,15]. Even though, short time experiments in its self could
the presence of 10% of H2 is imputable to the occurrence of dehy- not be representative to definitively establish what catalytic sys-
drogenation reactions. tem is more active and stable in the studied reaction, on the basis
Experiments performed in the presence of catalysts evidenced of both literature evidences and results here reported, there are
that the hydrogen yield increases from 6.8 mol% (blank test) to valid reasons to believe that Rh catalysts are more adequate to per-
40 mol% using Rh/Al2 O3 and to 19 mol% for Ni/Al2 O3 catalysts. form glycerol steam reforming. The reasons of such result are to
By considering the chemical–physical characterization data be searched in the composition of reaction stream that, depending
reported in Table 1, Ni/Al2 O3 catalyst has a MSA higher that Rh on reaction temperature dramatically changes since glycerol easily
catalyst, therefore, at least at the beginning of reaction, since Ni is decomposes before to reach the catalyst surface.
a excellent steam reforming catalyst, a higher or similar catalytic Olefins (C2 H4 , C3 H6 ) and carbon monoxide are the main chem-
behavior was expected, instead Rh was found to be more active. This ical species reaching the catalyst surface, therefore, it can be
is because in the steam reforming of glycerol what should primarily deduced that catalytic performance is straight depending upon the
be considered is that the composition of reaction stream is complex tendency towards: (i) C C cleavage, (ii) coke formation inhibition
and depends upon the extent of glycerol decomposition. By operat- by polymerization of CHx species and/or CO dissociation (Bouduard
ing at high temperatures (see Fig. 2) it is demonstrated that glycerol reaction) [11,13,23,24].
easily decomposes into CO, CO2 , olefins, water hydrogen and oxy- It is well known that Rh catalysts are active in C C bond
genates before to reach the catalytic surface, therefore, the catalytic activation and furthermore they are more resistance than Ni cat-
consideration should be done take into account the ability of active alyst towards coke formation because Rh is a very active metal in

Fig. 10. TEM image of Rh/Al2 O3 catalyst used in the endurance test: TR = 923 K; GHSV = 5000 h1 ; S/C molar ratio = 3.
 V. Chiodo et al. / Applied Catalysis A: General 381 (2010) 1–7 7

hydrogenation reactions that, in case of removal of unsaturated - Glycerol is subjected to pyrolysis phenomena and at 1073 K
compounds from catalyst surface, is a fundamental properties to drastically decomposes giving rising to the formation of CO,
depress the coke formation. The results shown in Fig. 4 clearly CO2 , olefins, water, hydrogen, acetone, acetaldehyde, ethanol,
demonstrate that on Rh supported catalyst a lower amount of coke propanol, acetic acid and 2,3-dyhydroxylpropanal;
forms respect to Ni/Al2 O3 catalyst. - Rh/Al2 O3 catalyst is more active and stable than Ni supported
From results reported in Fig. 6 it can be observe that, on the catalysts;
contrary of what the thermodynamic foresees, at the highest tem- - The steam reforming of glycerol, independently from the cata-
perature (1073 K) the catalyst activity is lower than that observed lyst used (Rh or Ni) and temperature investigated, is affected by
at 923 K. As a first hypothesis we have thought that the reason of coke formation due to the large presence of olefins in the reac-
such result was due to the metal sintering phenomenon that as tion stream, formed by glycerol thermal decomposition, in the
well know normally occurs at high reaction temperature, but TEM reaction stream;
investigation have revealed that particle size distribution of used - Even if for thermodynamic reasons the hydrogen production
catalyst does not change significantly respect to the fresh reduced should be favored by operating at high temperature results
sample. In particular the particle size distribution increases from obtained suggest that it is more convenient to operate at tem-
12 to 15 nm not sufficient to justify the deactivation observed, then perature not higher than 923 K;
our attention was addressed to investigate the morphology of coke - High reaction temperature promotes the formation of encapsu-
formed during reaction in order to try to find a relationship between lating carbon which negatively reflects on catalyst stability.
catalyst behavior and reaction temperature.
TEM images shown in Fig. 8 clearly demonstrate that as the reac- References
tion temperature increases the coke formation mechanism change,
in particular at temperature of 923 and 973 K is prevalent the for- [1] M.A. Pagliaro, R.B. Ciriminna, H.D. Kimura, M.C. Rossi, C. DellaPina, Angewandte
Chemie-Int. Edit. 46 (24) (2007) 4434–4440.
mation of filamentous carbon while at 1073 K the carbon formed [2] R.R. Soares, D.A. Simonetti, J.A. Dumesic, Angewandte Chemie-Int. Edit. 45 (24)
is mainly of encapsulating type. This is the key factor that could (2006) 3982–3985.
explain why at high temperature the catalyst performance worse [3] S. Adhikari, S. Fernando, S.R. Gwaltney, S.D. Filip To, R.M. Bricka, P.H. Steele, A.
Haryanto, Int. J. Hydrogen Energy 32 (2007) 2875–2880.
notwithstanding the steam reforming process is favored. The for- [4] X. Wang, S. Li, H. Wang, B. Liu, X. Ma, Energy Fuels 22 (2008) 4285–4291.
mation of encapsulating coke is deleterious for the Rh catalyst that [5] A. Iriondo, V.L. Bario, J.F. Cambra, P.L. Arias, M.B. Guemez, R.M. Navarro, M.C.
loses activity quickly and H2 yield result to be lower even when Sanchez-Sanchez, J.L.G. Fierro, Top. Catal. 49 (2009) 46–58.
[6] S. Adhikari, S. Fernando, A. Haryanto, Catal. Today. 129 (2007) 355–364.
the steam reforming is carried out at high temperature even if
[7] I.N. Buffoni, F. Pompeo, G.F. Santori, N.N. Nichio, Catal. Commun. 10 (2009)
the reaction is carried out in favorable thermodynamic conditions. 1656–1660.
Why at high temperature the formation of encapsulating carbon is [8] N. Luo, X. Fu, F. Chao, T. Xiao, P.P. Edwards, Fuel 87 (2008) 3483–3489.
prevalent it was well described in a previous paper [23–27]. What [9] T. Hirai, N. Ikenaga, T. Miyake, T. Suzuki, Energy Fuels 19 (2005) 1761–
1762.
governs the carbon formation mechanism is the elementary car- [10] F. Frusteri, S. Freni, L. Spadaro, V. Chiodo, G. Bonura, S. Donato, S. Cavallaro,
bon rate diffusion through the metal particle: an high diffusion Catal. Commun. 5 (2004) 611–615.
rate favors the formation of filamentous carbon while a low diffu- [11] N. Palmeri, V. Chiodo, S. Freni, F. Frusteri, J.C.J. Bart, S. Cavallaro, Int. J. Hydrogen
Energy 33 (22) (2008) 6627–6634.
sion rate of carbon favors the formation of encapsulating carbon [12] S. Adhikari, S. Fernando, A. Haryanto, Catal. Today 129 (2007) 355–364.
species. Naturally, when metal particle is encapsulated by coke [13] S. Cavallaro, V. Chiodo, S. Freni, N. Mondello, F. Frusteri, Appl. Catal. A 249 (2003)
immediately deactivates. 119–128.
[14] W. Buhler, E. Dinjus, H.J. Ederer, A. Kruse, C. Mas, Supercrit. Fluid 22 (2002)
The endurance test result performed at low GHSV (5000 h−1 ) 37–53.
and 923 K revealed that Rh catalyst is not so resistant in glycerol [15] S. Adhikari, S. Fernando, A. Haryanto, Renew. Energy 33 (2008) 1097.
steam reforming and as the coke accumulates on the surface it [16] R.D. Cortright, R.R. Davda, J.A. Dumesic, Nature 418 (2002) 964.
[17] J.R.H. Ross, M.C.F. Steel, A. Zeini-Isfahani, J. Catal. 52 (1978) 280.
deactivates on the time. What is important to underline in such [18] R.R. Davda, J.W. Shabaker, G.W. Huber, R.D. Cortright, J.A. Dumesic, Appl. Catal.
test is that, even the prevailing amount of coke formed is of fila- B 43 (2003) 13–26.
mentous nature (see Fig. 10A) even at such temperature, but with [19] J.W. Shabaker, J.A. Dumesic, Ind. Eng. Chem. Res. 43 (2004) 3105–3112.
[20] M.R. Nimlos, S.J. Blanksby, X. Qian, M.E. Himmel, D.K. Johnson, J. Phys. Chem. A
lower extent respect to the experiment performed at 1073 K, some
110 (2006) 6145–6156.
Rh particles resulted to be encapsulated by amorphous carbon with [21] C.-W. Chiu, M.A. Dasari, G.J. Suppes, AIChE J. 52 (2006) 3543–3548.
negative consequence on catalyst stability. [22] T.A. Semelsberg, R.L. Borup, J. Power Source 155 (2006) 340–352.
[23] F. Frusteri, L. Spadaro, F. Arena, A. Chuvilin, Carbon 40 (2002) 1063.
[24] C.H. Bartholomew, Catal. Rev.-Sci. Eng. 24 (1982) 67.
5. Conclusions [25] G. Italiano, C. Espro, F. Arena, F. Frusteri, A. Parmaliana, Appl. Catal. A: 357
(2009) 58–65.
From the results here reported the following conclusions can be [26] J.R. Rostrup-Nielsen, J. Catal. 33 (1974) 184.
[27] C.H. Bartholomew, Appl. Catal. A: Gen. 212 (2001) 17–60.
drown: