Effect of carbon chain length on catalytic C–O bond cleavage of polyols over Rh-ReOx/ZrO2 in aqueous phase

Article abstract of DOI:10.1016/j.apcata.2019.117213

Production of linear deoxygenated C4 (butanetriols, -diols, and butanols), C5 (pentanetetraols, -triols, -diols, and pentanols), and C6 products (hexanepentaols, -tetraols, -triols, -diols, and hexanols) is achievable by hydrogenolysis of erythritol, xylitol, and sorbitol over supported-bimetallic Rh-ReO_x (Re/Rh molar ratio 0.5) catalyst, respectively. After validation of the analytical methodology, the effect of some reaction parameters was studied. In addition to C–O bond cleavage by hydrogenolysis, these polyols can undergo parallel reactions such as epimerization, cyclic dehydration, and C–C bond cleavage. The time courses of each family of linear deoxygenated C4, C5, and C6 products confirmed that the sequence of appearance of the different categories of deoxygenated products followed a multiple sequential deoxygenation pathway. The highest selectivity to a mixture of linear deoxygenated C4, C5, and C6 products at 80percent conversion was favoured under high pressure in the presence of 3.7wt.percentRh-3.5wt.percentReO_x/ZrO₂ catalysts (54–71percent under 80 bar) at 200 °C.

Full text of DOI:10.1016/j.apcata.2019.117213

AppliedCatalysisA,General586(2019)117213
Contents lists available at ScienceDirect
Applied Catalysis A, General
journal homepage: www.elsevier.com/locate/apcata
Effect of carbon chain length on catalytic CeO bond cleavage of polyols over
Rh-ReOx/ZrO2 in aqueous phase
Achraf Sadiera, Noémie Perreta, Denilson Da Silva Perezb, Michèle Bessona, Catherine Pinela,⁎
a IRCELYON, Univ Lyon, Université Claude Bernard Lyon 1, CNRS, UMR 5256, 2 avenue Albert Einstein, F-69626, Villeurbanne, France
b Institut FCBA, InTechFibres, F-38044 Grenoble, France
A R T I C L E I N F O
A B S T R A C T
Keywords:
Aqueous phase
Polyols
Hydrogenolysis
Rh-ReOx/ZrO2 catalyst
Linear deoxygenated C4, C5, and C6 products
Production of linear deoxygenated C4 (butanetriols, -diols, and butanols), C5 (pentanetetraols, -triols, -diols, and
pentanols), and C6 products (hexanepentaols, -tetraols, -triols, -diols, and hexanols) is achievable by hydro-
genolysis of erythritol, xylitol, and sorbitol over supported-bimetallic Rh-ReOx (Re/Rh molar ratio 0.5) catalyst,
respectively. After validation of the analytical methodology, the effect of some reaction parameters was studied.
In addition to CeO bond cleavage by hydrogenolysis, these polyols can undergo parallel reactions such as
epimerization, cyclic dehydration, and CeC bond cleavage. The time courses of each family of linear deox-
ygenated C4, C5, and C6 products confirmed that the sequence of appearance of the different categories of
deoxygenated products followed a multiple sequential deoxygenation pathway. The highest selectivity to a
mixture of linear deoxygenated C4, C5, and C6 products at 80% conversion was favoured under high pressure in
the presence of 3.7wt.%Rh-3.5wt.%ReOx/ZrO2 catalysts (54–71% under 80 bar) at 200 °C.
1. Introduction
Under hydrogenolysis conditions, not only the desired CeO bond
cleavages occur together with the undesired CeC bond cleavage but
Hydrogenolysis reactions in aqueous medium of C4-C6 sugar alco-
hols including erythritol (ERY), xylitol (XYL), and sorbitol (SOR) are
important pathways for the production of high valuable chemicals. XYL
and SOR are considered among the twelve most relevant platform
molecules according to the US Department of Energy (DOE) [1], since
they are industrially obtained in large scale by catalytic hydrogenation
of the carbonyl groups of xylose and glucose, respectively [2]. They are
also important intermediates during the conversion of cellulose or xylan
[3–5]. Unlike the other polyols (XYL and SOR), the production of ERY
through direct catalytic hydrogenation is without interest due to the
high cost of erythrose substrate [6]. The standard industrial production
routes of this polyol involve the microbial fermentation of sugar and
sugar alcohols such as glucose and glycerol using various yeasts [7,8].
The selectivity of the catalytic hydrogenolysis reactions is a major issue
because both CeC and CeO bonds cleavage can take place during the
conversion of these highly functionalized molecules. The most chal-
lenging is the selective CeO bond cleavage without CeC bond cleavage
in order to produce partially deoxygenated chemicals used for in-
dustrial applications. They are key components in materials such as
polyesters, urethane foams and resins, and as intermediates for the
preparation of surfactants for which they can be used as isolated
compounds or as a mixture [9–12].
epimerization and cyclic dehydration reactions are also reported
[13–18].
Starting from glycerol (GLY), mainly, propanediols (1,2-PrDO, 1,3-
PrDO) and propanols (nPrOH, iPrOH) were obtained together with
small amount of ethanol (EtOH) and methanol [19]. Taking ERY as a
model molecule, butanetriols (1,2,3- and 1,2,4-BTO), butanediols (1,2-,
1,3-, 1,4-, and 2,3-BDO), and butanols (1- and 2-BuOH) were the main
deoxy products obtained from CeO bond hydrogenolysis [20]. In ad-
dition, ethylene glycol (EG), GLY, 1,2-PrDO and 1-PrOH were the sec-
ondary products obtained from competing CeC bond cleavage. More-
over, for polyols with four carbon atoms, intramolecular dehydration
reactions took place easily and formed thermodynamically stable het-
erocycles [21]. Indeed, dehydration of ERY yielded 1,4-anhydroery-
thritol (1,4-AE), whose successive CeO hydrogenolysis generated 3-
hydroxytetrahydrofuran (3OH-THF) and tetrahydrofuran (THF). Fi-
nally, ERY underwent epimerization reaction and formed threitol
(THR) [20].
Increasing further the number of carbon will increase the com-
plexity of the reaction resulting in the formation of more reaction
products. (Scheme 1)
Thus, C5 polyol (XYL) is converted to linear deoxygenated com-
pounds (pentanetetraols, -triols, -diols, and pentanols). Arabitol (ARB)
⁎ Corresponding author.
E-mail address: catherine.pinel@ircelyon.univ-lyon1.fr (C. Pinel).
https://doi.org/10.1016/j.apcata.2019.117213
Received 12 July 2019; Received in revised form 13 August 2019; Accepted 21 August 2019
Availableonline22August2019
0926-860X/©2019ElsevierB.V.Allrightsreserved.

A. Sadier, et al.
AppliedCatalysisA,General586(2019)117213
Scheme 1. Reaction pathways from polyols under hydrogenolysis conditions.
and adonitol (ADO) are formed after epimerization reaction, via de-
hydrogenation/hydrogenation reaction [13]. In addition, in-
tramolecular dehydration reactions of XYL yields 1,5- and 1,4-anhy-
droxylitol (1,5- and 1,4-AX); the successive CeO hydrogenolysis of 1,4-
AX generates tetrahydrofurfuryl alcohol (THFA). Moreover, CeC bond
hydrogenolysis reactions lead to many types of C1-C4 products such as
methanol, EG, EtOH, GLY, 1,2-PrDO, PrOH, ERY, and BuOH.
In turn, in addition to the desired deoxy products (hexanepentaols,
-tetraols, -triols, -diols, and hexanols), SOR can be converted to man-
nitol (MAN) and iditol (IDI) by epimerization. First dehydration yields
2,5-anhydrosorbitol (2,5-AS), 1,5-anhydrosorbitol (1,5-AS), and sorbi-
tans (1,4-S and 3,6-S), the successive second dehydration of sorbitans
generates isosorbide (IS). Also, the competing CeC bond hydro-
genolysis of SOR leads to many types of C1-C5 products including
methanol, EG, EtOH, PrOH, GLY, 1,2-PrDO, BDO, BuOH, pentanediols
(PDO), pentanols (PeOH), etc [14].
While, there is a large number of reports dealing with selective
hydrogenolysis of glycerol to propanediols [22–24], much less articles
focused on the hydrogenolysis of C4, C5 and C6 polyols and the lit-
erature dealing with the production of linear deoxygenated C4-C6
products (up to alkane) is rather limited [15–17]. The understanding of
the relative reactivity of each biobased polyols is mandatory for further
applications. So, CeO bond hydrogenolysis of XYL or SOR is an inter-
esting reaction to achieve and to compare with ERY hydrogenolysis.
Several groups reported that Cu-based catalysts are efficient for
CeO bond hydrogenolysis [9,17,25–27]. An hybridized Cu/CaO-Al2O3
catalyst (molar ratio of Cu/Ca2+ = 5.4) gave a selectivity of 39% to
1,2-BDO at 85% conversion in the hydrogenolysis of 2 wt% aqueous
solution of ERY at 230 °C under 28 bar of H2 [26]. Under optimum
conditions (180 °C, 130 bar of H2), in the presence of a commercial
CuO-ZnO (33:67) catalyst, a 21 wt% aqueous solution of SOR was hy-
drogenolyzed into a mixture of deoxyhexitols with a combined se-
lectivity of 63% at 92% conversion [17]. A solution containing 25 wt%
SOR, 25 wt% water, and 50 wt% 1,2-PrDO was hydrogenolyzed over a
Raney® Cu catalyst in a trickle-bed reactor (LHSV of 0.5 h−1 and hy-
drogen flow of 800 mL min−1, 210 °C, 124 bar) [9]. This patent claimed
78% selectivity towards deoxygenated C6 products at 99% conversion.
Very recently, the hydrogenolysis of 7.5 wt % aqueous solution of
mannitol (MAN) over Raney® Cu catalyst yielded a mixture of linear
deoxygenated C6 products with a molar composition of 76% (carbon
yield > 80%) at 180 °C under 200 bar of H2 [27].
Due to stability issues of Cu, alternative catalysts based on noble
metals are good candidates in the hydrogenolysis reactions. Ru-based
catalysts were efficiently reported for the hydrogenolysis of polyols
either in the presence of base, acid or neutral media. Significant C-C
cleavage were observed that could be reduced by addition of acid [18].
Furthermore, a number of previous studies showed that the co-
operation between an oxophilic promoter (Re, W, or Mo) and a highly
reducible noble metal (Rh, Ir, or Ru) enhanced the catalytic perfor-
mance in the selective CeO bond hydrogenolysis of polyols or cyclic
ethers (including tetrahydrofurfuryl alcohol and tetrahydropyran-2-
methanol) [15,16,20,28–35]. An Ir-ReOx/SiO2 (nRe/nIr = 1) catalyst
prepared by successive impregnation was investigated in the hydro-
genolysis of XYL and SOR in biphasic conditions (1 g substrate +4 mL
H2O + 4 mL n-dodecane). After 24 h at 140 °C under 80 bar of H2, al-
most full conversion of XYL and SOR was achieved; the maximum se-
lectivity towards linear deoxygenated C5 products from XYL reached
89%, and the selectivity to deoxy C6 products from SOR was 93% [16].
At 160 °C and under 80 bar of H2, a 2%ReOx-0.3%Pd/CeO2, was shown
to perform hydrodeoxygenation of vicinal diols, however the solvent
used was dioxane [15]. At nearly total conversion, the selectivity to
BDO was 93% in hydrogenolysis of ERY and only 1- and 3-PeOH were
detected with 99% selectivity from XYL, while SOR yielded 85% se-
lectivity to HDO. In our previous work on ERY hydrogenolysis in the
presence of a 3.7%Rh-3.5%ReOx/ZrO2 catalyst in aqueous phase, the
selectivity to BTO and BDO reached 37% and 29%, respectively, at 80%
conversion at 200 °C under 120 bar [20].
The objective of this paper was to compare hydrogenolysis of three
polyols (ERY, XYL, and SOR) in the presence of Re-promoted bimetallic
catalysts, in particular Rh-ReOx/ZrO2, and to study the effect of some
reaction parameters. The different number of carbon atoms in these
polyols will allow us to compare the relative ability of these substrates
2

A. Sadier, et al.
AppliedCatalysisA,General586(2019)117213
Scheme 2. Successive CeO bond hydrogenolysis of SOR under hydrogenolysis conditions. The framed molecules were commercially available.
to undergo parallel reactions such as epimerization, dehydration, CeC,
and CeO bond hydrogenolysis reactions.
products and detailed analysis of the products was performed (Figure
S1) [20]. Increasing the number of carbon (XYL and SOR) will increase
the possible formation of products especially dehydroxylated com-
pounds (Scheme S1, Scheme 2).
2. Results and discussion
Scheme S1 presents the successive CeO bond hydrogenolysis of XYL
to different categories of linear deoxygenated C5 products. Without
considering the configuration at each C, the first CeO cleavage of XYL
yields three pentanetetraols (PTeO). In addition, there are six possible
pentanetriols (PTrO) that can be formed through CeO bond cleavage of
PTeO. Subsequently, the CeO bond hydrogenolysis of the mixture of
pentanetriols generates six pentanediols (PDO). Finally, the hydro-
genolysis of PDO yields 1-, 2-, and 3-pentanols (PeOH) and then pen-
tane.
2.1. Identification of non-commercial deoxygenated products
Before being able to compare the catalytic hydrogenolysis results, it
was necessary to develop the analysis set-up for the numerous products
that can be obtained from the three reactants.
During ERY hydrogenolysis [20], four types of reactions were ob-
served which yielded four families of commercially available products.
HPLC analysis using Rezex ROA-Organic acid-H+ and Rezex RCM-
Monosaccharides-Ca2+ column allowed to separate all possible
Similarly, Scheme
2
presents the successive CeO bond
3

A. Sadier, et al.
AppliedCatalysisA,General586(2019)117213
hydrogenolysis of SOR with different categories of linear deoxygenated
C6 products. At the beginning, the first CeO bond cleavage of SOR
generates three hexanepentaols (HPO) without considering the stereo-
genic centers (HPO). Moreover, there are nine possible hexanetetraols
(HTeO) that can be formed through CeO bond cleavage of these hex-
anepentaols. Successively, the CeO bond hydrogenolysis of the mixture
of HTeO yields potentially ten hexanetriols (HTrO) and then, nine
hexanediols (HDO). Hydrogenolysis of HDO yields 1-, 2-, and 3-hexanol
(HeOH), and total hydrodeoxygation may finally produce hexane, that
is to say, that 34 partially oxygenated hexane derivatives can be formed
during the reaction.
Starting from XYL, a large number of linear deoxygenated C5 pro-
ducts is generated (Scheme S1). Among all the possible intermediates
(pentanetetraols, -triols, -diols, and pentanols), only a few of them were
commercially available (1,2-, 1,4-, 1,5-, and 2,4-PDO, 1-, 2-, and 3-
PeOH, Scheme S1). The further identification of the deoxygenated
products issued from XYL was then based on the following observa-
tions. First, previous work on hydrogenolysis of ERY clearly showed
that for a given number of carbon atoms, the more hydroxyl groups
present, the lower the retention time on the H+ column (Figure S1); in
contrast, no specific elution order was observed using the Ca2+ column.
Second, the catalytic hydrogenation reaction of three commercially
available deoxy-C5 sugars (2-deoxy-L-ribose, 3- deoxy-D-ribose, and 5-
deoxy-D-ribose) under mild conditions can yield the three pentanete-
traols (1,2,3,4-PTeO, 1,2,3,5-PTeO, and 1,2,4,5-PTeO, respectively)
shown in Scheme S1. Identification of the deoxygenated compounds
was essentially based on the hydrogenation of 2-deoxy-L-ribose.
Hydrogenation of 2-deoxy-L-ribose under mild conditions (0.17 mol
Fig. 1. Evolution of the peaks associated with the isomers of 1,2,3,5-pentane-
tetraol and pentanetriols during the hydrogenolysis of mixture of 1,2,3,5-PTeO
over Rh-ReOx/ZrO2 at 200 °C, with H+ column.
and (2S, 3S)-1,2,3,5-pentanetetraol (2′). On the other hand, CeO
cleavage of a secondary alcohol in XYL leads to the direct formation of
enantiomers (2) followed by isomer (1) after isomerization of one of the
chiral centers. Furthermore, the CeO cleavage of one of the primary
alcohol of xylitol produces both enantiomers (2R, 3R, 4S)-1,2,3,4-
pentanetetraol (3) and (2S, 3S, 4R)-1,2,3,4-pentanetetraol (3′). The
reaction medium of 2-deoxy-L-ribose hydrogenation after 24 h reaction
was analyzed on the H+ column. In the range 16.0–18.0 min, the major
peak at 16.4 min was assigned to (2S, 3R)-1,2,3,5-pentanetetraol (1)
and the minor peak at 17.0 min (shoulder) was attributed to the dia-
stereoisomers of 1,2,3,5-pentanetetraol (2) and (2′). In addition, in the
range 17.5–23.5 min, peaks with small areas started to appear from the
beginning of the hydrogenation reaction that could be attributed to
pentanetriols.
L
−1, 100 °C, 50 bar of H2, 5%Ru/C) allowed the identification of the
isomers of 1,2,3,5-pentanetetraols (Scheme 3). The direct hydrogena-
tion of 2-deoxyribose leads to the formation of (2S, 3R)-1,2,3,5-penta-
netetraol (1). The change in the conformation in position 2 and 3 yields
the second pair of diastereoisomers (2R, 3R)-1,2,3,5-pentanetetraol (2)
When harsh conditions were then used for hydrogenolysis of the
above mixture of PTeO isomers (200 °C over Rh-ReOx/ZrO2 catalyst),
the intensity of the main peak (1) at 16.4 min decreased to reach a
minimum after 9 h (green line) while the intensity of the peak at 17.0
increased to a maximum after 7 h. In parallel, six products were formed
that eluted between 17.5 and 23.5 min (Fig. 1). They were associated
with the family of pentanetriols, the monodeoxy compounds of PTeO.
1,2,3-, 1,2,5-, 1,3,5-, 1,3,4-PTrO are the possible triols formed from
1,2,3,5-PTeO and they correspond to the six different diastereoisomers
(Scheme S1).
For the analysis of the pentanediols (Figure S2), 2,4-PDO (two
diastereoisomers, 25.5 + 26.1 min), 1,4-PDO (29.8 min), 1,2-PDO
(34.2 min), and 1,5-PDO (36.1 min) are commercially available. So the
signals in the region 23.5–38.0 min could be assigned to the PDO fa-
mily. The pentanediols (1,2-, 1,3-, 1,4-, 1,5-, and 2,3-PDO) can be ob-
tained from 1,2,3,5-PTeO, while all these mentioned PDO as well as
2,4-PDO can be formed from XYL. The peaks observed in the zone
(23.5–38.0 min) using H+ column were thus attributed to the diaster-
eoisomers of the pentanediols (Scheme S1). THFA and 1,2-PDO were
eluted very close on the H+ column, and additional analysis was done
by GC.
On the other hand, the reaction of the most complicated model
(SOR) yields an even larger number of deoxygenated C6 compounds.
Among all the possible intermediates (hexanepentaols, -tetraols, -triols,
-diols, and hexanols), again, only a few of them were commercially
available (1,2,3-, and 1,2,6-HTrO, 1,2-, 1,5-, 1,6-, and 2,5-HDO, 1-, 2-,
and 3-HeOH, Scheme 2). Similar strategy for the identification of the
non-commercial linear deoxy C6 products was employed. The catalytic
Scheme 3. Ru/C catalyzed hydrogenation of 2-deoxy-L-ribose and CeO bond
hydrogenolysis of XYL to 1,2,3,5-PTeO. (2) and (2′) are enantiomers, (3) and
(3′) are enantiomers.
hydrogenation
reaction
of
three
commercially
available
4

A. Sadier, et al.
AppliedCatalysisA,General586(2019)117213
Scheme 4. Catalytic hydrogenation of sugars over Ru/C.
monodeoxysugars (L-rhamnose (6-deoxy-D-glucose), 2-deoxy-D-glu-
cose, and 3-deoxy-D-glucose) was performed under mild conditions
(0.02-0.55 mol L−1 sugar, 60 bar H2, 100 °C, 24 h, 5% Ru/C) to help in
the identification of HPO, the monodeoxy compounds of SOR. In ad-
dition to the hydrogenation reaction of the carbonyl function, iso-
merization was observed at the different hydroxyl positions yielding
eight diastereoisomers (with potentially 2 enantiomers for each) of
1,2,3,4,5-HPO starting from L-rhamnose, four diastereoisomers of
1,2,3,4,6-HPO starting from 2-deoxy-D-glucose, and four diastereoi-
somers of 1,2,3,5,6-HPO from 3-deoxy-D-glucose (Scheme 4).
As an example, the eight isomers of 1,2,3,4,5-HPO formed during
hydrogenation of L-rhamnose were clearly separated using the Ca2+
column (Fig. 2a).
The HPLC evolution of the concentration of the different products as
a function of time during hydrogenation of L-rhamnose allowed to
clearly identify the different diastereoisomers (Fig. 3).
Fig. 2. Chromatogram observed during catalytic hydrogenation of L-rhamnose
under mild condition (100 °C, 24 h), using (a) Rezex RCM-Monosaccharide
Ca2+ column, (b) Rezex ROA-organic acid H+ column. Signals 4–11 are at-
tributed to HPO, signals a–g are attributed to HTeO.
14.5 and 17.5 min which were attributed to 1,2,3,4,5-HPO isomers
(Fig. 2b). Furthermore, under more severe hydrogenolysis conditions,
the areas of these signals (14.5–17.5 min) decreased while new signals
appeared with retention times between 18.0 and 25.0 min, which were
attributed to HTeO, i.e. one CeO bond cleavage (Fig. 4).
Similar analysis was run starting from 2-deoxy-D-glucose and 3-
deoxy-D-glucose. It was shown that 1,2,3,4,6-HPO isomers eluted from
15.0 to 17.5 min while 1,2,3,5,6-HPO isomers eluted from 14.5 to
17.5 min. Unfortunately with this column, there is no possibility to
identify all the isomers formed during the transformation due to over-
lapping of the different signals.
Table S1 shows the deoxygenated products that can be obtained by
successive CeO bond cleavage of 1,2,3,4,5-, 1,2,3,4,6-, and 1,2,3,5,6-
HPO formed after the mild hydrogenation of L-rhamnose, 2-deoxy-D-
glucose, and 3-deoxy-D-glucose, respectively. As mentioned previously,
using H+ column for analysis of ERY hydrogenolysis, the more polar a
product is, the shorter the retention time is [20]. This was confirmed by
analysis of commercially available products (1,2,3-HTrO: 24.7 min,
1,2,6-HTrO: 27.7 min, 1,2-HDO: 49.4 min, 1,5-HDO: 41.3 min, 1,6-
HDO: 54.3 min and 2,5-HDO: 32.5 min).
Finally, as a result, Fig. 5 shows the range of retention time of each
family of linear deoxygenated C5 or C6 products using H+ column
during hydrogenolysis of XYL or SOR (Rh-ReOx/ZrO2, 200 °C, 50 bar
H2). The different linear deoxy products exhibit retention times that are
similar within a given family. The C5 products follow the order: XYL
and its isomers (15.0–15.5 min) < PTeO (16.0–17.5 min) < PTrO
(17.5–25.0 min) < PDO (25.0–38.0 min) (, Figs. 1, 5a and S2).
The compound 4 was the main product with a maximum con-
centration after 1 h and was directly obtained by hydrogenation of L-
rhamnose without epimerization. Then, 4 further isomerized to com-
pound 5 followed by 6 and 7. These three products correspond to iso-
merization at C2, C3 and C4 carbons, respectively. Isomers 8 to 11
corresponding to isomerization of a second carbon were formed, albeit
to a lesser extent. These results are in good agreement with previous
work of Palkovits et al [18]. In addition, minor peaks (a to g) appeared
after 24 h reaction time and can be attributed to further deox-
ycompounds (HTeO). To confirm this attribution, the reaction medium
resulting from the above hydrogenation reaction of L-rhamnose after
24 h (Fig. 2a) and containing mainly 1,2,3,4,5-HPO was further hy-
drogenolysed under more severe conditions (150 °C in the presence of
Ru/C catalyst). It was then observed that (i) the isomerization reaction
was more significant: the areas of signals 4 to 7 decreased while the
area of signals 8 to 11 increased; (ii) the area of signals attributed to
HTeO (a to g) increased and simultaneously the TOC value measured in
aqueous phase decreased (22% loss after 7 h). Using the Ca2+ column,
clear identification of the 1,2,3,4,5-HPO isomers was established. Un-
fortunately, during hydrogenolysis of sorbitol, there was also major
overlapping (i) between all the peaks related to C6 deoxygenated
compounds, i.e. HPO, HTeO, HTrO, and HDO and also (ii) between
these C6 deoxygenated compounds and the C-C hydrogenolysis pro-
ducts (C2-C5). This precludes any identification and quantification with
this column which was not used furthermore in the study.
In parallel, analysis was carried out using the H+ column. After mild
hydrogenation of L-rhamnose, broad signals were observed between
5

A. Sadier, et al.
AppliedCatalysisA,General586(2019)117213
Fig. 3. Evolution of the concentration of L-rhamnose and the different isomers of 1,2,3,4,5-hexanepentaols as a function of time. (♦) L-rhamnose; ( ) 4; (
6, and 7; (
) 8, 9, 10, and 11 (Ca2+ column analysis).
,
,
) 5,
,
,
,
Fig. 5. Response observed during the hydrogenolysis of (a) XYL or (b) SOR over
Rh-ReOx/ZrO2 at 200 °C under 50 bar of H2, with H+ column.
2.2. ZrO2- supported catalyst: influence of the metal
Fig. 4. Evolution of the peaks during hydrogenolysis of L-rhamnose over Ru/C
at 100 °C and 150 °C, with H+ column, in the range (a) 14.5–17.5 min and
b).17.5–25.0 min.
As mentioned in the introduction, several works have shown that
bimetallic catalysts based on a noble metal together with an oxophilic
promoter such as rhenium were efficient for hydrogenolysis reaction.
The effect of noble metal on sorbitol conversion and product selectivity
is presented in Table 1. Over bare supports (ZrO2) the catalytic activity
was negligible (5% SOR conversion after 24 h) as it was observed for
ERY. [20] Only cyclic C6 compounds (1,5-AS, 2,5-AS, 1,4-S, 3,6-S, and
isosorbide) were formed due to the intrinsic moderate acidity of the
support. Neither CeO nor CeC bond cleavage products were formed
due to the absence of metallic sites active in H2 dissociation.
Similarly, C6 products follow the same order: SOR and its isomers
(13.5–14.0 min) < HPO (14.5–17.5 min) < HTeO (17.5–24.5 min) <
HTrO (24.5–30.0 min) < HDO (32.0–56.0 min) (Figs. 5b and S3 to S6).
These retention time zones will be used to quantify the different fa-
milies of products.
ReOx/ZrO2 exhibited only 23% conversion after 24 h; the products
analyzed were mainly cyclic C6 compounds with a 54% selectivity.
6

A. Sadier, et al.
AppliedCatalysisA,General586(2019)117213
Table 1
Conversion and carbon selectivity to the different families during hydrogenolysis of SOR. Reaction conditions: SOR 0.27 mol L−1, 120 mL H2O, 200 °C, 80 bar of H2,
500 mg catalyst, nSOR/nRh = 200, nSOR/nPt = 342, nSOR/nIr = 377.
Catalyst
Molar ratio Re/Rh
Time [h]
Conv. [%]
Carbon selectivity (%)
C1-C5
Isomers
Cyclic C6
C6
ReOx/ZrO2
–
0.5
1
24
3
6
23
80
80
80
8
9
9
6
10
13
11
5
54
20
28
55
26
54
47
32
Rh-ReOx/ZrO2
Pt-ReOx/ZrO2
Ir-ReOx/ZrO2
1
24
Isomers = MAN + IDI, Cyclic C6 = 1,5-AS + 2,5-AS + 1,4-S + 3,6-S + IS, C6 = HPO + HTeO + HTrO + HDO + HeOH.
Deoxygenated C6 products (26% selectivity) and isomers (10% se-
lectivity) were also generated in addition to CeC bond cleavage pro-
ducts (8% selectivity). This result is in good agreement with the report
by Shao et al [36] who investigated the hydrogenation of 7.65 wt%
aqueous solution of 1,4-butanediol at 240 °C under 80 bar of H2 in the
presence of a 4 wt.%Re/C catalyst. THF was the main product formed
through internal dehydration with selectivity of 83.1% at 65% con-
version. Under the same reaction conditions, conversion was negligible
(< 1%) for the hydrogenation of a 6.1 wt.% solution of THF. In this
cited work, Re-based catalysts were mainly efficient for cyclization, but
they also catalyzed the dehydrogenation/hydrogenation steps (isomer
products) and hydrogenolysis reaction (linear deoxygenated C6 pro-
ducts) to a much less extent. It was also reported that ReOx species
exhibit surface acidity, as proved by IR spectra of pyridine adsorption
and NH3-TPD [37,38]. This suggests that ReOx species can trigger the
internal dehydration of SOR.
At 200 °C under 80 bar of H2, the catalytic activity for sorbitol
conversion followed the order Rh ≈ Pt > Ir, and the reaction time
needed for attaining ca 80% conversion was 3, 6, and 24 h, respectively.
Whatever the noble metal, the selectivity to C1-C5 products (in the
range 6–9%) remained roughly constant. Isomerization reaction was
more important in the presence of Rh-ReOx/ZrO2 and Pt-ReOx/ZrO2
catalysts (selectivity of 11–13%). In addition, Ir-ReOx/ZrO2 triggered
the internal dehydration of sorbitol to cyclic C6 products with a se-
lectivity of 55% at 80% conversion. Finally, the selectivity to linear
deoxygenated C6 products was favored in the presence of Rh-ReOx/
ZrO2 with a selectivity of 54%. The catalytic selectivity to C6 products
followed the order: Rh ≥ Pt > Ir. For that reason, Rh-based catalysts
were chosen in the selective hydrogenolysis of polyols.
2.3. Hydrogenolysis of XYL and SOR over Rh-ReOx/ZrO2 catalyst
Fig. 6. Evolution of the concentrations of the different families of products as a
function of time during hydrogenolysis of a) XYL and b) SOR. (♦) XYL or SOR;
Fig. 6 shows the evolution of the concentrations of the different
families of products defined in Scheme 1 as a function of time during
hydrogenolysis of XYL (a) and SOR (b) over Rh-ReOx/ZrO2 under si-
milar reaction conditions.
( ) Cyclic C5 or C6 products; ( ) Isomers; ( ) Deoxy C5 or C6 products; (
)
C5, ( ) C4, ( ) C3, and ( ) C2 products. Reaction conditions: reactant 0.27-
0.30 mol L−1, 120 mL H2O, 200 °C, 80 bar of H2, 500 mg Rh-ReOx/ZrO2,
nreactant/nRh =.200–230.
The hydrogenolysis reactions were carried out at 200 °C and 80 bar
of H2. Under these conditions, XYL and SOR were rapidly converted to
attain full conversion after 6–6.5 h (Fig. 6a and b).Similar initial reac-
(< 0.040 mol L−1, ˜ 5%).
The same trend of evolution of the families of compounds was ob-
served during SOR reaction (Fig. 6b). Linear deoxygenated C6 products
appeared from the initial stage of the reaction as main products and
their concentration attained a maximum of 0.155 mol L−1 after 5 h (˜
ca. 57% yield). Simultaneously, cyclic C6 compounds (mainly 1,4-S and
IS) were detected and reached a plateau after 3 h (˜ 0.040 mol L−1 or
15% of initial SOR). Isomerization of SOR to MAN and IDI occurred,
after 3 h, the maximum concentration reached 10% of initial SOR
concentration. In addition, products issued from C-C cleavage, in-
cluding C5 (THFA, PDO, PeOH), C4 (1,2-BDO, BuOH), C3 (GLY, 1,2-
PrDO, 1-PrOH), C2 (EG, EtOH), and C1 (methanol, not shown) were
formed continuously in low concentrations as secondary products.
To summarize, similar products distribution was observed during
hydrogenolysis of XYL and SOR, namely the products issued from CeO
tion rates were calculated for both substrates (ca 11 mol L−1 h−1
)
During XYL hydrogenolysis, linear deoxygenated C5 products appeared
as the main products from the initial stage of the reaction (Fig. 6a).
Their concentrations increased to a maximum of 0.150 mol L−1 after
5 h (yield ˜ 50%). In addition, the concentration of cyclic C5 compounds
increased and attained a maximum of 0.066 mol L−1 after 4 h (yield ˜
22%). These cyclic compounds were relatively resistant to hydro-
genolysis and they were still present after 6 h (0.050 mol L−1). Lower
reactivity of cyclic compounds was previously reported by Palkovits
et al via H/D exchange reaction [18]. Moreover, the concentration of
XYL isomer (mainly ARB) increased to a maximum of 0.027 mol L−1
after 3 h, and then it was also converted. Finally, C2 (EG, EtOH), C3
(GLY, 1,2-PrDO, 1-PrOH), C4 (ERY and 1,2-BDO), and C1 (methanol,
not shown) products were formed continuously in low concentrations
7

A. Sadier, et al.
AppliedCatalysisA,General586(2019)117213
Fig. 7. Comparison between (•,▲) TOC measured and (○, △) TOC calculated
during XYL (△,▲) and SOR (○,•) hydrogenolysis over Rh-ReOx/ZrO2
(Reaction conditions: see Fig. 6).
hydrogenolysis were formed predominantly up to 50-% yield. In addi-
tion, intramolecular dehydration occurred to give cyclic ether up to
15–22% yield. These cyclic ethers are much more resistant to CeO
hydrogenolysis and once they are formed, they do not further react.
Fig. 7 compares the evolution of TOC values (TOC measured and
TOC calculated) during XYL and SOR hydrogenolysis as a function of
time under these conditions.
At time zero, TOC measured was 18.5 and 19.5 gC L−1, respectively,
for the XYL and SOR solutions, respectively. The TOC values decreased
during the reaction (15–18% loss after 6–6.5 h) for both substrates due
to the formation of small amount of gaseous products via C-C hydro-
genolysis as mentioned previously. However, it is worth noting that, as
time proceeded, no significant difference was observed between the
measured TOC values and the calculated TOC values, which indicates
that the products present in aqueous phase were correctly analysed by
HPLC and GC. The mass balance was always over ca. 90%.
Then, for each substrate (XYL or SOR), we focused on the dis-
tribution of the different deoxy C5 or C6 products (main compounds),
respectively, as a function of reaction time, as reported in Fig. 8.
Under 80 bar of H2, full conversion of XYL was attained after 6 h and
PTeO (first CeO bond cleavage) is the major product among the cate-
gories of linear deoxygenated C5 compounds (Fig. 8a). It appeared from
the beginning of the reaction and the maximum concentration was
0.065 mol L−1 after 3 h (˜ 23% yield). On the other hand, PTrO ap-
peared after 0.5 h of the reaction to reach a plateau at 0.043 mol L−1
after 4 h (˜ 15% yield), whereas PDO was observed after 0.5 h and its
concentration attained 0.056 mol L−1 (˜ 19% yield). Lastly, the con-
centration of PeOH was negligible.
Fig. 8. Evolution of the concentrations of each family of deoxygenated C5 and
C6 products as a function of time during hydrogenolysis of a) XYL and b) SOR.
(♦) XYL or SOR; ( ) PTeO or HPO; ( ) PTrO or HTeO; ( ) PDO or HTrO; (
)
PeOH or HDO; ( ) HeOH. Reaction conditions: reactant 0.27-0.30 mol L−1, 120
mL H2O, 200 °C, 80 bar of H2, 500 mg Rh-ReOx/ZrO2, nreactant/nRh =.200–230.
expected during XYl and SOR hydrogenolysis.
Then, the influence of reaction conditions on activity and selectivity
for the hydrogenolysis of C4 to C6 polyols was studied.
2.4. Effect of substrate concentration over Rh-ReOx/ZrO2
The effect of substrate concentration (in the range 5–13 wt%) on
initial rate (Fig. 9) and on product selectivity at 70–86 % conversion
(Fig. 10) was investigated for ERY, XYL, and SOR hydrogenolysis at
200 °C, under 80 bar of H2 over 300–500 mg Rh-ReOx/ZrO2 (nRe/nRh
=
Similarly, HPO (one CeO bond cleavage) was the major family of
products formed from the beginning of reaction of SOR (Fig. 8b). The
concentration went through a maximum at 0.066 mol L−1 after 3 h,
then it decreased to 0.029 mol L-1 at 6.5 h. On the other hand, HTeO
started to appear after 0.5 h of the reaction, the concentration increased
up to a plateau at 0.030 mol L−1 after 5 h. HTrO and HDO were de-
tected only after 1 h and reached a concentration around 0.020 mol L−1
after 6.5 h. Lastly, HeOH was formed in traces. The formation of
polydeoxygenated C5 compounds must be generated from the succes-
sive CeO bond hydrogenolysis of XYL through PTeO as intermediates.
Similarly, polydeoxygenated C6 products from SOR should occur via
successive hydrogenolysis via HPO.
As mentioned previously, the analytical method using H + column
did not allow separating the different isomers (PTeO’s and HPO’s from
XYL and SOR, respectively). It is noteworthy during ERY hydro-
genolysis [20] that primary and secondary CeO hydrogenolysis oc-
curred simultaneously yielding to a mixture of 1,2,3-BTO and 1,2,4-
BTO. Similar absence of primary/secondary CeO selectivity was
Fig. 9. Effect of substrate concentration on initial rate of the reaction. ( ) ERY,
) XYL, ( ) SOR. Reaction conditions 120 mL H2O, 200 °C, 80 bar of H2,
300–500 mg Rh-ReOx/ZrO2.
(
8

A. Sadier, et al.
AppliedCatalysisA,General586(2019)117213
Fig. 11. Effect of temperature on selectivities at 80–95% conversion. ( ) %
TOC measured ( ) C-C cleavage products; ( ) Cyclic compounds; ( ) Isomers;
Fig. 10. Effect of substrate concentration on carbon selectivities to reaction
products during hydrogenolysis of ERY, XYL, and SOR at 70–86% conversion.
(
L
) Linear deoxygenated products. Reaction conditions: Reactant 0.27-0.40 mol
−1, 120 mL H2O, 80 bar of H2, 300–500 mg Rh-ReOx/ZrO2 catalyst; nreactant
/
(
) C-C cleavage products; ( ) Cyclic compounds; ( ) Isomers; ( ) Linear
nRh =. 200–450.
deoxygenated products. Reaction conditions: 120 mL H2O, 200 °C, 80 bar of H2,
300–500 mg Rh-ReOx/ZrO2.
effect on the selectivity to the reaction products at 70–86% conversion
is presented in Fig. 11 (Table S2). It is also worth mentioning that the
carbon loss (obtained from TOC measurement) at the defined conver-
sion of each substrate (ERY, XYL, and SOR) increased from 3, 8, and 4%
to 20, 16, and 10%, respectively as temperature increased. The se-
lectivity to C-C cleavage products was in the same range (5–12%), ex-
cept for ERY at 240 °C. The selectivity reached 16%, but it must be
noted that the mass balance in that case was lower due to significant
formation of products in gas phase.
0.5). In these experiments, the concentration of each substrate was
adjusted by varying the amount of the polyol, keeping constant the
volume of water (120 mL).
The reaction order with respect to all substrates (ERY, XYL, and
SOR) concentration was almost zero (Fig. 9). Indeed, this behavior was
observed in the literature for the hydrogenolysis of 1,4-anhydroery-
thritol [39] and tetrahydrofurfurylalcohol [40], in which the reaction
order with respect to 1,4-AE and THFA concentrations over Rh/SiO2
decreased from 0.6 and 0.4, respectively, to zero order over Rh-MoOx/
SiO2 and Rh-ReOx/SiO2. The authors suggested that MoOx and ReOx
species are responsible for the strong adsorption of 1,4-AE and THFA
over Rh-MoOx/SiO2 and Rh-ReOx/SiO2 catalysts, respectively. Similar
phenomenon could be suggested in the present work with adsorption of
substrates on ReOx species of Rh-ReOx/ZrO2 catalyst.
At 200 °C and under 80 bar of H2, whatever the substrate con-
centration (5–13 wt%), the selectivity to CeC cleavage products (C1-
C3, 11–12%; C1-C4, 6–8%; C1-C5, 5–9%) and to isomers
(THR = 7–13%; ARB = 9%; MAN + IDI = 9–13%) remained roughly
constant at similar conversion. Surprisingly, the selectivity to the cor-
responding C4, C5, and C6 cyclic compounds resulting from internal
dehydration of substrates increased moderately from 12, 26, and 21%
to 25, 41, and 38%, respectively, as substrate concentration increased
from 5 to 13 wt%. Indeed, this is expected since the cyclic dehydration
reaction is not catalyzed by metallic catalysts and it will be favored at
high concentration at the detriment of the CeO bond hydrogenolysis.
On the other hand, it is worth noting that the highest selectivity to
On the other hand, the highest selectivity to cyclic C4 from ERY, C5
from XYL and C6 from SOR was favored at 240 °C (selectivity = 25, 38
and 42%, respectively). These data are consistent with the observations
during hydrogenolysis of 9.4 wt% of SOR in ethylene glycol monoethyl
ether under 68 bar of hydrogen over a copper chromite catalyst [41]:
the yield to 2,5-bis(hydroxymethyl)-tetrahydrofuran obtained from in-
ternal dehydration of 1,2,5,6-hexanetetraol was 16.5% after 23 h at
200 °C; upon increasing the temperature from 200 to 260 °C, the yield
to the cyclic compound increased to 31.4%, which confirms that the
dehydration reactions are favored at high temperature; in addition, the
heterocyclic products are thermodynamically stable [41]. Moreover,
upon decreasing the reaction temperature from 240 to 160-150 °C, the
selectivity to linear deoxygenated C4, C5, and C6 products increased
from 33 to 34% to 80, 62, and 59%, respectively. The same trend was
observed in the dehydroxylation of MAN over Cu-Raney® catalyst [27]
in which the molar carbon selectivity to linear deoxygenated products
increased from 35 to 75% as temperature decreased from 220 to 180 °C.
These results confirm that upon decreasing the reaction temperature,
the selectivity towards CeO bond hydrogenolysis will be enhanced. It is
worth noting that the temperature has less influence on the selectivity
towards CeO bond cleavage for longer carbon chain. In parallel, iso-
mers formation was favored at low temperature; indeed the selectivity
to THR (isomer of ERY), ARB (isomer of XYL), and MAN + IDI (isomers
of SOR) increased from 6, 4, and 5% to 10, 13 and 16%, respectively, as
the temperature decreased. This trend in selectivity to isomers is in
accordance with the results obtained during hydrogenolysis of SOR to
ethylene glycol and propylene glycol over Ru/C catalyst, in which the
selectivity to C5 and C6 polyols (XYL + ARB + MAN + IDI) decreased
from 10 to 7% with increasing temperature from 180 to 220 °C [42].
linear
deoxygenated
C4
(BTO + BDO + BuOH),
C5
(PTeO + PTrO + PDO + PeOH), and C6 (HPO + HTeO + HTrO +
HDO + HeOH) products (58, 59, and 54%, respectively), was favored
for the lowest substrate concentration (5 wt%). This trend is in line with
the literature dealing with the selective hydrogenolysis of SOR to
1,2,5,6-hexanetetraols (200 °C, 136 bar of H2) over copper chromite
catalyst [9,41]. In these studies, the authors reported an increase of the
yield of 1,2,5,6-hexanetetraols from 28 to 38% as the concentration of
SOR decreased from 9.4 to 2 wt%, in ethylene glycol monoethyl ether as
solvent. Based on these results, substrate concentration of 5 wt% was
chosen for hydrogenolysis experiments, due to highest selectively to-
wards linear deoxygenated C4, C5, and C6 products.
2.6. Effect of H2 pressure over Rh-ReOx/ZrO2
2.5. Effect of reaction temperature over Rh-ReOx/ZrO2
The effect of H2 pressure on initial rate and product selectivity was
investigated for hydrogenolysis of the three substrates (ERY, XYL, and
SOR) at 200 °C, over Rh-ReOx/ZrO2 catalyst. During ERY, XYL, and SOR
The temperature was varied in the range of 150 °C to 240 °C, under
80 bar of H2, in the presence of 300–500 mg of Rh-ReOx/ZrO2. The
9

A. Sadier, et al.
AppliedCatalysisA,General586(2019)117213
Fig. 12. Effect of H2 pressure on selectivity to the reaction products and %TOC
measured at 80–95% conversion. ( ) % TOC measured; ( ) C-C cleavage
products; ( ) Cyclic compounds; ( ) Isomers; ( ) Linear deoxygenated pro-
ducts. Reaction conditions: reactant 0.27-0.40 mol L−1, 120 mL H2O, 200 °C,
300–500 mg Rh-ReOx/ZrO2 catalyst; nreactant/nRh =. 200–450.
reactions, the initial rate increased linearly from 300, 341, and 321
mmolSub gRh−1 h−1 to 800, 618, and 603 mmolSub gRh−1 h−1, respec-
tively, as pressure was varied from 30 to 80 bar.
The effect of hydrogen pressure (in the range 30–120 bar) on the
selectivity to the different products is presented in Fig. 12 (Table S2).
First, concerning the carbon balance, there was no difference between
TOC calculated and TOC measured in the hydrogenolysis of all sub-
strates. Also, the mass balance was > 90% under the range of H2
pressure applied, which means that the formation of gaseous com-
pounds such as CO2 or alkanes (methane, ethane) is limited, except for
ERY at 30 bar. Under these reaction conditions, minor C-C cleavage
occurred and the carbon selectivity to the products issued from C-C
cleavage was < 12% regardless the substrate and the H2 pressure
(30–120 bar). Secondly, the selectivity to cyclic C4, C5 and C6 products
was favored at low pressure (30 bar) (21, 48, and 35%, respectively).
These data are consistent with observations during the hydrogenolysis
of 4.8 wt% aqueous solution of SOR to hexane over Pt/NbOPO4 catalyst
at 250 °C. The yield to cyclic oxygenated products [1,4-sor-
Fig. 13. Dependence of product selectivity on ERY conversion under a) 80 bar
b) 120 bar. ( ) C3 products; ( ) Cyclic C4 products; ( ) Threitol; ( ) BuOH; (
) BDO; ( ) BTO. Reaction conditions: ERY 0.40 mol L−1, 120 mL H2O, 200 °C,
300 mg Rh-ReOx/ZrO2.
MAN + IDI; selectivity was of 9% at 120 bar starting from ERY,
whereas under 80 bar of H2, from XYL and SOR, the selectivity to iso-
mers reached 9 and 13%, respectively. This behavior is also in ac-
cordance with results obtained for the selective epimerization reaction
of 33.3 wt% isosorbide solution to isoidide and isomannide at 220 °C
over Ru/C; the selectivity to the isomers (isoidide + isomannide) in-
creased from 5 to 57% as hydrogen pressure increased from 23 to
58 bar [44].
Fig. 13 is an example of the evolution of carbon selectivity to the
different products as a function of ERY conversion under 80 and
120 bar of H2, at 200 °C, in the presence of Rh-ReOx/ZrO2 catalyst.
At low conversion < 40%, the total selectivity was in the range
105–115% due to the uncertainty of the analysis at low concentrations.
Under 80 bar, the selectivity to BTO decreased regularly from 47 to
22% as conversion increased from 21 to 92%, while at the same time
the selectivity to BDO remained constant around 26%. Differently,
under 120 bar, the selectivity to BTO decreased slowly at first (for
conversion < 80%), then sharply from 43 to 18%, as conversion in-
creased up to 100%. At the same time, the selectivity to BDO increased
from 25 to 35%. Moreover, the selectivity to BuOH increased from 0 to
4% under 80 bar, and from 4% to 8% under 120 bar. In addition, the
selectivity to THR decreased from 22 to 5% and from 14% to 3% under
80 and 120 bar, respectively, with increasing conversion. The se-
lectivity to C3 did not vary with conversion (7–12% under 80 and
120 bar). Finally there was no effect of pressure on the selectivity to
cycles which remained constant (9–12% under 80 bar, slightly lower at
ca. 6–7% under 120 bar) as a function of conversion.
bitan + isosorbide
+ 2-(tetrahydrofuran-2-yl)ethan-1-ol] decreased
from 60 to 10% upon increasing the hydrogen pressure from 10 to
50 bar [43]. Indeed, these two reactions (intramolecular dehydration
and hydrogenolysis) are parallel reactions and the latter being favoured
by increasing hydrogen pressure. In parallel to the observed decrease of
cyclic compounds, the selectivity to linear deoxygenated products of
ERY, XYL, and SOR increased from 43, 32, and 43% to 71, 55, and 54%,
respectively, when the pressure increased. More specifically, the se-
lectivity to mono C-O hydrogenolysis product was higher when the
pressure increased, regardless the nature of the polyol (Table 2). In
parallel, the selectivity to the other consecutive CeO cleavage products
remained roughly constant except for XYL, where it increased from 16
to 31% as the pressure increased from 30 to 80 bar. Even if iso-
merization reaction remained moderate, high pressure favors slightly
the epimerization reaction of ERY to THR, XYL to ARB, and SOR to
Table 2
Effect of H2 on the carbon selectivity to the mono- and poly-deoxygenated
products (reaction conditions see Fig. 12).
PH2 (bar)
ERY (%)
Mono
XYL (%)
Mono
SOR (%)
Mono
Poly
Poly
Poly
30
80
120
13
28
37
30
30
34
16
20
24
16
21
31
20
22
31
23
23
23
10

A. Sadier, et al.
AppliedCatalysisA,General586(2019)117213
corresponding XYL or SOR under identical reaction conditions (Rh-
ReOx/ZrO2 at 200 °C, under 50 bar of H2). 1,2,3,5-PTeO and 1,2,3,4,6-
HPO were obtained from the mild hydrogenation reaction over Ru/C at
100 °C under 50 or 60 bar of H2, of 2-deoxy-L-ribose or 2-deoxy-D-
glucose, respectively (see experimental section).
2.7. Insight into the deoxygenation pathway
Let us first consider the case of XYL and PTeO. Under these condi-
tions, 95% XYL conversion was attained after 9 h with up to 40%
carbon selectivity to linear deoxygenated products (Fig. 15a) and PTeO
(first CeO bond cleavage) is the major product among the linear
deoxygenated C5 compound (Fig. 15b). That compound appeared from
the beginning of the reaction and the maximum concentration was
0.053 mol L−1 after 5 h (˜ 18% yield). On the other hand, PTrO ap-
peared after 0.5 h of the reaction to reach a plateau at 0.028 mol L−1
after 7 h (˜ 10% yield), whereas PDO was observed after 1 h and its
concentration attained 0.032 mol L−1 (˜ 11% yield). Lastly, the con-
centration of PeOH was 0.003 mol L−1 (˜ 1%) after 9 h of reaction. In
comparison, reaction of PTeO (0.16 mol L−1, nPTeO/nRh = 190) shows
that it is a little less reactive than XYL and the conversion reached 85%
after 9 h (Fig. 15a). This lower conversion can be explained by the
formation of only trace amounts of cyclic C5 compounds (mainly
THFA, < 0.005 mol L−1) during PTeO reaction (Figure S7), whereas
this reaction was observed from XYL (Fig. 6a). It could be suggested
that decreasing the number of hydroxyl groups in the molecule, low-
ered the probability to form cyclic compound. The evolution of each
category of linear deoxygenated C5 products shows that in parallel,
PTrO (first CeO bond cleavage) is the major product formed from PTeO
(Fig. 15c).
During the hydrogenolysis reaction, the concentration of PTrO at-
tained a maximum of 0.061 mol L−1 after 7 h (˜ 38% yield) and de-
creased to 0.033 mol L−1 at the end of the reaction. On the other hand,
PDO appeared after 0.5 h of the reaction and reached a maximum of
0.059 mol L−1(˜ 37%), whereas traces of pentanols were only formed
after 24 h of the reaction. The mass balance at the end of both reactions
was around 90%. The evolution of the concentration of each category of
C5 products during hydrogenolysis of XYL and PTeO confirms that the
formation of linear deoxygenated C5 products occurred through con-
secutive CeO bond cleavages of XYL. The analysis of the sequence of
appearance of the different categories of C5 deoxygenated products
clearly showed the sequential deoxygenation pathway:
Fig. 14. Dependence of product selectivity on SOR conversion under a) 30 bar
and b) 80 bar. ( ) C1-C5 products, ( ) Cyclic C6 products, ( ) Isomers, (
)
Deoxygenated C6 products without HPO, ( ) HPO. Reaction conditions: SOR
0.27 mol L−1, 120 mL H2O, 200 °C, 500 mg Rh-ReOx/ZrO2 catalyst.
Similar progression was observed during the hydrogenolysis of SOR.
Fig. 14 is an example of the evolution of carbon selectivity as a function
of SOR conversion under 30 and 80 bar of H2, at 200 °C. Under 30 bar,
as conversion increased from 23% to 89%, the selectivity to HPO de-
creased strongly from 40% to 18% whereas the selectivity to the other
C6 linear derivatives (with less oxygenated functions) increased from
10% to 24%. Similarly, under 80 bar the selectivity to HPO decreased
from 45% to 22% and the selectivity to deoxygenated linear C6
(without HPO) increased from 12% to 33% as conversion increased
from 33% to 95%. IDI and MAN underwent the same types of reactions
(dehydration, CeO and CeC bond cleavage) as SOR, but with lower
reactivity compared with SOR (Fig. 6b). Regardless of H2 pressure, the
selectivity to cyclic C6 products decreased slightly with the increase of
SOR conversion (from 22% to 17% under 80 bar and from 40% to 34%
under 30 bar H2). Independent hydrogenolysis of cyclic isosorbide was
conducted at 200 °C and exhibited low activity (26% conversion after
24 h). This is in good agreement with the relative stability of cyclic C6
compounds during the conversion of SOR. Finally, the selectivity to C1-
C5 products increased from 8% and 4% to 12% and 15%, respectively.
This is expected as there are secondary products obtained from CeC
bond cleavage reactions. These results indicate that successive deox-
ygenation reactions occur with time. We have previously evaluated the
hydrogenolysis of butanetriol: 1,2,4-BTO was almost as reactive as ERY
while 1,2,3-BTO was much less reactive. [11] To evaluate the relative
reaction rate of C5 and C6 polyols, we also compared the temporal
evolution of monodeoxygenated C5 or C6 products with the
This result agrees with that found for the transformation of ERY to
butanediols over Ir-ReOx/SiO2 [45] or Rh-ReOx/ZrO2 catalyst [20].
Butanediols were formed from ERY passing through butanetriols as
intermediates.
The reaction pathway of formation of the linear deoxygenated C6
compounds was investigated, by similarly comparing the temporal
evolution of each category of deoxygenated C6 products during the
hydrogenolysis of SOR and 1,2,3,4,6-HPO. HPO is the major family of
products formed from the beginning of reaction of SOR (Fig. 15e). The
concentration went through a maximum at 0.047 mol L−1 after 5 h,
then it decreased to 0.032 mol L-1 at 9 h. On the other hand, HTeO
started to appear after 0.5 h of the reaction, the concentration increased
up to a plateau at 0.020 mol L−1 after 7 h. HTrO and HDO were de-
tected only after 1 h and reached a concentration around 0.025 mol L−1
after 9 h. Lastly, HeOH was formed in traces.The final TOC measured
was of 88% (i.e. 12% loss).
During the hydrogenolysis of a mixture of hexanepentaols (HPO)
(Fig. 15f), first of all, HPO was converted gradually up to 96% con-
version after 24 h. So, under similar reaction conditions, HPO is much
less reactive than SOR (Fig. 15d), the difference was more significant
than in the case of XYL and PTeO (Fig. 15a). Once again we attributed
the difference to the absence of formation of cyclic compounds during
HPO reaction (Figure S8). HTeO, the major monodeoxy family formed
from HPO, appeared from the initial stage of the reaction and was
11

A. Sadier, et al.
AppliedCatalysisA,General586(2019)117213
Fig. 15. Evolution of the selectivity as a function of time: a) carbon selectivity during XYL and PTeO hydrogenolysis; d) carbon selectivity during SOR and HPO
hydrogenolysis. Evolution of the concentrations of each family of deoxygenated C5 products as a function of time during hydrogenolysis of b) XYL and c) 1,2,3,5-
PTeO; (◆) XYL; ( ) PTeO; ( ) Deoxy C5 from XYL; ( ) Deoxy C5 from PTeO; ( ) PTrO; ( ) PDO; ( ) PeOH. Evolution of the concentrations of each family of
deoxy C6 products as a function of time during hydrogenolysis of e) SOR and f) 1,2,3,4,6-HPO. (◆) SOR; ( ) HPO; ( ) Deoxy C6 from SOR; ( ) Deoxy C6 from HPO;
(
) HTeO; ( ) HTrO; ( ) HDO; ( ) HeOH. Reaction conditions: 200 °C, 50 bar of H2; a) and b) XYL 0.30 mol L−1, 120 mL H2O, 500 mg Rh-ReOx/ZrO2, nXYL/nRh
=
230; a) and c) PTeO 0.16 mol L−1, 33 mL H2O, 75 mg Rh-ReOx/ZrO2, nPTeO/nRh = 190; d) and e) SOR 0.27 mol L−1, 120 mL H2O, 500 mg Rh-ReOx/ZrO2, nSOR/nRh
= 200; d) and f) HPO 0.16 mol L−1, 30 mL H2O, 50 mg Rh-ReOx/ZrO2, nHPO/nRh = 270.
further converted after a maximum of 0.042 mol L−1 after 9 h (yield of
26%). HTrO and HDO appeared after 1 h of the reaction, while HTrO
was obtained in larger amounts than HDO. Traces of hexanols were
detected after 24 h of the reaction. The TOC measured after 24 h was
72%, so more gaseous compounds were formed compared to SOR re-
action.
As expected, the evolution of the concentration of each category of
C6 products during hydrogenolysis of SOR and HPO confirms the
multiple consecutive CeO bond cleavages of SOR
This multiple consecutive CeO bond cleavage was also described
very recently for the hydrogenolysis of mannitol or xylitol at 180 °C and
150 bar of H2 over Cu Raney® [27]. Moreover, this catalyst was shown
to be selective for a specific CeO bond cleavage, i.e. it mainly occurred
at the 3-position of mannitol to yield 1,2,4,5,6-HPO, and at a lesser
extent at the 6-position (1,2,3,4,5-HPO). The authors proposed a pre-
ferential adsorption via vicinal OH-groups and cleavage of OH-groups
in close proximity. These results are in good agreement with the
evolution of the selectivity with polyol conversion (Fig. 14).
2.8. Effect of support (TiO2 vs ZrO2)
The support may play a role on activity and selectivity of the re-
action, so we evaluated TiO2 which is also stable under the reaction
conditions. The effect of H2 pressure (in the range 30–120 bar) was also
investigated for ERY, XYL, and SOR hydrogenolysis at 200 °C over a
3.4 wt%Rh-2.7 wt%ReOx/TiO2 catalyst. At low pressure (30 bar),
during the hydrogenolysis of ERY, XYL, and SOR, initial reaction rates
were of 211, 160, and 157 mmolSub gRh−1 h−1, respectively. However,
as pressure increased from 30 to 120 bar, the initial rate over Rh-ReOx/
TiO2 increased slowly to 550, 319, and 296 mmolSub gRh−1 h−1, re-
spectively. These catalysts were used in a previous study and fully
characterized by XRD, TEM, XPS and CO chemisorption. The difference
in catalytic activity (initial rate) between the bimetallic catalysts over
both supports can be explained by the difference in metallic dispersion
12

A. Sadier, et al.
AppliedCatalysisA,General586(2019)117213
Fig. 16. Effect of H2 pressure on selectivity to the reaction products and %TOC
measured at 72–84% conversion. ( ) % TOC measured; ( ) C-C cleavage
products; ( ) Cyclic compounds; ( ) Isomers; ( ) polydeoxygenated products
C4 without BTO, C5 without PTeO, or C6 without HPO; ( ) monodeoxygenated
Fig. 17. Catalyst recycling study. ( ) Run 1; ( ) Run 2. Reaction conditions:
SOR 0.27 mol L−1 120 ml H2O, 200 °C, 80 bar of H2, 500 mg of Rh-ReOx/ZrO2.
C6 products without HPO: HTeO + HTrO + HDO + HeOH.
products BTO, PTeO, or HPO. Reaction conditions: reactant 0.27-0.40 mol L−1
,
3. Conclusion
120 mL H2O, 200 °C, 300–500 mg 3.4 wt%Rh-2.7 wt%ReOx/TiO2 catalyst;
reactant/nRh =. 200–450.
n
The effects of some reaction parameters were investigated for polyol
hydrogenolysis using three different substrates. Regardless of the sub-
strate, the responses in terms of catalytic activity and product se-
lectivity were similar. The polyols can undergo parallel reactions such
as epimerization, dehydration, CeC, and CeO bond cleavage. At 80%
conversion, the highest selectivity to linear deoxygenated C4, C5, and
C6 products was favored at high pressure of H2 in the presence of 3.7 wt
%Rh-3.5 wt%ReOx/ZrO2 catalyst (54–71% under 80 bar) at 200 °C. The
evolution of each family of linear deoxygenated C5 and C6 products
during the hydrogenolysis reactions of (XYL or SOR) and (HPO or
PTeO) confirms that the formation of polydeoxygenated C5 and C6
products was generated from the successive CeO bond hydrogenolysis
of XYL and SOR through HPO and PTeO as an intermediates. As a
general rule, the reactivity of partially deoxygenated compounds is
lower compared to the alditols. The optimization of reaction conditions
are needed to improve furthermore the selectivity to linear deox-
ygenated products and further work are under progress.
of rhodium over both supports. The Rh-ReOx/ZrO2 displayed a higher
CO/Rh ratio (0.80) than Rh-ReOx/TiO2 (0.25) [20].
Fig. 16 presents the effect of H2 pressure on the selectivity to dif-
ferent products, using the TiO2-supported catalyst. With respect to the
carbon balance, there is no difference between calculated TOC and
measured TOC in the hydrogenolysis of all susbstrates. Also, the mass
balance was > 85% under the range of H2 pressure applied (30 to
120 bar) in the presence of Rh-ReOx/TiO2 catalyst with < 10% carbon
selectivity to C-C cleavage products. As over ZrO2 support, the se-
lectivities to cyclic C4, C5 and C6 products were favored at low pressure
over Rh-ReOx/TiO2 catalyst (13% at 30 bar, 24% at 30 bar, and 21% at
80 bar, respectively). Significant high pressure (120 bar) was required
to observe lower amounts of cyclization products. The selectivity to the
monodeoxygenated products increased from 10 to 11% to 28–35%,
when the pressure was increased. This enhancement went along with a
small decrease in selectivity to polydeoxygenated products. Finally, the
epimerization reactions occurred under high pressure (120 bar, se-
lectivity of 12–14%). To conclude, there is no significant difference
between both supports neither concerning the activity nor the se-
lectivity to the different families of products.
4. Experimental section
4.1. Preparation of catalysts
Rh-ReOx/ZrO2 and Rh-ReOx/TiO2 (nRe/nRh = 0.4-0.5) were pre-
pared by consecutive impregnation, as reported previously [20]. In
summary, the first impregnation step was conducted by suspending the
support (ZrO2 MEL chemicals, SBET = 129 m2/g or DT-51D, Cristal,
2.9. Stability of bimetallic catalysts
The test of recyclability was investigated during SOR hydro-
genolysis in the presence of 3.4 wt.%Rh-3.2 wt.%ReOx/ZrO2 (Fig. 17).
After the first reaction, the solid was collected, washed with distilled
water, and dried at 110 °C overnight. The catalyst was re-activated
under a flow of H2 at 450 °C before the second run. The initial catalytic
activity was quite similar. Full conversion was attained after 6.5 h in the
presence of fresh catalyst (run 1), vs. a conversion of 93% after 11 h
during the second run. No appreciable leaching of rhodium and rhe-
nium metals was detected in the aqueous phase (< 0.06%) after the
first run. The deactivation of the catalyst may be due to the subsequent
washing of the filtrated catalysts, several times with distilled water
which caused the leaching of some rhenium as it was observed else-
where with bimetallic catalysts used for succinic acid hydrogenation
[46]. Analysis of the solid after washing revealed 2.5 wt.% Re vs.
3.2 wt.% in the fresh catalyst. Aqueous solubilization of rhenium spe-
cies was also reported In addition, no significant changes were observed
concerning the selectivity ( 7%) to deoxy C6 products (C6 without
HPO or HPO) in run 1 and 2.
S
BET = 91 m2/g) in an aqueous solution of RhCl3 (10 g L−1) in a flask;
the solution was stirred at room temperature for 7 h. After evaporation
of the solvent, the material was dried overnight in an oven at 110 °C
and calcined under air (60 mL min−1) at 4 °C min−1 to 500 °C for 3 h.
The second impregnation was conducted with an aqueous solution of
NH4ReO4 (10 g L−1). After calcination, the solids were reduced under
pure H2 (30 mL min−1) at 3.5 °C min−1 to 450 °C for 3 h; after cooling
and flushing with Ar, they were passivated under 1% v/v O2/N2 (30 mL
min−1) at room temperature for 30 min. The loading amount of Rh in
the bimetallics was between 3.4 and 3.7 wt.%. The additive (Re) was
added to give a nominal molar ratio (Re/Rh) around 0.5. The mono-
metallic 3.7 wt.% Rh/ZrO2 and 3.8 wt.% ReOx/ZrO2 were reduced di-
rectly after the first impregnation and calcination steps and then pas-
sivated. A commercial 5 wt.% Ru/C catalyst was purchased from Alfa
Aesar.
3.4 wt.%Pt-3.4 wt.%ReOx/ZrO2 and 3.7 wt.%Ir-3.9 wt.%ReOx/ZrO2
catalysts were prepared by consecutive impregnation as Rh based cat-
alysts. The basic characterization such as ICP, BET, and XRD were done.
13

A. Sadier, et al.
AppliedCatalysisA,General586(2019)117213
concentration at time t. The carbon selectivity Sit to a desired product
The percentage weights of noble metals (Pt and Ir) and Re were in the
range 3.4–3.7 wt% and 3.4–3.9 wt%, respectively. The BET surface area
of the catalysts were in the range 70-92 m2 g−1. Finally, the powder X-
ray diffractions of Pt-ReOx and supported on ZrO2 (Figure S9) showed
no diffraction peaks that can be assignable to the metallic phase sug-
gesting small crystallite sizes (< 5 nm).
was calculated according to Eq. (2):
[Product]ti × nCproduct i
Sit (%) =
× 100
([Sub]0 − [Sub]t ) × nCSub
(2)
where [Product]ti is the concentration of product i formed at time t and
nCproduct i is the number of carbon atoms in the product i.
The initial reaction rate was calculated using Eq. (3) (slope of the
curve at low conversion):
4.2. Catalytic testing
The hydrogenolysis reactions of ERY (Alfa Aesar, 99%), XYL (Sigma,
99%), and SOR (Sigma, 99%), were performed in a 300 mL batch
Hastelloy Parr 4560 autoclave set-up. In a typical experiment, an
aqueous solution (120 mL) of substrate (5 wt%) and 500 mg of catalyst
were loaded into the reactor. After sealing, the autoclave was purged
three times with Ar, pressurized with H2 to the required pressure
(PH2 = 30–120 bar), and heated to the reaction temperature (T
=200 °C) under stirring (1000 rpm). Liquid samples (1 mL) were col-
lected periodically during the course of the reaction. Small amounts of
H2 were added periodically to maintain the H2 pressure constant. At the
end of the reaction, the autoclave was cooled, the pressure was re-
leased, and the suspension was collected and filtered.
mmole of Sub
V (mmolSub g−Rh1 h−1 ) =
0
(3)
mass of Rh× time
The calculated TOC in liquid phase is determined by the sum of the
carbon concentrations of all compounds quantified by HPLC and GC at
time t (4):
n
⎡
⎛
⎞
Calculated TOCt ( g L )=MC ⎜∑ Nicarbon×Cti HPLC
−1
⎢
⎣
⎟
⎠
i
⎝
n
⎤
⎥
⎦
⎛
⎞
+ ∑ Nicarbon×Cti GC
⎜
⎝
⎟
⎠
i
(4)
1,2,3,5-PTeO product was formed by the reduction reaction of
0.17 mol L−1 2-deoxy-L-ribose (2 g, 40 mL H2O) under 50 bar of H2 at
100 °C for 24 h over 5%Ru/C. The product was analyzed by H+ column
with a purity of ˜97% after 24 h. Under similar reaction conditions
(60 bar of H2, 100 °C, 24 h over 5%Ru/C). 1,2,3,4,5-, 1,2,3,4,6-, and
1,2,3,5,6-HPO were synthesized by hydrogenation of L-rhamnose, 2-
deoxy-D-glucose, and 3-deoxy-D-glucose, respectively. The purity of the
products analyzed by H+ column after 24 h was in the range 95–97%.
Linear deoxygenated C4 (BTO and BDO), C5 (PTeO, PTrO, and
PDO), C6 products (HPO, HTeO, HTrO, and HDO), in addition to C2-C5
products (ERY, THR, 1,4-AE, 3OH-THF, GLY, 1,2-PrDO, and EG) were
analyzed using a Shimadzu LC 20A HPLC connected to a refractive
index detector (RID-10A). The separation of the products was achieved
using a column with a stationary phase (PS/DVB with SO3−, Rezex
ROA-Organic Acid H+, Ø 7.8 mm x 300 mm) heated at 40 °C. A solution
of H2SO4 (0.005 N) in ultra-pure ELGA water was used as a mobile
phase at a flow rate of 0.5 mL min-1. The separation of XYL and its
isomer (ARB), SOR and its isomers (MAN and IDI), cyclic C5 and C6
products [(1,4- and 1,5-AX), (1,4- and 3,6-S) and (2,5- and 1,5-AS)]
were analyzed using a Shimadzu LC 20A HPLC connected to a refractive
index detector (RID-10A). The separation of these products was
where MC is the molar mass of carbon (12 g mol−1), Ncarbon is the
number of carbon atom in compound i, and Cti is the concentration of
compound i at time t (in mol L−1). The comparison between TOC cal-
culated and TOC measured gives an indication about carbon balance. At
the end of each catalytic test, the final solution was analyzed by ICP to
measure the % of leached metal.
i
Acknowledgements
The financial support of this work was provided by Rhône-Alpes (AS
PhD grant), FCBA, and ANR (French Agence Nationale de la Recherche,
ANR-15-CE07-0017) within the French-Luxembourg Programme.
Appendix A. Supplementary data
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.apcata.2019.117213.
References
[1] T. Werpy, G. Petersen, A. Aden, J. Bozell, J. Holladay, J. White, A. Manheim,
D. Eliot, L. Lasure, S. Jones, Top Value Added Chemicals From Biomass. Volume I-
Results of Screening for Potential Candidates from Sugars and Synthesis Gas, DTIC
Document, (2004).
[2] A. Corma, S. Iborra, A. Velty, Chem. Rev. 107 (2007) 2411–2502.
[3] S. Liu, M. Tamura, Y. Nakagawa, K. Tomishige, ACS Sustain. Chem. Eng. 2 (2014)
1819.
achieved using a column with a stationary phase (PS/DVB with SO3−
,
Rezex RCM-Monosaccharide Ca+2 (8%), Ø 7.8 mm x 300 mm) heated at
60 °C. Elution was performed by 0.4 mL min-1 with water as a mobile
phase. The analysis of the C1-C6 mono alcohols in the liquid phase, in
addition to other C4 and C5 products [THFA, 1,2-PDO, 1,5-PDO, 2-
methyl-tetrahydrofuran (2-MTHF), and THF] was performed on a
Shimadzu GC-2010 system equipped with a flame ionization detector
and an autosampler. A Phenomenex FFAP 30 m x0.25 μm column was
used for separation of the products. The temperature was increased
from 40 °C to 100 °C at 5 °C min-1 (hold for 2 min), and then up to 250 °C
at 8 °C min-1. The response factor was determined experimentally for
the commercial compounds. The same value was then used for a family
of compounds.
[4] S. Liu, Y. Okuyama, M. Tamura, Y. Nakagawa, A. Imai, K. Tomishige,
ChemSusChem 8 (2015) 628.
[5] S. Liu, Y. Okuyama, M. Tamura, Y. Nakagawa, A. Imai, K. Tomishige, Green Chem.
18 (2016) 165.
[6] V. Monedero, G. Pérez-Martínez, M.J. Yebra, Appl. Microbiol. Biotechnol. 86
(2010) 1003–1015.
[7] M. Jeya, K.-M. Lee, M.K. Tiwari, J.-S. Kim, P. Gunasekaran, S.-Y. Kim, I.-W. Kim, J.-
K. Lee, Appl. Microbiol. Biotechnol. 83 (2009) 225–231.
[8] W. Rymowicz, A. Rywińska, M. Marcinkiewicz, Biotechnol. Lett. 31 (2009)
377–380.
[9] A. Sanborn, T. Binder, US2014/033586, 2015.
The concentration of Total Organic Carbon (TOC) was measured
using a Shimadzu TOC-VCSH analyzer, to check the possible formation
of gaseous products. This measurement was also compared to the cal-
culated TOC which is the total concentration of carbon in the liquid
phase quantified by HPLC and GC. If the values are close, the quanti-
fication by HPLC and GC is complete.
[10] Y. Cao, W. Niu, J. Guo, M. Xian, H. Liu, Sci. Rep. 5 (2015) 18149–18159.
[11] H. Mueller, W. Mesch, K. Broellos, US4973769 (1990).
[12] R. Hasegawa, K. Hayashi, US 4596886 (1986).
[13] P.J.C. Hausoul, A.K. Beine, L. Neghadar, R. Palkovits, Catal. Sci. Technol. 7 (2017)
56–63.
[14] Y.T. Kim, J.A. Dumesic, G.W. Huber, J. Catal. 304 (2013) 72–85.
[15] N. Ota, M. Tamura, Y. Nakagawa, K. Okumura, K. Tomishige, Angew. Chem. Int. Ed.
54 (2015) 1897–1900.
The substrate conversion was calculated from Eq. (1):
[16] K. Chen, M. Tamura, Z. Yuan, Y. Nakagawa, K. Tomishige, ChemSusChem 6 (2013)
613–621.
[Sub]0 − [Sub]t
[17] B. Blanc, A. Bourrel, P. Gallezot, T. Haas, P. Taylor, Green Chem. 2 (2000) 89–91.
[18] P.J.C. Hausoul, L. Negahdar, K. Schute, R. Palkovits, ChemSusChem 8 (2015)
3323–3330.
Conversion (%) =
× 100
[Sub]0
(1)
where [Sub]0 is the initial concentration of substrate, and [Sub]t is the
14

A. Sadier, et al.
AppliedCatalysisA,General586(2019)117213
[19] I. Furikado, T. Miyazawa, S. Koso, A. Shimao, K. Kunimori, K. Tomishige, Green
Chem. 9 (2007) 582–588.
R.J. Davis, J.A. Dumesic, J. Am. Chem. Soc. 133 (2011) 12675–12689.
[33] Y. Nakagawa, S. Liu, M. Tamura, K. Tomishige, ChemSusChem 8 (2015)
[20] A. Said, D. Da Silva Perez, N. Perret, C. Pinel, M. Besson, ChemCatChem 9 (2017)
2768–2783.
[21] C. Montassier, J.C. Ménézo, J. Moukolo, J. Naja, L.C. Hoang, J. Barbier,
J.P. Boitiaux, J. Mol. Catal. 70 (1991) 65–84.
[22] Y. Nakagawa, K. Tomishige, Catal. Sci. Technol. 1 (2011) 179–190.
[23] Y. Wang, J. Zhou, X. Guo, RSC Adv. 5 (2015) 74611–74628.
[24] M. Besson, P. Gallezot, C. Pinel, Chem. Rev. 114 (2014) 1827–1870.
[25] C. Montassier, J.C. Menezo, L.C. Hoang, C. Renaud, J. Barbier, J. Mol. Catal. 70
(1991) 99–110.
1114–1132.
[34] K. Tomishige, Y. Nakagawa, M. Tamura, Green Chem. 19 (2017) 2876–2924.
[35] Y. Nakagawa, M. Tamura, K. Tomishige, Res. Chem. Intermed. 44 (2018) 3879.
[36] Z. Shao, C. Li, X. Di, Z. Xiao, C. Liang, Ind. Eng. Chem. Res. 53 (2014) 9638–9645.
[37] T. Kawai, K.M. Jiang, T. Ishikawa, J. Catal. 159 (1996) 288–295.
[38] F. Schekler-Nahama, O. Clause, D. Commereuc, J. Saussey, Appl. Catal. Gen. 167
(1998) 237–245.
[39] T. Arai, M. Tamura, Y. Nakagawa, K. Tomishige, ChemSusChem 9 (2016)
1680–1688.
[26] X. Jin, J. Shen, W. Yan, M. Zhao, P.S. Thapa, B. Subramaniam, R.V. Chaudhari, ACS
[40] S. Koso, Y. Nakagawa, K. Tomishige, J. Catal. 280 (2011) 221–229.
[41] D. Urbas Branko, US4820880 (1989).
Catal. 5 (2015) 6545–6558.
[27] B. Kühne, H. Vogel, R. Meusinger, S. Kunz, M. Kunz, Catal. Sci. Technol. 8 (2018)
755–767.
[28] L. Ma, D. He, Catal. Today 149 (2010) 148–156.
[29] Z. Wang, B. Pholjaroen, M. Li, W. Dong, N. Li, A. Wang, X. Wang, Y. Cong, T. Zhang,
J. Energy Chem. 23 (2014) 427–434.
[42] Y. Jia, H. Liu, Catal. Sci. Technol. 6 (2016) 7042–7052.
[43] J. Xi, Q. Xia, Y. Shao, D. Ding, P. Yang, X. Liu, G. Lu, Y. Wang, Appl. Catal. B
Environ. 181 (2016) 699–706.
[44] E. Hagberg, K. Martin, E. Jan Van, J. Le Nôtre, D. S. Van Es, J. Van Haveren,
WO2013125950 A1 (2013).
[30] K. Chen, S. Koso, T. Kubota, Y. Nakagawa, K. Tomishige, ChemCatChem 2 (2010)
[45] Y. Amada, H. Watanabe, Y. Hirai, Y. Kajikawa, Y. Nakagawa, K. Tomishige,
ChemSusChem 5 (2012) 1991–1999.
547–555.
[31] T. Buntara, S. Noël, P.H. Phua, I. Melián-Cabrera, J.G. de Vries, H.J. Heeres, Top.
Catal. 55 (2012) 612–619.
[46] B.K. Ly, B. Tapin, F. Epron, C. Pinel, C. Especel, M. Besson, Catal. Today (2019),
https://doi.org/10.1016/j.cattod.2019.03.024.1.
[32] M. Chia, Y.J. Pagán-Torres, D. Hibbitts, Q. Tan, H.N. Pham, A.K. Datye, M. Neurock,
15