Hydroconversion of xylose derived ketals: A key strategy for producing A broad range of green-hydrocarbons suitable as fuels and petrochemicals

Article abstract of DOI:10.1016/j.apcata.2020.117911

Recently, sugarcane bagasse was transformed by ketalization reaction into a biocrude with physical-chemical properties adequate for further conversion in typical refining processes. Herein, a representative compound of this biocrude, 1,2:3,5-di-O-isopropylidene-α-D-xylofuranose, DX, in n-hexane was hydro processed in the presence of several palladium-catalysts. Screening in a flow reactor andamong Pd supported on SiO₂, γ-Al₂O₃, HUSY, HY, HBEA and HZSM-5, Pd/HZSM-5 showed the least susceptibility to coke deposition and good deoxygenation capacity, yielding significant amounts of hydrocarbons containing seven or more carbons. Additionally, in a batch and milder conditions, this catalyst was highly selective towards hydrocarbons with high carbon yield and remarkably low coke, i.e., 86 % and <5 % of carbon from DX, respectively. Water was the major product of deoxygenation. The hydrocarbon products included branched and linear alkanes, cycloalkanes and aromatics, whose distribution was sensitive to the catalyst properties, such as support's mesoporosity.

Full text of DOI:10.1016/j.apcata.2020.117911

Journal Pre-proof
Hydroconversion of xylose derived ketals: a key strategy for producing a
broad range of green-hydrocarbons suitable as fuels and petrochemicals
S e´ rgio Castro Pereira, Matheus Souza, Laura M. Esteves, Nuno
Batalha, Yiu Lau Lam, Marcelo Maciel Pereira
PII:
S0926-860X(20)30504-4
DOI:
https://doi.org/10.1016/j.apcata.2020.117911
APCATA 117911
Reference:
To appear in:
Applied Catalysis A, General
Received Date:
Revised Date:
Accepted Date:
25 August 2020
16 October 2020
1 November 2020
Please cite this article as: Pereira SC, Souza M, Esteves LM, Batalha N, Lam YL, Pereira MM,
Hydroconversion of xylose derived ketals: a key strategy for producing a broad range of
green-hydrocarbons suitable as fuels and petrochemicals, Applied Catalysis A, General
(2020), doi: https://doi.org/10.1016/j.apcata.2020.117911
This is a PDF file of an article that has undergone enhancements after acceptance, such as
the addition of a cover page and metadata, and formatting for readability, but it is not yet the
definitive version of record. This version will undergo additional copyediting, typesetting and
review before it is published in its final form, but we are providing this version to give early
visibility of the article. Please note that, during the production process, errors may be
discovered which could affect the content, and all legal disclaimers that apply to the journal
pertain.
©
2020 Published by Elsevier.

Hydroconversion of xylose derived ketals: a key strategy for
producing a broad range of green-hydrocarbons suitable as
fuels and petrochemicals.
1
2
3
2
Sérgio Castro Pereira , Matheus Souza , Laura M. Esteves , Nuno Batalha , Yiu Lau
Lam , Marcelo Maciel Pereira1*,2
2
1
Instituto de Química, Universidade do Estado do Rio de Janeiro, Rio de Janeiro, RJ,
Brazil- * maciel@iq.ufrj.br
2
Instituto de Química, Universidade Federal do Rio de Janeiro, Rio de Janeiro, RJ,
Brazil
3
Departamento de Engenharia Química e de Petróleo, Universidade Federal
Fluminense, Niterói-Rio de Janeiro, Brazil.
Graphical abstratc
Highlights


Hydro-conversion of sugar –acetal was explored in fixed and batch reactors.
Sugar-acetal showed remarkably high green-hydrocarbon yields.
1



Obtention of Green iso-paraffin, cyclo-alkanes, and aromatic hydrocarbon.
Pd/ZSM-5 was the best catalysts used.
Abstract
Recently, sugarcane bagasse was transformed by ketalization reaction into a
biocrude with physical-chemical properties adequate for further conversion in typical
refining processes. Herein, a representative compound of this biocrude, 1,2:3,5-di-O-
isopropylidene-α-D-xylofuranose, DX, in n-hexane was hydro processed in the
presence of several palladium-catalysts. Screening in a flow reactor andamong Pd
supported on SiO
2
, γ-Al
2
O , HUSY, HY, HBEA and HZSM-5, Pd/HZSM-5 showed the
3
least susceptibility to coke deposition and good deoxygenation capacity, yielding
significant amounts of hydrocarbons containing seven or more carbons. Additionally, in
a batch and milder conditions, this catalyst was highly selective towards hydrocarbons
with high carbon yield and remarkably low coke, i.e., 86 % and <5 % of carbon from
DX, respectively. Water was the major product of deoxygenation. The hydrocarbon
products included branched and linear alkanes, cycloalkanes and aromatics, whose
distribution was sensitive to the catalyst properties, such as support’s mesoporosity.
Keywords: green-hydrocarbons, hydro-conversion, ZSM-5, palladium, biomass, sugar.
1
Introduction
The massive use of fossil fuels for hydrocarbons and energy generation led to
several concerns in the last decades. Beyond environmental and health hazards, the
depletion of petroleum sources and unstable oil prices, have negatively affected
economies worldwide. Global population growth and the increasing energy demand
required for the many human beings activities are at the center of this problem.
2

Transportation is one of the activities dependent on fossil fuels, representing a large
consumption of primary energy, approximately 28% [1]. Nowadays, renewable energy
represents a relatively small portion of the total energy demand, particularly comparing
to the photosynthesis potential [2, 3]. Thus, biofuels production from biomass is a
promising alternative that can reduce carbon footprint [4].
Among all types of biomass, lignocellulosic biomass is emerging as the main
source for biofuels production. Lignocellulosic biomass is the most abundant renewable
source on earth, is available worldwide, and can be be a carbon-neutral source [5]. The
great complexity of lignocellulosic biomass structure has presented a significant
challenge to developing processes capable of promoting its conversion to chemicals
and fuels [6, 7]. Fast pyrolysis is the most commonly used technique for the production
of second-generation biofuels [8]. Despite the simple structure and flexibility to process
different biomass sources, biomass pyrolysis produces a bio-oil with several negative
points, including high instability, acidity and oxygen contents, making its transformation
into hydrocarbons challenging.
The removal of heteroatoms, like oxygen, from feedstocks, is often achieved by
hydrotreating processes or hydroprocessing (HDP) [5, 9]. For instance, catalytic
upgrading for oxygen removal has been carried out in pyrolysis bio-oil [10], in
oxygenated compounds pure or mixed with hydrocarbons [11, 12]. The catalysts used
for bio-oil hydroconversion are quite diverse. However, it has been proven that noble
metal catalysts like (Ru, Rh, Pt and Pd) present a better performance regarding oil
yield and hydrodeoxygenation degree [13]. γ-alumina is the most commonly used
support for hydroprocessing catalysts. However, in recent years, different porous
materials such as activated carbon and zeolites have been studied. These catalysts
have already shown potential in various processes of refining and hydroprocessing [14-
16]. Zeolites are used in several processes due to their high specific area, adsorption
capacity and the possibility of tailor-making the number and type of acid sites and
textural properties. Also, the pore structures (the size and connectivity of channels and
3

cavities) can control the conversion of molecules involved in the various chemical
processes [13].
An alternative approach to biofuel production based on second-generation
biomass was proposed in two steps [17]. Firstly, sugarcane bagasse was successfully
transformed into a bio-crude by combining acid-catalyzed hydrolysis with organic
reactions, like ketalization and acetylation under mild reaction conditions, i.e.
temperature between 60-150 °C and pressure up to 12 bar [18-21]. The biocrude,
produced from the ketalization of the sugarcane cellulose and hemicellulose sugars
with acetone, is mainly composed of isopropyl-ketals, such as 1,2:3,5 -di-O-
isopropylidene-α-D-xylofuranose
(DX)
and
1,2:5,6-
di-O-isopropylidene-α-D-
glucofuranose (DG), and represented up to 40 wt.% of the total composition [20]. Our
group has recently developed a method for full biomass conversion into biocrude using
a continuous flow reactor [22]. This method proposed that first hemicellulose is
converted, followed by amorphous and crystalline cellulose with 25 substances being
found, incluiding and molecules with up to five O-isopropylidene units such as
Tetrapentose pentacetonide – m/z 764.37 were identified. Additionally, we
demonstrated that this type of biocrude and its representative components, DX and
DG, can be converted into green hydrocarbon using simple catalytic tests. Further
study, with 13C labeled DX revealed a significant contribution of carbohydrates to the
hydrocarbons formation [23]. Indeed, a mixture of DX mixture in n-hexane up to 50
wt.% was recently converted into aromatics in fluidized catalytic cracking (FCC), and
carbon from DX conversion into hydrocarbons was estimated to be higher than 70%
[
24-26].
Herein, DX was used as a representative compound of this new biocrude. DX
was mixed and processed with n-hexane under hydroprocessing conditions. Pd
supported on γ-alumina, silica-alumina and zeolites (HUSY, HY, HBEA, HZSM-5) were
first compared using a fixed-bed flow reactor (350 ºC, 50 bar H
2
). Then, the Pd/HZSM-
5
was studied in further detail in a batch reactor at different reaction conditions (250 ºC,
4

4
0 bar H ). The obtained results were also compared with those obtained with
2
mesoporous ZSM-5 support. n-hexane hydroconversion under similar conditions was
carried out for comparison.
2
Experimental
2.1 Catalyst preparation
The γ-alumina, silica, and the zeolites (Y, USY, BEA and ZSM-5) used as
catalyst supports were supplied by Petrobras. The acid form of the zeolites was
obtained by ion exchange using a (NH SO 1M solution at 70 ºC for 1 h in a 50 mL/g
4
)
2
4
zeolite proportion. This procedure was performed four times. After the ionic exchange,
the samples were dried at 120 ºC overnight and then calcined in air at 450 ºC for 4 h.
Pd was introduced into the support by incipient wetness impregnation using
Pd(NO
3
)
2
.2H O (Sigma-Aldrich) as a precursor. The amount of precursor was
2
estimated to obtain 0.5 wt% of Pd. Then, the samples were dried at 120 °C overnight
and calcined in air at 450 °C for 3 h. The catalysts were named Pd/γ-Al
2
O
3
, Pd/SiO ,
2
Pd/HUSY, Pd/HY, Pd/HBEA and Pd/HZSM-5.
ZSM-5 (SAR = 41) was treated by an alkaline treatment using a NaOH 0,5 M
solution. The mixture (ZSM-5 and NaOH) was heated until 70 ºC and remained at this
temperature for 20 min under agitation. Then, the sample was filtered and washed with
(
NH₄)₂SO₄ 1M solution. After this step, the sample was washed with deionized water at
0ºC. Finally, the sample was dried at 120 ºC overnight and calcined at 450 ºC for 3 h.
6
Based on the zeolite yield after alkaline treatment, the SAR was estimated to be 14.
The catalyst was named Pd/m-HZSM-5.
2.2. Catalyst characterization
Nitrogen adsorption and desorption process were carried out in a Micromeritics
ASAP 2020 apparatus. Firstly, the catalysts were pretreated at 220 °C under vacuum
for 4 h. After that, the nitrogen adsorption-desorption isotherms were performed at -196
5

ºC. The specific surface area (SBET) and external surface area (Sext) were determined
by the Brunauer–Emmett–Teller (BET) and t-plot method, respectively. The total
volume of pores was determined at 0.98 P/P
method.
0
and the micropore volume by the t-plot
X-ray powder diffraction (XRD) patterns were obtained using a Rigaku X-ray
diffractometer with a nickel-filtered CuKα1 radiation source (λ= 0.15406 nm) and a
graphite monochromator. XRD profiles were collected by scanning with 0.02°
increments (counting 1 s between each increment) in the 2θ angle region between 5°
and 80°.
Energy dispersive X-ray spectrometry (EDX) analyses were performed on a
Shimadzu EDX-720 spectrometer under vacuum (<30 Pa) to evaluate the chemical
composition of the calcined palladium catalysts. Also for these catalysts, SAR is
represented by SiO
2
/Al
2
O molar ratio, which divided by two gives the Si/Al molar ratio.
3
X-ray photoelectron spectroscopy (XPS) analysis was performed with a Thermo
Scientific ESCALAB 250Xi spectrometer (Al-Kα X-ray source, 1486,7ꢀeV). During
-8
analysis, the pressure was kept around 10 ꢀmbar inside the analysis chamber.
Charging effects were minimized with low energy electron flood gun. XPS data were
calibrated using the Cꢀ1s peak (284.6ꢀeV) as reference.
NH -TPD analysis was performed on a Micromeritics AutoChem II 2920
3
equipment. Before the analysis, the samples were dried at 150 °C under He flow and
reduced at 450 °C for 2 hours. The samples were then purged with He at 450 °C and
then cooled to 70 °C to adsorb NH
by purging with He. NH desorption profile was carried out at a constant flow of He from
0 °C to 900 °C with a heating rate of 10 °C.min .
The coke loading in the spent catalysts was determined by thermogravimetric
analysis in a Netzsch TG-IRIS apparatus. The samples were dried under nitrogen flow
3
. The excess NH
3
in the the samples was removed
3
-1
7
-1
at 250 °C for 30 min at a heating rate of 10 °C.min . Afterward, the sample was heated
6

-1
1
0 °C.min up to 700 °C under airflow. The coke loading corresponded to the sample
weight loss under airflow.
2.3 Hydroprocessing
Firstly, hydroprocessing was performed in a fixed-bed reactor. For each test, 1
g of catalyst was loaded into the reactor. Before reaction, the catalyst was reduced by
-1
flowing H
2
(100 mL.min ) heating from room temperature to 450 ºC at a rate of 10
-1
ºC.min and remained at this temperature for 60 min. The reaction was performed for 2
-1
h at 350 ºC under 50 bar of hydrogen pressure, using hydrogen flow (100 mL.min )
measured at the condenser exit and atmospheric pressure. DX liquid, 20 wt.% in n-
hexane, was admitted with a flow rate of 0.1 mL.min-1 using an HPLC pump. For
comparison, pure n-hexane hydroconversion was performed under the same
conditions for relevant catalysts.
The hydroconversion of DX in the batch reactor was performed in a stirred 450
mL autoclave fabricated by Parr Instrument Company. Before reaction, the catalyst
was reduced at 300 ºC for 3 h with an H
desired amount of DX (15 to 30 g) and 90 g of n-hexane was admitted to the reactor.
The reactor was flushed three times with H and then filled with 40 bar H . The reactor
2
flow of 50 mL/min. Then, a mixture of the
2
2
temperature was increased from room temperature to 250 ºC while the feed was stirred
at a rate of 300 rpm. Finally, at this final temperature and pressure the reaction was
carried out for 24 h.
Two sets of DX/catalyst ratios were used. First, the amount of DX (in grams)
was explored (30, 20, and 15 g), keeping constant the mass of the catalyst (5 g). These
experiments were named as T30/5, T20/5, T15/5, respectively. Second, the DX mass
was fixed at 15 g and the catalyst amount varied (5, 10, and 15 g). Finally, we explored
the effect of improving the mesopores of the ZSM-5, preparing an additional catalyst.
Gas products were analyzed on-line during the fixed-bed reaction and at the
end of the hydroconversion reaction by using a MicroGC (Agilent Technologies
7

MicroGC 490). Liquid products, both organic and aqueous, were collected by
condensation (0 ºC), under pressure, with a condenser placed right after the reactor
exit. The liquid yield was obtained by the weight of the liquid fraction recovered in the
condenser. Then, the liquid phase was frozen at -20 ºC, and the frozen aqueous
fraction was measured by weighing after separation from the organic fraction. The
organic fraction was analyzed by using GC-MS (Agilent Technologies, 7890A and
5975C) for product identification and by using GC-FID (Agilent Technologies 7890A
coupled with FID detector) for product quantification, as described in Supplementary
Information. Selectivity was determined by wt.% of carbon. The same equipment
estimated oxygenates, their yield were underestimated as they derived from carbon
intensity without correction of the weight of oxygen in these components. Thus, values
represent an estimated amount. Still, for the the tests where no oxygenates were
observed, namely T15/15, the selectivity is in wt.%.
The amount of coke in the spent catalysts was determined by thermogravimetric
analysis (Netzsch TG-IRIS). The samples were heated from 308 to 523 K at a rate of
-1
1
0 °C min under N
2
atmosphere. The temperature was kept at 523 K for 30 min, after
in N ) and the
temperature increased to 973 K at a rate 10 K min and then maintained at 973 K for
0 min. The amount of coke in the catalyst corresponded to the weight loss at
temperatures higher than 523 K. The coke yield was estimated comparing this value
considering the amount of catalyst in each reaction) to the total feed introduced into
the reactor.
Conversion, yield, and selectivity were estimated by the equation 1, 2 and 3,
respectively.
which the atmosphere was changed to synthetic air (20.9% O
2
2
-1
3
(
(
1)
(
2)
8

(3)
Relative selectivity to liquid and gas phase, when applied, is defined by
equation 4.
(4)
2.4 Synthesis of 1,2:3,5-di-O-isopropylidene-α-D-xylofuranose (DX)
DX used in the reaction was synthesized as described elsewhere [12][23].
Further informations are presented in Supplementary Information. After the synthesis,
pure DX was dissolved in n-hexane to obtain the desired mixture used in the reactions.
The purity of the final mixture was verified after each synthesis by using GC-MS. It is
important to observe that DX can be produced from biomass as recently
demonstrated[22].
Results and discussion
3.1 Catalyst support screening in a flow reactor
o
A mixture of 20 wt.% of DX in n-hexane was converted at 350 C and 50 bar of
hydrogen in the presence of Pd/γ-Al , Pd/SiO , HZSM-5 + Pd/SiO , Pd/HUSY and
2
O
3
2
2
Pd/HZSM-5. The coke and liquid yield and produtct selectivity in the liquid phase (by
carbon intensity, %) are presented in Table 1. It was difficult to quantify the gas
9

products under continuous flow, and consequently, a good weight balance could not be
attained. Thus, this fraction was not considered. Nevertheless, it was considered that
the weight balance could be completed by the gas collected and the light compounds
dissolved in the liquid phase, which vaporized during the collection of the liquid from
the condenser and were not captured.
Table 1: Hydroconversion of 20% DX in n-hexane with Pd based catalysts in a fixed
bed reactor. Coke and liquid yield (wt.% of the feed) and distribution of the products in
the organic liquid phase (by carbon intensity), A6- are branched hydrocarbons
containing six carbons and hydrocarbon with 5 carbons or less, CyA6+ are naphthenics
containing cyclopentane and cyclohexane with 6 or more carbons, A7+ are
hydrocarbons with 7 carbons or more.
HZSM-5 +
2
Pd/γ-Al O
3
Pd/SiO
2
Pd/HUSY
Pd/HZSM-5
Pd/SiO
2
Yield (wt.%)
19.4
69.9
Coke
Organic Liquid
Aqueous phase
11.6
62.8
16.3
76.4
4.1
66.9
3.2
0.8
76.2
1.0
Selectivity in the liquid phase (organic compounds in carbon intensity wt.%)
−
6.2
0.4
−
1.2
−
−
18.7
3.8
A
6-
74.4
16.3
66.0
23.5
CyA6+ + A7+
Oxigenates and
non-identified
93.4
98.8
77.5
3.2
8.1
When γ-Al
2
O
3
and SiO were used as support a considerable amount of coke
2
was produced. Oxygenates was the major product and hydrocarbons were formed in
very small amounts in the liquid phase. With a mixture of HZSM-5 + Pd/SiO catalysts
2
oxygenates slightly decreased, and a corresponding increase in hydrocarbons was
observed. Also hydrocarbons containing seven and more carbons were observed.
Coke yield remained high. Contrarily, when Pd/HUSY and Pd/HZSM-5 catalysts were
used coke yield and oxygenates in the liquid phase drastically decreased. Notably,
hydrocarbons were the primary products and a substantial increase of hydrocarbons
containing seven or more carbons was observed. Besides, an aqueous phase was
observed on both tests.
1
0

After the initial tests, a screening of different zeolites supports in flow conditions
was carried out. This included previous tested catalysts Pd/HZSM-5 and Pd/HUSY and
also Pd/HBEA and Pd/HY. The catalysts were tested with 20% DX in n-hexane and
compared with pristine n-hexane at the same conditions. The results are presented in
Table 2.
Table 2: Hydroconversion of 20% DX in n-hexane with Pd/zeolite catalysts in a fixed
bed reactor. n-Hexane and DX conversion (%). Coke and liquid yield (wt.% of the
feed). Liquid products selectivity (wt.% of carbon). A5- are hydrocarbons with 5 carbons
or less, branched A are branched hydrocarbons containing six carbons, CyA6+ are
6
naphthenics containing cyclopentane and cyclohexane with 6 or more carbons, A7+ are
hydrocarbons with 7 carbons or more.
Pd/HUSY
Pd/HY
Pd/HBEA
Pd/HZSM-5
DX
+
nA
6
+ DX nA + DX nA
6
6
+ DX nA
6
DX conv. (%)
n-hexane conv. (%) 79
−
100
24
−
51
97
28
−
62
100
29
−
65
100
19
Yield (wt.%)
3.3 0.1 1.7
Coke
0.1
0.1
0.7
0.1
0.6
Organic Liquid
Aqueous phase
67.8 67.4 72.6 63.2 75.1 69.0 67.5 73.4
−
2.8
Selectivity (by carbon intensity in wt.%)
0.5 8.5 0.9 4.0 0.2
−
2.0
−
2.6
−
1.6
0.6 11.9
A
5-
6.0
branched A
6
58.8 3.9 45.6 0.7 58.3 18.3 46.7 4.85
CyA6+
−
−
−
−
2.1
1.4
0.2
−
−
−
−
−
1.2
0.8
0.3
−
−
−
−
−
2.4
1.7
0.3
−
−
−
−
−
3.7
2.1
0.8
0.1
A
7+
Oxygenates
Aromatics
All Pd/zeolite catalysts were active for n-hexane conversion. The activity
follows: USY > ZSM-5  BEA > Y. This sequence will be further discussed in the next
section of zeolite characterization. The presence of DX reduced the n-hexane
conversion in 70% for both Pd/HUSY and Pd/HZSM-5 and 50% for both Pd/HBEA and
Pd/HY. For all tests, DX was fully converted. Both the higher molecular weight and the
presence of reactive bonds such as the C-O links could be responsible for the more
successful occupation of the active sites (both acidic or metallic) by the DX molecule
and oxygen intermediates as compare to n-hexane. This could explain the decrease in
the activity of n-hexane with the presence of DX in the mixture.
1
1

The liquid fraction was the major fraction (yield between 65 and 75 wt.%) and a
very small yield in coke was again confirmed (between 0.1 and 3.3 wt.%). A clear
tendency on decreasing coke formation can be observed: USY > Y > BEA ZSM-5.
This tendency will also be discussed in the next section.
As shown in Table 2, an aqueous phase was observed when DX was present in
the feed. It was colorless with no organic products detected by GC-MS. The phase’s
amount was slightly superior for tests using Pd/HUSY and Pd/HBEA catalysts.
Remarkably, both Pd/HUSY and Pd/HBEA also produced lower oxygenates than
Pd/HY and Pd/HZSM-5. Nonetheless, oxygenates were produced only in a small
amount in all cases. These intermediate products resulted from the incomplete
conversion of DX into hydrocarbons, and probably the smaller pores of Pd/ZSM-5 and
the lower activity Pd/HY led to less complete conversion.
The main products of pristine n-hexane were isomers and followed by a small
amount of light products. Usually the overall selectivities to isomerization of n-hexane
in hydro-isomerization in the presence of metal-zeolite catalysts are very high [27-29].
Yet, both cracking by -scission and hydrogenolysis are unfavorable under the reaction
conditions [30]. In the presence of DX, light hydrocarbons (A5-), increased. Remarkably,
cycloalkanes with six or more carbons and alkanes with 7 or more carbons were
produced. The amount of these compounds obtained were between 13.9 and 29 wt.%
of the liquid product which indicatesas a first supposition, DX is responsible for their
formation.
Most of A +
compounds contains seven and eigtht carbons (Supplementary
7
Table 2 and 3). Alternatively, CyA6+ compounds contained mainly hydrocarbons with
six, seven and eight carbons, except for Pd/HY, which showed high hydrocarbons with
nine carbons selectivity. Also, the highest amount alkanes and cycloalkanes (up to 10
and 11 carbon atoms, respectively) were observed in the presence of Pd/HY. This
suggests that this catalyst might be more suitable to produce high carbon chain
products, favored by the large pore zeolites and low acidity of Y zeolite [31].
1
2

Gas products produced in the flow reaction are methane, ethylene, ethane,
propane, propylene, butanes and pentanes, CO, and CO , but they are not quantified,
table S4. The CO/CO ratio was higher than one for metal zeolite catalysts. Also, the
CO + CO / methane ratio following Pd/USY> Pd/Beta > Pd/HY > Pd/ZSM-5. Water
2
2
2
yield follows the same order. These results suggest that deoxygenation is favored in
large porous size (like Pd/USY) and external accessibility (like Pd/Beta). But a lower
CO and CO formation would indicate preliminarily less carbon loss.
2
In summary, through these results, we observed that DX is readily fully
converted, indicating we can test milder conditions. However, the material weight
balance was not satisfactory, probably due to the loss of light products (the material
weight balances are 80% or less). The Pd/ZSM-5 catalyst was a promising catalyst due
to the lower coke deposition and higher yield of A7+ hydrocarbons. Also, when biomass
is converted into biocrude by acetalization reaction sugar-acetal derivative, up to 700
daltons are observed [22]. Hence, batch reaction conditions would be more suitable to
convert these large molecules than a fixed bed reactor. Thus, we would test Pd/ZSM-5
in a batch reactor and vary more testing conditions to obtain a better insight into the DX
transformation.
3.2 Zeolite suppoted catalyst: characterizations suggested interpretations
For better interpretations of the tests results, the four Pd zeolite catalysts were
characterized in more detail. XRD patterns of all catalysts are presented in
Supplementary Figure 1. In all cases, the X-ray diffraction patterns corresponded to
0
what was expected for each zeolite framework [32]. No PdO or Pd crystallite peaks
were observed. The physicochemical properties of Pd/Zeolite catalysts and their
respective supports are described in Table 3 (nitrogen adsorption-desorption isotherms
are shown in Supplementary Figure 2). N physisorption and textural properties data
2
showed no significant changes by the impregnation of palladium on the zeolites. EDX
results showed that the metal content (wt.%) is similar to the nominal value (0.5 wt.%)
1
3

for all Pd/Zeolite catalysts. The distribution of acid sites of catalysts and their respective
supports were determined by NH
Supplementary Figure 3.
3
-TPD. The obtained profiles are shown in
Table 3. Physicochemical properties of all Pd/Zeolite catalysts and their respective
support, acidity obtained by NH
and Supplementary Table 1).
3
-TPD and XPS results (see Supplementary Figure 4
N
2
physisorption
3
NH -TPD
XPS
Catalyst
Si/Al Si/Al
S
S
V
V
Weak Strong
mmol.gcat-1
(m .g ) (m .g ) (cm .g ) (cm .g )
BET
-1
ext
p
3
micro
-1
PdXPS PdFRX Pdxps/PdFRX
2
2
-1
-1
3
FRX
HUSY
Pd/HUSY
HY
Pd/HY
HBEA
6
6
5
5
14
-
4.9
-
4.3
-
651
633
692
684
653
64
66
74
82
221
0.37
0.39
0.30
0.30
0.75
0.27
0.26
0.24
0.24
0.20
0.90
0.98
1.29
1.09
0.82
1.21
1.09
1.32
1.14
0.80
0.66
-
-
0.4
-
0.41
-
-
0.3
-
0.1
-
0.75
-
0.24
-
0.85
0.4
0.45
0.89
Pd/HBEA
14
16.2
645
235
0.87
0.17
0
.58
0.59
n.d.
0.38
n.d.
-
-
-
-
HZSM-5
23
14
-
-
394
484
118
237
0.21
0.50
0.12
0.10
m-ZSM-5
n.d.
-
0.52
-
-
0.58
-
0.70
0.3
-
Pd/HZSM-5
23
18.8
-
378
488
137
241
0.21
0.53
0.10
0.10
Pd/m-HZSM-5 14
n.d.
First, we observed that while the total acidity decreased in the order Pd/Y 
Pd/USY > Pd/HBEA > Pd/ZSM-5, and the n-hexane cracking followed the order: USY >
ZSM-5 ~ BEA > Y. Since n-hexane cracking demands strong acid strength sites, this
may explain Y zeolite smaller activity and ZSM-5 zeolite higher conversion when
compared to the BEA zeolite.
Next, even though DX was totally converted in all cases, in fact Pd/HUSY and
Pd/HBEA carried out more transformation of the intermediates, producing lower
amount of oxygenates and higher aqueous phase. Here, in addition to the high acidity,
we would note that BEA zeolite has the highest external area. This may facilitate the
transformation of the larger DX molecule. As shown in the table, the external area
decreased in the order BEA > ZSM-5 > Y  USY. The alkaline treatment to produce the
m-HZSM-5 increases the external area due to the creation of mesopores. Additionally,
the internal area was slightly lower in m-ZSM-5 compared to the start zeolite. This
1
4

slight reduction indicates that the alkaline treatment slightly damaged the zeolite
framework.
The tendency on coke formation observed USY > Y > BEA ZSM-5 can be
readily rationalized. The low yield of coke for ZSM-5 (0.6 wt.%) is most likely due to the
smaller pore size of the MFI framework, which limits bimolecular reaction [25, 26, [32,
33]. On BEA zeolite, the high external area enables a faster diffusion of the coke
precursors from the zeolite grain leading to a smaller coke formation [34]. By contrast,
the larger pore volume (Table 3) and higher acid site density of USY and Y zeolites
increased the coke yield. In particular, on USY, the extra-framework aluminum may
promote coke formation [35].
Pd determination by XPS and FRX is presented in Table 3 along with Pd ratio
obtained by XPS and FRX. This ratio follows the order Pd/BETA > Pd/USY > Pd/ZSM-
5
>> Pd/HY. This order represents the relative accessibility of the metal sites and could
be interpreted in terms of zeolite properties. For example, higher external area and
higher SAR limit ionic exchange, both properties favor metal on the external surface of
the catalyst. These are in line with the properties of BEA and USY. By contrast, Y
zeolite with high internal exchange sites, favor Pd in the interior of the zeolite grain. If
we speculate external Pd sites favor hydrogenolysis, then, Pd/HY catalyst would
provide the least hydrogenolysis of long-chain products as observed. Note that in this
case, we speculated that the metal effect is acting on the DX derivatives and not on the
initial step of the transformation of the DX molecule.
3.3 Pd/ZSM-5 testing in batch conditions: Effect of catalyst to DX ratio
The DX and n-hexane conversions, yields of coke, gas, and liquid fractions of
tests with varying DX/catalyst ratios are presented in Table 4 (detailed information can
be found in Supplementary Table 6). In the batch process, the weight balance closed
above 92%. Pure n-hexane was almost fully converted, but in the presence of DX, n-
1
5

hexane conversion remarkably decreased. For example, in the test T30/5, only 2.5% of
n-hexane was converted. Increasing the amount of catalyst (decreasing the
DX/Catalyst ratio), the n-hexane conversion increased. In all cases, DX was fully
converted. The higher reactivity of DX compared to n-hexane suggests that by
changing the reaction conditions, the conversion of the co-reactant can be minimized.
Table 4: Hydroconversion in batch condition ((n) g of DX + 90 g of n-hexane using (m)
g of Pd/HZSM-5, Tn/m). N-hexane and DX convesioin. Coke and liquid yield (wt.%).
Products selectivity the liquid phase (wt.% based on carbon intensity), A5- are
hydrocarbons with 5 carbons or less, branched A
6
are branched hydrocarbons
containing six carbons, CyA6+ are naphthenics containing cyclopentane and
cyclohexane with 6 or more carbons, A7+ are hydrocarbons with 7 carbons or more.
n-hex
T30/5
T20/5
T15/5
T15/10
T15/15
Conversion (%)
DX
n-hexane
-
98
100
2.5
100
2.3
100
3.0
100
4.4
100
10.2
Yield (wt.%)
Coke
Liquid
Organic liquid
phase)
0.1
3.0
1.71
0.78
92.0
0.83
89.3
0.71
90.2
(
29.9
90.8
91.6
-
0.8
0.6
0.2
-
-
Aqueous phase
Gas
67.2
97.2
1.9
96.5
2.4
96.2
2.0
94.9
2.0
92.1
2.5
93.4
Total
Selectivity (by carbon intensity in wt.%)
A
5-
88.9
7.9
<0.1
-
-
-
-
-
4.5
1.4
1.2
1.3
0.2
-
7.0
1.1
0.4
0.5
0.2
-
10.4
1.7
2.3
3.5
1.0
-
17.8
11.0
3.0
7.7
4.9
1.0
1.2
1.6
16.8
28.4
3.1
11.1
2.1
0.9
-
branched A
6
A
7+
CyA6+
Mono-Aromatics
Polyaromatics
Oxygenates
56.0
4.3
51.4
3.6
31.6
9.4
Non-identified
0.7
The coke yield is an important consideration in this hydroconversion process.
Naturally, pristine n-hexane produces a very small coke yield. When converting the DX
and n-hexane mixtures, a high DX/catalyst ratio gave rather high coke. For example,
the test T30/5 produced 3 wt.% of coke. However, keeping the amount of DX constant
at 15 grams, and increasing the catalyst did not affect the coke yield (T15/5-T15/15).
Indeed, a remarkable low coke yield could be obtained. For instance, 0.7 wt.% yield of
1
6

coke from a feed with 15 wt.% DX means approximately 5 wt.% of DX is converted into
coke.
The liquid fraction analysis in Table 4 shows that the mixtures of DX and n-
hexane produced mainly liquid organic phase. With a high DX/catalyst ratio, oxygenates
were the main products. Yet, increasing the amount of the catalyst remarkably
decreased oxygenates, and they were absent in the test T15/15. Probably, oxygenates
were intermediates, and more catalyst present transformed them into hydrocarbon.
The aqueous phase was observed in the experiments with a high DX/catalyst
ratio. Lower or even absence of the water phase was observed by increasing the
amount of catalyst per DX. We suggest that the water produced was trapped in the
zeolite catalyst. Worth mentioning is that the presence of water may have several
effects. It may affect the performance of the catalysts and inhibit the reaction of the
intermediate compounds. Further, the presence of acid sites and water may revert an
acetal of sugar into the corresponding ketone and sugar. Thus, we evaluated the
conversion of a mixture of xylose and acetone in the following section.
In short, even though DX is 100% converted in all tests, using a small catalyst
amount in relation to DX resulted in a high amount of oxygenates, probably being
intermediates to hydrocarbons. We attained total conversion to hydrocarbon by
increasing the catalyst to the DX ratio up to 1. It is important to mention that this was
attained at a lower temperature by the batch reactor (250 °C) than the flow reactor (350
°
C). Also high liquid yield of 90 wt.% was obtained. Further, we had a good weight
balance using the batch reactor.
Pristine n-hexane produced mainly hydrocarbons containing five and fewer
carbons. Isomers with six carbons represent 8 wt.%. In Table 4, comparing pristine n-
hexane to tests T30/5, T20/5, and T15/5, we observed that the branched isomers
containing six carbons drop 5 fold in intensity, considering n-hexane gave less than 30
wt.% liquid yield. In contrast, the other test gave ~ 90 wt.% liquid yield, in fact, the yield
of isomer dropped even more drastically. Yet, increasing the catalyst amount branched
1
7

isomers, A
6
, increased and the test T15/15 showed a higher A selectivity than pure n-
6
hexane. As n-hexane was poorly converted, we were led to conclude that DX is
converted to these compounds. Indeed, A
product of the conversion of DX.
6
could be considered as one of the major
Pristine n-hexane did not produce cycloalkanes and hydrocarbons containing
seven or more carbons are less than 0.1 wt%. In contrast, both cycloalkanes (CyA6+
and alkanes (A7+) were observed in the presence of DX and their amount increased
with catalyst amount. Particularly, on T15/10 and T15/15, CyA6+ was superior to A7+
)
.
This result was also observed in the fixed bed tests. The former showed most
hydrocarbons containing 8 carbons, such as ethyl-cyclohexane, but compounds up to
1
1 carbons could be observed (see Supplementary Table 5 and Supplementary Figure
). Regarding A7+, compounds up to 10 carbons were observed and alkanes with 7
5
carbons such as heptane and 2-methylhexane were the main products. In general, the
type of these products was similar to those observed in fixed-bed conditions. Aromatics
were not observed in pristine n-hexane and observed in a lower amount for both
T15/10 and T15/15 tests. They are mainly mono-aromatic compounds with 7 to 10
carbons, such as 1,3-dimethylbenzene and 1,2,4-trimethylbenzene. Also, in lower
amounts, polyaromatics like indenes and naphthalenes were observed. These
compounds were absent in a higher DX/ catalyst ratio or in the fixed bed tests.
From the above results, it is worth indicating that this possibility of producing
several classes of hydrocarbons (branched and linear alkanes, cycloalkanes,
aromatics) implies that green fuels like gasoline, jet-fuel, for example, could be
produced from DX.
The fraction of A5- in the liquid organic phase also contains dissolved light
hydrocarbons such as propane and butane. We combined these dissolved species with
the collected gas products and presented them in Table 5 (also in Supplementary
Table 4). Pristine n-hexane produces mainly propane, in general the light produts are
related to the hydrogenolysis of n-hexane [36, 37]. The amount of A5- products
1
8

remarkably increased compared to a fixed-bed reactor, as higher residence time was
used in bath reactor, thus favor sequentiall reactions of the isomers of n-hexane.
Decreasing the DX/catalyst ratio butanes and pentanes are main products, as n-
hexane conversion remarkably decreased, DX had an important contribution to these
products. Also, the conversion of DX into methane, ethane, and propane cannot be
ruled out.
Table 5: Hydroconversion in batch condition (90 g of n-hexane with (n) g of DX using
(
m) g of Pd/HZSM-5, Tn/m). Selectivity (wt.%) of CO, CO
2
and hydrocarbons with 5
carbons or less and CO /CO ratio.
2
n-hexane
T30/5
0.8
0.6
1.0
0.9
1.1
1.2
2.8
2.4
T20/5
1.0
0.8
1.7
2.5
0.9
1.3
4.5
3.5
T15/5
1.3
1.1
2.1
2.9
3.1
0.7
3.7
5.3
T15/10
0.1
0.9
3.1
6.7
7.0
0.6
2.9
T15/15
0.9
1.2
1.9
6.0
6.9
0.3
4.9
Methane
Ethane
Propanes
Butanes
Pentanes
CO
4.0
6.5
41.1
20.2
17.2
-
CO
2
-
CO /CO
2
-
4.7
16.8
Three important observations could be made from the analysis of CO and CO
in this light fraction. First, CO+CO represent an important fraction of the light gases.
Second, CO was always superior to CO, and the CO /CO ratio considerably increased
with increasing catalyst amount, up to 16.8 for T15/15. Hence, the catalyst promoted
more formation of CO and removed oxygen from the feed with less carbon loss. Third,
considering the last case (T15/15), CO + CO was 1.25 g, i.e. 8.3 wt% of DX. From an
2
2
2
2
2
2
oxygen balance, we deduced that water was the main product of deoxygenation. The
amount of water produced will be further discussed in the subsequent tests.
To enhance hydrocarbon production with special emphasis on the heavier
hydrocarbons (A7+ and CyA6+) fractions, we explored the effect of increasing
mesopores of the catalyst on the DX conversion. The ZSM-5 zeolite with mesopores
generated from the original ZSM-5 used by alkaline treatment was used as support.
The resulted catalyst was named Pd/m-HZSM-5, with characteristics already presented
1
9

in Table 3. The comparison of the test results with the original Pd/ZSM-5 catalyst was
shown in Table 6.
Table 6: Hydroconversion in batch condition (15 g of DX + 90 g of n-hexane using 15 g
of Pd/HZSM-5 or Pd/m-HZSM-5). DX and n-hexane conversion. Selectivity (wt.%),
products yield (wt.%), and amount of products (g). A5- are hydrocarbons with 5 carbons
or less, branched A are branched hydrocarbons containing six carbons, CyA6+ are
6
cycloalkanes containing cyclopentane and cyclohexane with 6 or more carbons, A7+ are
hydrocarbons with 7 carbons or more.
Pd/HZSM-5
100
Pd/m-HZSM-5
Pd/HZSM-5
Pd/m-HZSM-5
DX conv. (%)
n-hex. conv. (%)
100
3.3
-
-
-
-
10.2
Selectivity (wt.%)
Product (g)
Coke
Liquid
Gas
Water *
Total mass
Feed converted
2.9
n.a
10.3
17.8
-
3.5
n.a.
7.8
26.2
-
0.71
90.2
2.5
4.3
97.7
24.18
0.63
93.1
1.4
4.7
99.8
17.97
-
-
Selectivity (wt.%)
Product (g)
CO
CO
Methane
Ethane
Propanes
Butanes
0.3
4.9
0.9
1.2
1.9
6.0
6.9
28.4
3.1
11.1
2.1
0.9
0.7
0.1
3.7
0.4
2.2
2.3
12.6
6.6
11.8
5.2
15.6
4.8
0.7
0.07
1.18
0.22
0.29
0.45
1.45
1.66
6.88
0.76
2.69
0.51
0.22
0.18
0.02
0.66
0.07
0.39
0.41
2.26
1.18
2.11
0.93
2.79
0.87
0.13
0.41
2
Pentanes
branched A
6
A
7+
CyA6+
Aromatics
Polyaromatics
Non-identified**
2.3
CO
2
/CO
16.8
16.8
30.2
24.1
-
-
A
5-
4.07
4.31
*
water formation was estimated to close the oxygen balance.
** estimated by carbon intensity
It is observed that n-hexane conversion decreased in the presence of Pd/m-
HZSM-5 compared to Pd/HZSM-5. The gas yield decreased and the liquid yield
increased. The coke yield also showed a slight decrease. These observations were
consistent with the facilitated DX accessibility into the large porous volume, and
2
0

favored competition with n-hexane for active sites. Thus, lowering n-hexane
conversion. Mesopores may also favor the release of intermediate products, hence
decreasing gas and coke yields and increasing liquid products.
Pd/m-HZSM-5 favored more deoxygenation by forming water instead of CO
2
and CO. CO being the least amount. It is also important to point out that the amount of
oxygen estimated as water, from the oxygen balance between the feed DX, and the
oxygen containing products, such as water, CO
balance for both tests as presented in Table 6.
2
, and CO, nearly closed the material
Mesopore supported catalyst gave higher yields of A7+ and CyA6+. The
selectivity of alkanes and cycloalkanes products for both tests are presented per
carbon number for Pd/HZSM-5 and Pd/m-HZSM-5 in Figure 1. In the Pd/HZSM-5 a
maximum was observed for compounds with eight carbons, while for Pd/m-HZSM-5, a
maximum was observed for nine carbons. The mesopores presence also slightly
increased the mono-aromatics and slightly decreased the polyaromatics. These results
were again consistent with the increased pore size of the support, giving higher
molecular weight hydrocarbon products on one hand yet avoiding polycondensation of
intermediate aromatic products. Besides, as the silica to alumina ratio was lower for
this support, more site density could favor hydrogen transfer reaction, thus increasing
the aromatic compounds.
2
1

Figure 1: Alkanes and cycloalkanes selectivity (wt.%) obtained from the
hydroprocessing of 15g DX in 90 g n-hexane using 15 g of Pd/ZSM-5 and Pd/mZSM-5.
Making use of the complete analysis of Table 6, we can further estimate the
distribution of DX into products for both Pd/HZSM-5 and Pd/m-HZSM-5 catalysts to
emphasize the effect of increased mesoporosity of the support. This is presented in
Figure 2, which also shows Lewis structure of hydrocarbons per group. This was
derived considering the only contributions of n-hexane were to the hydrocarbons
containing six or less carbons (isomers A + A5-). These were discounted using the
6
value observed in the test of pristine n-hexane multiplied by the n-hexane conversion in
the mixture, assuming that n-hexane behaves equally when pure or in mixtures with
DX. Non-identified products were not taken into consideration. Water was estimated to
close the oxygen balance, i.e., discounting oxygen in CO and CO from the oxygen in
2
DX. Firstly, the amount of oxygen estimated as water to close the oxygen balance
almost closed the material balance for both tests as presented in Table 6.
From Figure 2, we could observe two important features that are
complementary. First, the presence of mesopores favored, even more, the main
product of deoxygenation, water, rather than CO
2
or CO. Hence, it is more efficient in
deoxygenation in terms of preserving carbon in the hydrocarbon product. Consistent
2
2

with this, we observed all the hydrocarbon products increased. Indeed, the amount of
green carbon of DX converted into hydrocarbons could be estimated: using Pd/HZSM-
5
, 70% of carbon from DX was converted into hydrocarbons, this value increased to
6% using Pd/m-HZSM-5.
8
The changes of the product profiles of A7+ have already been discussed and
were consistent with the presence of mesopores. However, we did not have enough
results to interpret the changes in the product profiles of branched A
present. We observe a shift from A to A5-. This could be a combination of factors.
From the heavier hydrocarbons formed, there could be more cracking by acid sites
lower SAR of m-HZSM-5) and more hydrogenolysis due to more accessible metallic
6
and A5- at
6
(
sites (due to mesopores). Further, there was a more direct contribution from the
transformation of DX or the intermediate oxygenates. Indeed, the conversion of
carbons from xylofuranose of DX into both A5- and A7+ compounds in hydrotreatment
conditions was demonstrated by using labeled molecules [12].
Figure 2: Products yield (wt.%) on a DX basis obtained from the hydroprocessing of 15 g of DX
in 90 g n-hexane using 15 g of Pd/HZSM-5 and Pd/HZSM-5 MESO. Error bars were determined
by three repetitions using only the Pd/HZSM-5 catalyst.
2
3

3
.4 A comparison with conversion of cellulose and sugars without
protection.
We can only give a general discussion because comparing this work with
relevant works is difficult due to the great variety in the experimental conditions used
and in the type of target products. The conversion of sugar and second-generation
biomass into furfural and HMF (5-hydroxylmethylfurfural) was widely investigated [38].
Furfural, HMF, and other derivatives are explored as a platform for chemicals and fuels
[
39-42]. Less amount of work was devoted to obtain green-hydrocarbon, particularly
with a longer carbon chain. Here, we will restrict our discussion on the type of
hydrocarbon products.
Simple, one-step processes used cellulose, hexoses, and pentoses as
platforms for producing hydrocarbons in hydro-de-oxygenation conditions (temperature
o
in the range 250-350 C and low to high H
2
pressure). In the presence of Pt/USY
catalyst, glucose and fructose were converted with high selectivity into hydrocarbons
containing four carbons and xylose into hydrocarbons with three carbons. The reaction
pathway was proposed to occur through furfural derivatives and acids [43]. In another
approach, Dumesic and co-workers developed aqueous-phase catalytic processes to
convert sugars [44]. The overall strategy consists of mainly three-step process: 1-
Hydrolysis to mono-saccharides; 2- hydroconversion of mono-shacharades into to a
hydrophobic organic liquid containing alcohols, ketones, carboxylic acids, and alkanes
containing four, five, or six carbon atoms, as well as heterocyclic tetrahydrofuran and
tetrahydropyran compounds; 3 – further, sequential processes (cracking, aldol
condensation ketonization, for example) convert the organic liquid into several types of
fuel, [45], which includes compounds in gasoline, or longer-chain useful as diesel [46].
Recently, in a biphasic reactor, green naphtha containing mainly hydrocarbons with five
and six carbons, including cyclo and branched alkanes [47], was obtained by a
combination of sequential processes (including hydrolysis, hydrogenation, acylation,
2
4

alkylation among other reactions). Then, compatible with further downstream petro-
refinery operations.
These deoxygenation processes involve dehydration, hydrogenation,
hydrogenolysis (C-O and C-C bonds) and their combinations. Thus, limiting the
hydrocarbon carbon chain equal to or less than the precursor sugar. Longer carbon-
chain and cyclic-alkanes are obtained using sequential reactions, mostly related to acyl
intermediates [48]. In general, multiple and sequential processes may reduce the yield
of a key product. This is particularly important for a carbohydrates’s deoxygenation
reaction (which contains lower carbons like hexoses and pentoses) into long carbon-
chain hydrocarbons (useful as regular fuels like gasoline, jet fuel, and diesel).
Finally, worth mentioning an alternative approach, alkanes containing 2-9
carbons were the main products, yet, C10-C20 compounds were also obtained. Various
biomass-materials, including different wood flours and cellulose were transformed in
the presence of Pt/HZSM-5 [49]. The biomass was pre-treated in alcohol at high
temperatures for hours. The pretreated material was suspended in water and subject to
hydrocracking at 673 K. The yields of C
2
-C hydrocarbons were very high, and much
9
higher than the reference, untreated case. According to the authors, the reaction
pathway includes lowering the molecular weight of cellulose materials, producing
oxygenates intermediates which are further converted to alkane by successive
hydrocracking and condensation.
At first glance, the pretreatment and the product profile prompted a better
comparison with the present ketalization and hydrotreament. However, the material
obtained after solvolysis by alcohol was not described in the report. Moreover, the
hydrotreatment was made in the presence of a much larger amount of water than our
present case where water is only one of the products. This will promote hydrolysis as
the initiation step different from our transformation. Further, the amount of coke
produced by this approach was not discussed.
2
5

3.5 Insights on pathway of DX conversion into hydrocarbons in hydro-
conversion condition
The liquid products of test T15/15 (in the presence of HZSM-5) were monitored
from 1-11 h on time on stream (TOS), as presented in Figure 3. Oxygenates decreased
during TOS. The intensity of these products was obtained by carbon intensity
(
measures in the GC-FID). Though not in weight %, they represent the contribution of
each type of product to the carbons in the mixture.
Figure 3: Product distribution in the liquid phase by carbon intensity (wt.% of carbon of a
product / total wt.% of carbon in the products) obtained from the hydroconversion of 15 g of DX
in 90 g of n-hexane using 15 g of Pd/HZSM-5. A5- (alkanes with 5 or less carbons), A
Branched hydrocarbons with 6 carbons, A7+ (Alkane with 7 carbons or more), C
Cycloalkane with 6 carbons or more).
6
y
A
6+
(
As observed from the previous sections of the catalyst to DX ratio study,
oxygenates were intermediates that decreased with the reaction's progress. The
identification and quantification of the oxygenated products are under investigation and
presented in our next publications.
As the oxygenates decreased, all hydrocarbons increased but in a different
manner. One of our main interests, cycloalkanes and alkanes (containing seven of
2
6

more carbons) increasing from 1 to 3 h of time of reaction, then remaining constant.
The aromatics behave similarly. Branched isomers containing six carbons and A5-
(
propane, butanes, and pentanes) increased with time of reaction.
It is well known that a sugar acetal can be reverted into sugar and acetone in
the presence of water and acid catalyst [50, 51]. We evaluated the importance of this
contribution to the oxygenate pool. A preliminar test converted a mixture of xylose,
acetone and n-hexane in the presence of Pd/ZSM-5, as presented in Figure 4. More
detailed data are presented in Supplementary Table 7. A radically different product
distribution was observed when compared to the DX conversion in the test T15/15. For
example, coke yield increased 5-fold, and oxygenates are observed. More strikingly,
cyclo-alkanes decreased 4-fold. These differences showed that the reverse reaction of
DX (for example, into xylose and two acetones) contributed mainly to light
hydrocarbons, aromatics, and coke, but did not lead to a broad range of hydrocarbons
as observed to DX. Thus, compared to unprotected sugar, the presence of an
isopropylidene group in DX decreased undesirable reactions like coke formaton and
improved the flexibility in obtaining different hydrocarbons types and long carbon-chain.
2
7

Figure 4: Products yield (g) produced by converting 15 g of DX + 90 g of n-hexane and 10 g of
xylose + 5 g of acetone + 90 g of n-hexane (It was used 15 g of Pd/ZSM-5 and 24 hours of
reaction). Note, the yield of oxygenates fraction was underestimated as it was derived from
carbon intensity without correction of the weight of oxygen in these components.
An attempt to sumarise the possible routes of transformation of DX into
hydrocarbons are presented in Scheme 1. DX can be converted into oxygenated
compounds (a pool of oxygenates) and sequentially into alkanes, cycloalkanes and
aromatics. Speculatively this first step obtaining oxygenate derivatives is mainly
promoted by acid sites. These products can undergo further reactions. Now metal sites
are operative in addition. First it plays a role in preventing coke deposition. Other
examples are: aromatics can be converted to cycloalkanes and vice-versa by
hydrogenation and dehydrogenation respectively, cycloalkanes and alkanes can be
converted to lighter hydrocarbons by hydrogenolysis, cracking and hydrogen transfer
reactions via acid catalysis also occur.
During the generation of oxygenates, water is also produced. Once this is
present, the reverse reaction of ketalization could take place, giving Xylose and
2
8

acetone. Their subsequent transformation to hydrocarbons could take place, but
probably this is not the main reaction pathway.
Yet, it is not clear if DX itself can be converted directly to some types of
hydrocarbons bypassing the oxygenates pool. Hence, a possible link is also drawn in
Scheme 1.
Scheme 1: General reaction pathway of DX conversion on Pd/HZSM-5 catalyst.
As demonstrated DX, itself or via an oxygenated pool, produced a broad range
of hydrocarbons containing up to 11 carbons. Potentially, the yield in target products
may be improved in further works. As attempts to suppress further transformation of
heavier hydrocarbons to lighter products, we have explored the modification of Pd to
decrease hydrogenolysis [52] and reduce acid sites by increasing the SAR of the
zeolites. These catalysts optimizations are subjects of our coming publications.
2
9