1.

Introduction

5-hydroxymethylfurfural (HMF) is one of the top biomass-derived building blocks, as

reported by the U.S. Department of Energy [1]. HMF can be converted into many promising
chemical intermediates, such as 2,5-dimethyltetrahydrofuran (DMTHF) [2] and [3], 2,5dimethylfuran (DMF) [4], [5] and [6], 2,5-diformylfuran (DFF) [7], and 2,5furandicarboxylic acid (FDCA) [8], [9] and [10]. Generally, HMF can be generated from the
conversion of carbonhydrates with the assistance of catalysts [11]. So far, various catalytic
systems have been reported to achieve maximum yields of HMF, and these systems use
different catalysts (e.g., mineral and organic acids [12], [13] and [14], salts [15],
[16] and [17], and zeolite [18]) in different solvents (e.g., ionic liquids [19], [20], [21],
[22] and [23], organic solvents (DMSO [24], [25] and [26], acetone [26] and [27]) and biphasic solvents [28] and [29]. For examples, Riisager and co-workers obtained a high yield of
HMF (i.e., 40 mol%) by using boric acid as a promoter in ionic liquids [13]. Another high
HMF yield was reported by the Zhao group. They demontrated that 68 mol% of HMF could
be achieved at 100 C for 3 h by using 1-ethyl-3-methylimidazolium chloride [(EMIM)Cl] as
the solvent and CrCl2 as the catalyst [30].
HMF has been produced through the conversion of different feedstocks, including
monosaccharides (fructose or glucose) [21], [31], [32] and [33], disaccharides (maltose,
sucrose, or cellobiose) [34], and polysaccharides (insulin, starch, or cellulose) [35],
[36] and [37]. Up to now, most studies have still focused on using feedstocks exhibiting
simple and regular molecular structures, due to easier catalytic mechanisms. Tong et al.
reported maximum yield of HMF (72 mol%) with 87% selectivity from fructose dehydration
with Brnsted acidic ionic liquids (N-methyl-2-pyrroli-donium methyl sulfonate [NMP]+
[CH3SO3]) at 90 C for 2 h [31]. The production of HMF from sucrose has been studied by
Qi et al. [34] who showed that HMF yield of 76 mol% could be observed in the [BMIM]Cl
and CrCl36H2O catalytic system with microwave heating at 100 C for 5 min. A more
complex feedstock such as cellulose has also been used because it is more abundant and has
lower value than fructose. The conversion of cellulose to HMF in a one-pot reaction resulted
in 52.7 mol% yield when using Cr[(DS)H2PW12O40]3 as the catalyst at 150 C for 2 h [35].
However, from the viewpoint of practical applications, direct conversion of raw
lignocellulosic biomass such as rice straw and wood is more applicable and highly demanded.
The studies on the direct utilization of raw lignocellulosic biomass are quite limited [38],
[39] and [40]. Binder et al. [38] reported the first direction conversion of cellulose in
lignocellulosic in a complex catalytic system of CrCl3-HCl-DMA-LiCl-[Emim]Cl at 140 C
for 2 h, and HMF yield of 48 mol% could be obtained. Another example was reported by
Zhang and Zhao [39] where CrCl36H2O and ionic liquid ([C4mim]Cl) were used as the
catalyst and solvent, respectively, with the assistance of microwave irradiation. HMF yields
of 45, 47 and 52 mol% were obtained from corn stalk, rice straw and pine wood, respectively.
Recently, untreated wheat straw was also investigated as a practical source of raw biomass
from which to produce HMF, as shown by the Wang group [40]. The formation of HMF from
wheat straw was carried out at 160 C using a complex catalytic system of [BMIM]Cl- CrCl3-

LiCl. From this, an HMF yield of 57.9 mol% was achieved. Despite these pioneering studies;
recurring disadvantages of the experimental systems were the utilization of high temperature
(> 140 C) and complicated catalytic systems. Thus, the challenge of developing a simple
catalytic system and a reduced reaction temperature for producing HMF from raw
lignocellulosic biomass still remains.
Here we propose combinational treatments including pre-treatments of raw lignocellulosic
biomss (i.e., straw and wood chip) and the subsequent catalytic conversion. After
optimization of reaction conditions, catalysts, and solvents, we concluded that a HMF yield
as high as 79 mol% could be achieved when the raw lignocellulosic biomass (wood chips)
were pre-treated in diluted NaOH solution (3 wt%) and then catalyzed in the system using
CrCl36H2O as the catalyst and 1-butyl-3-methyl imidazolium chloride ([BMIM]Cl) as the
solvent at 120 C for 2 h. We have also studied the influence of pre-treatments and various
catalysts on the yields of HMF, together with the investigation of the activation energy of the
catalytic conversion.
2. Materials and methods
2.1. Chemicals

Glucose (99%), fructose (99%), 5-hydroxymethylfurfural (HMF) (98%), Chromium(III)

chloride hexahydrate (96%), 1-Butyl-3-methylimidazolium chloride ([BMIM]Cl) (98%), 1Ethyl-3-methylimidazolium chloride ([EMIM]Cl) (98%), 1-Ethylpyridinium chloride
([Epy]Cl) (98%), dimethyl sulfoxide (DMSO) (99%), N,N-dimethylacetamide (DMA) (99%),
N,N-dimethylfomamide (DMF) (99%), amberlyst 15 hydrogen form, boric acid (99.5%),
phosphotungstic acid hydrate were purchased from Sigma-Aldrich and Acros Company and
used without further purification.
2.2. Raw lignocellulosic biomass

The wood was obtained from a polywood factory. The wood with a moisture content of 10%
was chopped into smaller pieces (2 cm) with a shredder and used as the feedstock. The
composition of the wood chip was as follows: glucan 44.2 0.4%, xylan 15.7 0.1%,
arabinan 0.8 0.0%, acetate 2.9 0.0%, extractive 3.2 0.4%, ash 0.6 0.0% and lignin
30.5 0.1%. The rice straw was gathered from a private farm in Taiwan, and its composition
was analyzed as: glucan 32.9 0.1%, xylan 18.7 0.1%, arabinan 3.2 0.1%, extractive
10.1 0.2%, ash 11.3 0.1% and lignin 19.1 1.1%. The details of the analytic procedure of
the raw biomass were described in the supporting information, and the composition was
calculated on a dry weight basis [41].
2.3. Pretreatment

Typically, 0.3 g of raw lignocellulosic biomass materials (i.e., wood and straw) was treated by
using 30 mL of diluted acids (HCl or H2SO4) or and alkaline (NaOH) with a concentration of
1 or 3 wt%. The mixture was stirred at RT or 60 C for 24 h. After that, the mixture was
filtered by vacuum filtration and dried at 50 C for 24 h prior to use.

2.4. Catalytic conversion

The catalytic conversion was carried out in a 20 mL glass vial containing 0.05 g of raw
material after pretreatment, various kinds of catalysts and varying amounts of solvents for the
optimization of reaction conditions. After reaction at various temperatures for various
periods, the mixture was cooled down to room temperature, diluted into 25 mL water and
filtered with a 0.22 m syringe filter for products analysis.
2.5. Kinetics studies

For studying reaction kinetics, the reaction was carried out at three different temperatures (80,
100, and 120 C) with [BMIM]Cl (5 g) as the solvent and CrCl36H2O with the concentration
of 20 mol% were used as the catalyst. At different time intervals, aliquots (0.5 mL) was
withdrawn from the reaction solution, diluted into 25 mL, filtered and then analyzed with
HPLC.
2.6. Analysis of products

HPLC equipped with a refractive index detector (HPLC-RI) and a Shodex Asahipark NH2P50 4E column was employed for the analysis of the amounts of glucose and fructose present
in solution. The mobile phase was an acetonitrile:water (70:30) solution at 35 C with a flow
rate of 1 mL min1. The concentration of HMF in solution was quantifed using an ICECoregel 87H3 column (operated at 35 C). The mobile phase was 8 mM sulfuric acid equeous
solution at a flow rate of 0.6 mL min1. The conversion is calculated using the following
formula, where [] is concentration.
Because the production of HMF is mainly the result of the dehydration of monosaccharides
and polysaccharides, acidic catalysts, either Brnsted or Lewis acids, are considered useful
for this reaction [11]. Here, we try six different types of acidic catalysts including CrCl3,
CrCl36H2O, CrCl2, H3BO3, amberlyst 15 hydrogen form, and phosphotungstic acid hydrate
to catalyze HMF production. To avoid the complexity of different feedstocks and
pretreatment methods, here we chose wood chips as the feedstock, and the wood chips were
mixed with 0.9% (w/w) dilute sulfuric acid solution at 110 C. Then, the pre-soaked wood
chips were continuously transferred to a high-pressure reactor that preceded the acidcatalyzed steam explosion at 185 C for 1 min [41]. The composition of the resulting solution
(pretreated wood chip) contained 14.7 wt% of glucose.
The reaction was carried out inthe presence of [BMIM]Cl (1 g), a fixed amount of catalysts
(i.e., 0.02 g) and the pre-treated biomass solution (1 mL) at 100 C for 6 h. As shown in Fig.
1(a), excluding the case without use of catalysts, there were conversion of glucose ranging
from 19 to 94 mol% for the six cases with the presence of catalysts. However, HMF could
only be found when using CrCl36H2O, amberlyst-15, and H3PW12O40 as catalysts, with the
corresponding yields being 35, 25, and 13 mol%, respectively. Remarkably, CrCl36H2O
demonstrated good catalytic activity for producing HMF. This result was in a good agreement
with the previous study by the Zhang et al. [30] who showed that Cr3+-based catalyst
exhibited higher activity and selectivity toward HMF than Cr2+-based catalyst because Cr3+
has the higher Lewis acidity as compared to Cr2+. Moreover, we also confirmed that Lewis

acids could assist glucose-to-fructose isomerization and fructose-to-HMF dehydration

[42] and [43].
Besides the effectiveness of catalysts, solvents also played a principal role in the yield and
selectivity of HMF [1] and [44]. Here we use six kinds of solvents including organic solvents
(i.e., DMA, DMF, and DMSO) and ionic liquids (i.e., [BMIM]Cl), [EMIM]Cl), and [Epy]Cl).
We then fixed the amounts of the catalyst CrCl36H2O (0.02 g), the pre-treated biomass
solution (1 mL), and solvents (10 mL for organic solvents and 1 g for ionic liquids) at 100 C
for 6 h. As shown in Fig. 1b, interestingly, no HMF was formed when using the organic
solvent systems, even when conversions of glucose were found in the range among 21
48 mol%. We suggested that it due to the presence of water in the pre-treated biomass
solution. Dam et al. have shown that the presence of water in the processes promotes the
hydrolysis of glucose but leads to the formation of undesired by-products, especially levulinic
acid and formic acid [14]. Additionally, the Qi group also reported that it would significantly
reduce HMF when there was more than 5 wt% of water in the reaction system [45]. It should
be noted that the water content in our pre-treated biomass solution was 85.3 wt%, so it is
reasonable that organic solvents could not work here. In contrast, the ionic liquid [BMIM]Cl
exhibited the highest HMF yield, up to 35 mol%, as compared to the cases of [EMIM]Cl
(15 mol%) and [Epy]Cl (11 mol%). According to the results above, we decide to use
CrCl36H2O as the catalyst and [BMIM]Cl as the solvent for further studies. Several previous
reports have also showed that CrCl36H2O and [BMIM]Cl were the best catalyst and solvent
for generating HMF from glucose, sucrose, and cellulose [20], [32] and [46].
With such results in hand, we then investigated the effect of the amounts of CrCl36H2O on
the HMF yield. Other conditions such as the amounts of [BMIM]Cl (1 g) and pre-treated
biomass solution (1 mL) were fixed while varying amounts of CrCl36H2O from 0.02 to
0.05 g were used for the HMF production at 100 C for 6 h. As shown in Fig. 1c, the
conversion increased from 50% to 100% as the amout of CrCl36H2O was increased from
0.02 g to 0.05 g, respectively, indicating more acidic sites in the catalyst would help the
hydrolysis of glucose. Similarly, HMF yield also gradually increased from 35% to 49% as the
CrCl36H2O amount was increased. However, the HMF yield seemed to reach a maximum
when the amount of CrCl36H2O exceeded 0.04 g, even as the conversion increased. It has
been reported that a high concentration of acidic catalyst would promote the conversion of
glucose to levulinic acid and formic acid [47]. Consequently, the optimized amount of
catalyst was decided to be 0.04 g.
3.2. Study of reaction kinetics

To study the reaction kinetics of HMF production, we investigate the influence of reaction
temperatures (i.e., at 80 C, 100 C and 120 C) on glucose conversion and HMF production.
As shown in Fig. 2, it clearly indicates that both glucose conversion and HMF production
increased as the reaction temperature was increased from 80 C to 120 C with a fixed
reaction time. The maximum conversion and HMF yield was found to occur at a reaction
temperature of 120 C. At this case, both of the glucose conversion and HMF yield reached
their maximums (i.e., 100% and 58.5%, respectively) at 6 h. The glucose conversion

remained at almost 100% but the HMF yield started to decrease when the reaction time was
over 8 h. The results obtained in this study are consistent with previous studies. Qi et al. [34]
examined the conversion of glucose in the range of temperatures between 90 C to 140 C,
and they also found that the HMF yield would increase with the increase of temperature up to
120 C. Other studies also found that longer reaction time periods would reduce the HMF
yield due to the occurrence of the side reactions of HMF [45], [47] and [48].
In general, the conversion of raw lignocellulosic materials to HMF has resulted in
low HMF yield due to the recalcitrant and heterogeneous features of the raw
materials [51] and [52]. Hence, a proper pre-treatment to decompose the rigid
structure of raw materials is indispensable [53] and [54]. In this study, the raw
lignocellulosic biomass (i.e., rice straw and wood chops) was pretreated in an
either dilute acid (H2SO4, HCl) or alkaline (i.e., NaOH) solution with a
concentration of either 1 or 3 wt% at 60 C for 24 h. After the pretreatment, the
catalytic conversion was performed at the reaction conditions as optimized
previously (i.e., [BMIM]Cl as the solvent, CrCl 36H2O as the catalyst, 120 C as the
reaction temperature, and 2 h as reaction time period). HMF yields were not
significantly different between wood and straw at the same reaction conditions.
As shown in Table 1, given no catalyst in the reaction system, we found that
there were still HMF yields for wood (i.e., 41%) and rice straw (i.e., 35%), owing
to the catalytic ability of CrCl36H2O catalyst in [BMIM]Cl. However, as expected,
higher HMF yields could be obtained from the samples with pre-treatment. At the
same reaction conditions (i.e., room temperature for 24 h), the HMF yields of
samples with the pretreatment were around 50% that was 10% higher than
those of samples without pretreatment. In addition, we found that there was no
obvious difference of HMF yields when using either acids (either H 2SO4 or HCl) or
base (NaOH) for pretreatment, although higher concentrations (i.e., 3%) of these
acids and bases did slightly enhance the HMF yields.

When the reaction temperature was increased to 60 C, the HMF yields further increased for
both raw biomass materials with pretreatments. Remarkably, wood chops and rice straw
pretreated with 3% NaOH and catalyzed under the conditions of 60 C for 24 h with the
presence of CrCl36H2O exhibited the highest HMF yields (i.e., 79 and 76%, respectively). To
the best of our knowledge, these values are the highest among all conversions of raw
lignocellulosic biomass materials. Previous literatures have reported that pretreatments of raw
biomass materials could remove a certain part of hemicellulose, thus efficiently increasing the
swelling of cellulose [54]. In addition, Ibrahim et al. has shown that alkaline pretreatments
could lead to the degradation of ester and glycosidic side chains. Particularly, alkaline could
cause structural alteration of lignin, cellulose swelling, partial de-crystallization of cellulose,
and partial solvation of hemicellulose [55]. However, previous pretreatments were all
conducted at higher temperatures (>120 C) [54]. In this study, we first demonstrate that rice
straw and wood chop could be effectively converted to HMF with over 75% yield after
pretreated at the relatively low temperature of 60 C for 24 h and then subsequently reacted at
120 C for 2 h with the assistance of CrCl36H2O as the catalyst and [BMIM]Cl as solvent.
4. Conclusion

We have demonstrated a combinational treatment including diluted acid/base pretreatments

and efficient catalytic conversion to produce HMF from raw lignocellulosic biomass. The
pretreatment conditions were optimized to be 3% NaOH at 60 C for 24 h, and the reaction
conditions were optimized to be [BMIM]Cl (1 g) as the solvent and CrCl36H2O (0.04 g) as
the catalyst at 120 C for 6 h. Under the optimized conditions, we successfully produced
HMF from wood chops with a maximum yield of 79%, which is believed to be the highest
yield so far. Additionally, we also investigated the reaction kinetics and activation energy of
HMF production, and it is proved that the activation energy can be reduced when an efficient
catalyst such as CrCl36H2O is used. The results obtained in this study would be useful for
producing HMF with an enhanced yield over 75%, and a subsequent conversion of HMF to
other useful fine chemicals can be highly expected.

Highlights

Non-pretreated catalysts showed better DMF selectivity than pretreated ones.

CuZr catalyst presented the best DMF selectivity probably due to the neutral nature of
ZrO2 and Cu hydrogenation capacity.

The presence of impurities in a real feed reduced the DMF selectivity.

Pressure increase improved H2 mass transfer, resulting in higher DMF selectivity.

Lower space velocity increased DMF selectivity

Abstract

Catalytic hydrogenolysis reaction of 5-hydroxymethylfurfural platform molecule to produce

2,5-dimethylfuran conversion was studied. For that purpose noble (Pt and Ru) and non-noble
(Ni and Cu) metal catalysts supported on acid (HYAl2O3 and Al2O3) and basic (ZrO2 and
TiO2) supports were used. All of the tested catalysts were able to convert completely HMF.
However, among the mentioned catalysts, the Cu catalyst supported on ZrO2 showed the best
behavior in terms of DMF selectivity, probably due to the neutral nature associated to ZrO2
support. Moreover, this catalyst was studied in order to know the influence of some reaction
parameters on DMF selectivity. As results obtained with CuZr catalyst concluded, a
temperature increase had not influence on the aforementioned parameter because the reaction
is exothermic. However, the type of feed, the increment of the pressure and the space velocity
decrease improved the DMF selectivity.
1. Introduction

Nowadays fossil resources, such as petroleum, coal and natural gas, are the most important
suppliers to world energy system, around 80% of the primary energy consumption
[1] and [2]. This fact involves serious problems related to the concerns about depletion of this
kind of energies [2], [3], [4] and [5], energy security and economics due to their centralized
production, and greenhouse emissions, which are related with the increase of global warming
[5] and [6]. However, it is important to remark that petroleum plays dominant role in
chemical industry because of its use in a wide range of products including plastics, carpets,
curtains, wall paints, varnishes, soaps, perfumes and hairsprays [7], and in transportation
sector in which properties of the used fuels are fulfilled by non-renewable petroleum-derived
liquids fuels [8]. In order to reduce the problems derived from uses of petroleum, searching
and developing of sustainable and environmental friendly energy sources have become
essential topics to achieve.
According to the literature [1] and [9], the biorefinery, and therefore highly abundant and
carbon-neutral biomass [4], can be a good alternative to the current needs of energy,
chemicals and unconventional fuels. Regarding to the first generation biofuels, biodiesel and
bioethanol, they had a great introduction on transportation sector until a few years ago.
However, their production from food crops came in conflicts with food industry [10].
Moreover, these biofuels present several drawbacks that limit their use as transportation fuels,
such as low oxidation stability, corrosive nature and poor cold flow properties [11]. This fact
leads to the development of a second generation of biofuels or fuels bioadditives, which are
derived from non-food residual lignocellulosic biomass [4], [10] and [11].
Although there is a wide variety of chemical building blocks obtained from lignocellulosic
biomass, 5-hydroxymethylfurfural (HMF), which is also considered chemical platform,
seems to be the most promising intermediate as it can be produced from different raw
materials (glucose, fructose, sucrose), and it can also be converted into biofuels, such as 2,5dimethylfuran (DMF), and other fine chemical molecules (levulinic acid (LA), 2,5furandicarboxilic acid (FDCA), ethyl levulinate (EL)) [1], [3], [8], [9] and [12]. The
production of HMF dates back to 19th century when the HMF was separated from reaction

mixture of fructose, sucrose and oxalic acid [12]. Nowadays, the most extended reaction to
produce HMF is dehydration of the aforementioned carbohydrates over acid, homogeneous or
heterogeneous catalysts [1], [8], [9] and [12], because its high volume production from
petroleum derived feedstocks suppose a great cost and therefore a low availability [13].
Regarding the DMF, it is produced from selective hydrogenation of HMF [11] under H2
atmosphere, also called hydrogenolysis process [2] and [8]. This reaction involves different
pathways [1], [11] and [14] in which dehydration and hydrogenation reactions are
predominant. These pathways are shown in Fig. 1. Nevertheless, according to the literature
[2], [15] and [16], apart from dehydration and hydrogenation reactions, secondary reactions
could occur during the HMF conversion, such as demethylation and decarbonylation. Fig. 2
shows a complete reaction scheme of HMF conversion, which have been adapted from the
aforementioned literature.
DMF high-quality fuel is considered, as ethanol, an ideal renewable and sustainable substitute
or additive of the conventional gasoline [2], because of its high energy density, similar to that
of gasoline and higher than ethanol [3] and [11], and research octane number, higher than
gasoline [1]. Moreover, DMF shows very low solubility in water and therefore it can be used
as a blender in transportation fuels [11] and [16]. In this sense, some literature reports the
good performance of the DMF as a fuel on direct injection spark ignition (DISI) type engines
[3] and [6] without important modifications of the engine.
As reported in the recent literature about this topic, DMF production from HMF catalytic
hydrogenolysis is mainly carried out by autoclave systems [1], [2], [3], [8], [14],
[16] and [17], it means, in discontinuous systems, using synthetic HMF as reactant and noble,
such as Pd, Pt, Ru, Rh and Au [1], [2], [14], [16] and [17], and non-noble, such as Cu, Co, Ni
and Fe [3], metal-based catalysts. Only few research works report information about
hydrogenolysis of HMF over continuous systems [8] and [18] using real feed of HMF [18].
Based on this background, the aim of this work was to compare the activity toward the
production of 2,5-DMF from 5-HMF of catalysts containing noble and non-noble metals
supported on carriers showing different degrees of acidity and basicity.
2. Experimental procedure
2.1. DMF production from HMF hydrogenolysis

Two types of feeds were used in the DMF production designed as real and synthetic
HMF. The first one was obtained from synthetic fructose (Sigma-Aldrich, 99%) dehydration
over Amberlyst 70 catalyst in autoclave reactor, where an organic phase containing around
0.7 wt% of HMF in 1-butanol was obtained. Taking into account this result, the second feed
was prepared using 0.7 wt% of synthetic HMF (Sigma-Aldrich, 99%) diluted in 1-butanol.
This compound was selected as most of the studies about DMF production reported in the
literature [8], [16], [17] and [18] used it as a solvent agent.

The gas phase hydrogenolysis of the different HMF feeds was carried out in a continuous
bench-scale fixed bed reactor in which 0.2 and 0.5 g of catalyst were diluted with inert SiC
(catalyst/SiC = 1:9 wt), corresponding to WHSV (g of HMF/g of catalyst h) of 0.12 h1 and
0.06 h1, respectively. Prior to hydrogenolysis reaction, all tested catalysts were pretreated
and non-pretreated. The pre-treatment were performed in-situ under 100 Nml/min of H2 at
0.1 MPa and 400 C for 2 h.
HMF conversion was carried out for 4 h under a feed flow of 0.1 ml/min, temperatures of 200
and 225 C and pressures from 0.5 to 3 MPa. These pressures were achieved using
37 Nml/min of pure H2. The feed and output liquid streams were collected and analyzed by
off-line high-performance liquid chromatography (HPLC) and gas chromatography (GC).
Concretely, HMF quantity on feed and liquid product stream was analyzed by an Agilent
HPLC chromatograph (1260 Infinity) equipped with Hi-Plex H column and infrared detector,
and DMF and other products on the liquid product stream were quantified using a Hewlet
Packard GC equipped with Suprawax 280 capillary column.
For better understanding of catalytic activity and product distribution, parameters such as
HMF conversion (%) and DMF selectivity (%) were calculated.
2.2. Catalysts preparation

The selected catalysts to carry out HMF hydrogenolysis were noble and non-noble metal
catalysts supported on acid (Al2O3 designed as Al) and neutral-basic (TiO2 and ZrO2 designed
as Ti and Zr, respectively) supports. The noble metal catalysts were PtHYAl, RuTi and
RuCuTi. The Pt catalyst was prepared by wet impregnation of a physical mixture of HY
zeolite and -Al2O3 (Girdler Sudchemie) [19]. This HYAl support was prepared by
suspension of HY (Conteka) and -Al2O3 powder. The Ru and RuCu catalysts were prepared
by incipient wetness impregnation [20]. In particular, 2Ru5CuTi catalyst was prepared by
sequential impregnation of TiO2 (Degussa, Aeroxide titanedioxide P25) with copper
(Cu(NO3)23H2O (Alfa Aesar, 98%)) precursor salt. After drying, the solid was impregnated
with RuCl3 (Johnson Matthey, metallic phase Ru 40%) precursor salt and then calcined at
conditions reported in the literature [20]. The nominal content for the aforementioned metal
catalyst was 1 wt% of Pt, 2 wt% of Ru and 5 wt% of Cu. Regarding non-noble catalysts,
these were designed as NiAl, CuAl and CuZr. The NiAl and CuAl catalysts were prepared
following the procedure reported in the literature [21] and [22] in order to achieve nominal
contents of 35 wt% of Ni and Cu. In the case of the CuZr catalyst, this catalyst was prepared
by co-precipitation of aqueous solution of Cu(NO3)25H2O and ZrO2(NO3)xH2O salt
precursors. The amount of metal precursor was selected to achieve a content of
approximately 30 wt% of Cu. Then the solution was stirred at 500 rpm, meanwhile NH4OH
aqueous solution (excess of 20%) was added in order to achieve a pH above 9, followed by
aging, vacuum filtering and washing at pH around 7. Finally, catalyst was dried at 110 C
overnight and calcined at 300 C for 1 h.

2.3. Catalysts characterization

The calcined catalysts were characterized by several physic-chemical techniques such as

plasma atomic emission spectroscopy (ICP-AES), surface area (BET method), temperatureprogrammed reduction (H2-TPR), temperature-programmed desorption of NH3 (NH3-TPD)
and temperature-programmed desorption of CO2 (CO2-TPD). Analyses conditions of these
techniques and the used equipments for characterization of calcined PtHYAl, NiAl, CuAl and
RuTi are published elsewhere [19], [20], [21] and [22]. In the case of the CuZr and RuCuTi
catalysts, the characterization equipment and conditions are explained below.
2.3.1. Plasma atomic emission spectroscopy

Metal contents were determined by ICP-AES, using a Perkin-Elmer Optima 3300DV

apparatus, previous dissolution of the ground samples in acid solutions.
2.3.2. Textural properties

BET surface area and pore characteristics of the calcined fresh catalysts were evaluated from
the N2 adsorptiondesorption isotherms obtained at 77 K over the whole range of relative
pressures. In the case of the CuZr catalyst, this sample was characterized using Autosorb-1C/TCD (Quantachrome, USA) after outgassing solid samples at 300 C for 3 h.
2.3.3. H2-TPR analysis

This technique was used to study the reducibility of the calcined catalysts. The measurements
were carried out using an AutoChem II instrument (Micromeritics, USA) equipped with a
TCD detector. TPR profiles were obtained by heating the samples from room temperature to
900 C at a linearly-programmed rate of 10 C/min, while 50 ml/min of reduction gas (5%
v/v H2 diluted in Ar) was passed through the sample.
2.3.4. NH3-TPD analysis

Temperature programmed desorption of NH3 was used in order to know the acid
characteristics of the calcined samples. The used equipment was AutoChem II instrument.
First of all, samples were flushed by He at 250 C for 30 min, followed by cooling at 40 C
and loading of NH3 for 30 min. Then the physically absorbed NH3 was removed using He at
85 C until no further desorption was recorded and release of chemically adsorbed NH3 was
collected increasing temperature from 85 C to 800 C at a rate of 10 C/min.
2.3.5. CO2-TPD analysis

The basicity of calcined samples were determined by temperature programmed desorption of

CO2 probe molecule. The CO2-TPD measurements were carried out also with AutoChem II
instrument and the same procedure as NH3-TPD. First of all, samples were flushed by He at
250 C for 30 min, then cooled down to 40 C and loaded with CO2 for 30 min. Then the
physically absorbed CO2 was removed using He at the same temperature until no further
desorption was recorded. Finally, the samples were heated from 40 C to 600 C at a rate of
10 C/min, while the release of CO2 data was collected.

2.3.6. SEM and TEM analysis

The morphological appearance of the fresh and used catalysts was studied by SEM analysis
in a JEOL JSM-6400 with W filament and a resolution of 3.5 nm, and TEM images were
obtained in a Philips SuperTwin CM200 apparatus operated at 200 kV and equipped with
LaB6 filament and EDAX EDS microanalysis system. The fresh and used samples were
prepared via dispersion into ethanol solvent and placed on a carbon-coated copper grid (300
Mesh) followed by drying under vacuum.
3. Results and discussion
3.1. Characterization results
3.1.1. Chemical and textural properties of catalysts

Table 1 shows the metal content measured by ICP-AES and the specific surface area (SBET)
and pore volume data from N2 physisorption for the calcined catalysts. The PtHYAl and NiAl
catalysts showed experimental metal loads close to the nominal ones [19], [20], [21] and [22].
In the case of the CuAl catalyst, the active metal content was slightly lower than the
theoretical one. The same phenomenon was also detected for CuZr catalyst, because the
measured load was 23.7 wt%, 5 points lower than the theoretical one. In the case of RuTi and
RuCuTi catalysts, the measured Ru content was around 0.07 wt%, while in the case of Cu,
the metal loading was similar to the theoretical one.