Selective C?O Hydrogenolysis of Erythritol over Supported Rh-ReOx Catalysts in the Aqueous Phase

Article abstract of DOI:10.1002/cctc.201700260

Bimetallic Rh-ReO_x (Re/Rh molar ratio 0.4–0.5) catalysts supported on TiO₂ and ZrO₂ were prepared by the successive impregnation of dried and calcined unreduced supported Rh catalysts. Their catalytic performances were evaluated in the hydrogenolysis of erythritol to butanetriols (BTO) and butanediols (BDO) in aqueous solution at 150–240 °C under 30–120 bar H₂. The activity depended on the nature of the support, and the highest selectivity to BTO and BDO at 80 % conversion was 37 and 29 %, respectively, in the presence of 3.7 wt %Rh-3.5 wt %ReO_x/ZrO₂ at 200 °C under 120 bar. The characterization of the catalysts by CO chemisorption, TEM with energy-dispersive X-ray spectroscopy, thermogravimetric analysis with MS, and X-ray photoelectron spectroscopy suggests a different distribution and reducibility of Re species over the supported Rh nanoparticles, which depends on the support.

Full text of DOI:10.1002/cctc.201700260

Accepted Article
Title: Selective C-O hydrogenolysis of erythritol over supported Rh-
ReOx catalysts in aqueous phase
Authors: Achraf Said, Denilson Dasilva Perez, Noémie Perret,
Catherine Pinel, and Michèle Besson
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemCatChem 10.1002/cctc.201700260
Link to VoR: http://dx.doi.org/10.1002/cctc.201700260
A Journal of
www.chemcatchem.org

ChemCatChem
10.1002/cctc.201700260
Selective C-O hydrogenolysis of erythritol over supported Rh-ReO
x
catalysts in aqueous phase.
[
a]
[b]
[a]
[a]
Achraf Said, Denilson Da Silva Perez, Noémie Perret, Catherine Pinel,* and Michèle
Besson*
[
a]
[a] – Univ Lyon, Univ Claude Bernard, CNRS, IRCELYON, UMR5256, 2 Avenue Albert
Einstein, F-69626 Villeurbanne, France
[b] – Institut FCBA, InTechFibres, 38044 Grenoble, France
*
E-mail: catherine.pinel@ircelyon.univ-lyon1.fr, michele.besson@ircelyon.univ-lyon1.fr
Scissors for C-O bonds. Selective C-O hydrogenolysis is a challenging strategy to remove the
hydroxyl groups of oxygenated biomass-derived molecules. Reaction pathways for erythritol
in water were studied over TiO
2
- and ZrO
2
-supported Rh-ReOx catalysts. Using ZrO as
2
support, a selectivity up to 66% butanetriols and butanediols was achieved at 80% conversion.
1
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
Abstract
Bimetallic Rh–ReO
x
(Re/Rh molar ratio 0.4-0.5) catalysts supported on TiO
2
and ZrO were
2
prepared by successive impregnation of dried and calcined unreduced supported Rh catalysts.
Their catalytic performances were evaluated in the hydrogenolysis of erythritol to butanetriols
(
BTO) and butanediols (BDO) in aqueous solution at 150-240°C under 30-120 bar H . The
2
activity was dependent on the nature of the support and the highest selectivity to BTO and BDO
at 80% conversion reached 37 and 29%, respectively, in the presence of 3.7wt%Rh–
3
.5wt%ReO
x
/ZrO catalyst at 200°C under 120 bar. Characterizations of the catalysts by CO
2
chemisorption, TEM-EDX, TGA-MS, XPS suggest a different distribution and reducibility of
Re species over the supported Rh nanoparticles depending on the support.
Keywords
Biomass; aqueous phase; butanetriols and butanediols; erythritol hydrogenolysis; supported
Rh-ReOx catalysts
2
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
1
. Introduction
Over the last few decades, society has been moving toward the use of lignocellulosic biomass
on a large scale for the production of chemicals and fuels.[1–4] Indeed, platform chemicals
derived from the carbohydrate fraction (cellulose and hemicellulose), such as sugars (e.g.
glucose), sugar alcohols (e.g. sorbitol) and cyclic ethers (e.g. tetrahydrofurfuryl alcohol via
furfural) are promising alternatives to unsustainable fossil resources for the production of high-
value added chemicals. The platform molecules are highly functionalized and contain an O/C
ratio higher than most of the commodity chemicals, thus, their transformation requires to lower
the O content. Among the deoxygenation methods, hydrogenolysis of alcoholic C-O bond is of
particular significance. Performing this reaction efficiently, and also selectively eliminating
either a primary or a secondary alcohol, is necessary, however C-O bond hydrogenolysis is not
an easy reaction to achieve. A number of supported catalytic systems have been studied and
reported to be selective for the partial dehydroxylation of glycerol towards 1,2-propanediol[5,6]
(
up to 98%) and 1,3-propanediol[7–9] (up to 67%). Nevertheless, the efficient and selective
removal of OH groups in superior polyols with more OH groups such as erythritol, xylitol, and
sorbitol remains very challenging.[10–12]
Erythritol (ERY) is a promising platform chemical for the production of butanetriols and diols
as C4 building blocks. Chemical[13] and fermentative[14,15] processes have been investigated for
its production in large-scale. ERY can be produced by catalytic hydrogenation of dialdehyde
starch over nickel catalyst at 160-200°C under 140 bar of H , but this process has not been
2
industrialized due to safety and environmental reasons.[13] The standard industrial production
routes involve the microbial fermentation of sugar and sugar alcohols such as glucose and
glycerol using various yeasts. The highest yields reported for the production of ERY are of 61%
from glucose[14] and 56% from glycerol.[15]
The literature dealing with the production of C4-triols and diols from sugar alcohols is
limited.[10–12] During erythritol hydrogenolysis, four types of reactions can occur: C-O bond
hydrogenolysis, C-C bond hydrogenolysis, dehydration and epimerization as shown in Scheme
1
. Erythritol is converted to the desired butanetriols (1,2,3-BTO, 1,2,4-BTO) and the series of
butanediols (BDO) by C-O bond cleavage. Over-hydrogenolysis of butanediols yields 1-
butanol (1-BuOH)[16] and 2-butanol (2-BuOH).[17] In addition, dehydration of erythritol can
occur and yields 1,4-anhydroerythritol (1,4-AE); the successive C-O hydrogenolysis generates
the formation of 3-hydroxytetratydrofuran (3OH-THF) and tetrahydrofuran (THF). C-C
3
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
hydrogenolysis reactions lead to many types of C2 and C3 products such as ethylene glycol
EG), glycerol (GLY), propylene glycol (PG) and propanol (PrOH). Finally, erythritol can also
(
undergo epimerization reaction and form threitol.
The butanetriols and butanediols are all valuable products. 1,2,4-BTO, currently obtained
industrially by catalytic hydrogenation of malic esters, is used in the production of polyurethane
foams, butanetriol trinitrate (a propellant), pharmaceuticals, and high quality inks.[18,19] 1,4-
BDO is employed as solvent, in the manufacture of various polymers, and for the synthesis of
-butyrolactone and THF. 1,2-BDO is generated as a by-product during the fabrication of 1,4-
butanediol from butadiene. It is a precursor for the production of low viscous polyester and
plasticizers.[20] 2,3-BDO has many applications in the printing inks, pharmaceuticals, perfumes,
and pesticides industries. It is produced via pyruvate from the microbial fermentation of
monosaccharides (hexoses or pentoses).[21]
1
4
2
2
2
3
2
4
2
2
1
2
3
4
Epimerization reaction
C-O bond hydrogenolysis
Dehydration reaction
C-C bond hydrogenolysis
Scheme 1. Possible reaction routes associated with the conversion of erythritol under
hydrogenolysis conditions.
4
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
The first studies dealing with the conversion of erythritol were conducted over conventional
hydrogenolysis catalysts such as supported Ru and Cu.[22,23] Montassier et al[22] reported that
Cu-Ru/C catalyst favored predominantly dehydration reaction, with formation of 1,4-
anhydroerythritol, and traces of C2-C3 products by C-C bond hydrogenolysis, when working
at 260°C under 40 bar of H . Moreover, under similar reaction conditions, Ru/S-modified
2
carbon (S/Ru=1) and Raney Cu exhibited initial selectivities towards butanediols (1,2-
BDO+2,3-BDO) of 33% and 54%, respectively.[23]
Since, it has been shown that the combination of an oxophilic promoter (Re, Mo or W) and a
highly reducible noble metal (Ir, Rh, or Ru) enhanced the catalytic performance in the selective
C-O bond hydrogenolysis of polyols or cyclic ethers (tetrahydrofurfuryl alcohol and
tetrahydropyran-2-methanol).[24–28] For example, C and SiO
shown to be active and selective in C-O hydrogenolysis of 1,2,6-hexanetriol to 1,6-
hexanediol[27] and 1,2,4-butanetriol to 1,4-butanediol.[28] In Dumesic’s group work, Rh-ReO
/C
-supported Rh-ReO catalysts were
2
x
x
proved to be selective in the hydrogenolysis of secondary C-O bonds during hydrogenolysis of
polyols and cyclic ethers derived from biomass.[28] There are fewer studies dealing with the use
of these supported bimetallic catalysts for hydrogenolysis of erythritol. Tomishige’s group
studied the effect of the nature of the noble metal (Ir and Rh) and of the oxophilic promoter
(
Re, W, and Mo) over SiO
including H pressure (20-80 bar), temperature (80-140°C), erythritol concentration (5-67wt%
aqueous solution), acid additive (nH+/nIr = 1), and metal particle size, on the selectivity towards
butanediols over Ir-ReO /SiO catalysts. Under optimum conditions (100°C, 80 bar of H
H+/nIr=1), the maximum selectivity towards butanediols in water reached 48% at 74%
conversion. In another study, a selectivity of 40% towards butanediols was observed at 91%
conversion when using Ir-ReO /SiO combined with H-ZSM-5 in water, at short reaction time
120°C, 80 bar of H
).[12] Really high selectivity in butanediols was reported in the presence of
ReO -Pd/CeO (93% selectivity at 98% conversion), however the reaction was conducted with
2
support.[10] They also investigated the effect of reaction parameters
2
x
2
2
,
n
x
2
(
2
x
2
dioxane as solvent.[11]
In this work, mesoporous TiO
in water under hydrogenolysis conditions.[29,30] Rh-ReO
2
and ZrO
2
were chosen as the supports because of their stability
/TiO and Rh-ReO /ZrO were
investigated for the selective hydrogenolysis of erythritol to butanetriols and diols in batch
reactor. The influence of reaction conditions (e.g. temperature, pressure of H , acid additive,
x
2
x
2
2
effect of support) were studied in details. We also tried to get better insight on the reaction
mechanism associated with the formation of butane- triols and diols.
5
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
2
. Results and discussion
2
.1. Catalyst characterization
The properties (chemical composition, BET surface area, CO chemisorption) of the synthesized
catalysts are presented in Table 1.
Table 1. Physicochemical properties of the synthesized catalysts
[b]
Catalyst
Rh-Re loading [a]
wt%)
n
Re/nRh
S
BET
CO/Rh
2
-1
[c]
(
(molRe/molRh
-
)
(m .g ) (mol/mol)
Rh/TiO
2
3.7-0.0
3.6-0.0
3.2-2.5
3.7-2.6
3.7-3.5
83
97
0.27
0.74
-
Rh/ZrO
2
-
Rh-ReO
Rh-ReO
Rh-ReO
x
x
x
/SiO
/TiO
/ZrO
2
0.4
0.4
0.5
200
85
2
0.25
0.80
2
98
[
[
[
a]
b]
c]
based on ICP-OES analysis
from N physisorption, SBET TiO
Calculated from CO chemisorption measurements
2
-1
2
-1
2
-1
2
2 2 2
= 91m .g , SBET ZrO = 129 m .g , SBET SiO = 200 m .g
The analysis of the metal and promoter contents of the catalysts showed that complete
deposition occurred. The Rh and Re loadings were in the range 3.2-3.7wt% and 2.5-3.5wt%,
respectively. The molar ratio between metallic Rh and oxophilic promoter (Re) was fixed at
0
.4-0.5 based on the literature. Indeed, the catalytic performances of Re promoted catalysts for
hydrogenolysis reactions have been shown to depend on Re content; the curve usually exhibits
a volcano shaped trend, where the maximum depends on the nature of the noble metal. For the
conversions of glycerol to 1,3-propanediol[31] and tetrahydrofurfuryl alcohol to 1,5-
pentanediol,[32] optimum Rh/Re ratios of 0.5 were determined; the presence of high amount of
Re can decrease the amount of Rh surface atoms, hence the catalytic activity.
The BET surface area of mesoporous (TiO
2
and ZrO
2
) and non-porous (SiO -Aerosil) supports
2
2
-1
were of 91, 129 and 200 m g , respectively; the BET surface area of the catalysts were not
significantly changed after incorporation of the metals.
6
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
Figure 1 shows the diffraction patterns of Rh/TiO
2
x
and Rh-ReO /TiO
2
catalysts compared to
the pristine support. The two patterns exhibited no peaks other than those associated to the
anatase TiO support, i.e. no discrete Rh and Re containing phases were observed suggesting
high dispersion. Figure 2 shows the patterns of Rh/ZrO and Rh-ReO /ZrO catalysts. No size
of Rh metal crystallites could be estimated due to the overlap between peaks of ZrO and the
main peak of metallic Rh at 41°, whereas no diffraction peak assignable to Re metal or oxide
was observed. Finally, the XRD pattern associated with Rh-ReO /SiO (Figure S1) exhibited a
2
2
x
2
2
x
2
wide and weak diffraction peak at 2-Theta = 41°. The crystallite size of metallic Rh was
approximately 5 nm as determined by using the Scherrer equation from the half-width of the
XRD broadening peak.
Figure 1. XRD patterns for TiO
2
(1), Rh/TiO
2
(2), Rh-ReO
x
/TiO (3).
2
Figure 2. XRD patterns for ZrO
2
(1), Rh/ZrO
2
(2), Rh- ReO
x
/ZrO (3).
2
7
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
Figure 3 shows TEM images of the mono- and bi-metallic catalysts supported on TiO
corresponding size histograms. The images for the ZrO -supported catalysts are shown in
Figure S2. For each sample, approximately 250 particles were measured for estimation of the
average particle sizes. The average particle size in Rh/TiO was 1.6 nm. On the other hand, the
Rh/ZrO sample showed no evidence of any particles, indicating the presence of tiny
2
and the
2
2
2
nanoparticles of Rh, below ca. 1 nm. After deposition of rhenium (Fig. 3(b) and Fig. S2(b)),
particles were detected on both supports. However, the particle size distribution of these
particles showed that the mean size on ZrO
2
(3.1 nm) was larger than on TiO (1.5 nm).
2
50
40
30
20
10
0
1
.62 ± 0.37 nm
0
.5-1 1-1.5 1.5-2 2-2.5 2.5-3 3-3.5 3.5-4
d/ nm
50
40
30
20
10
0
1.50 ±0.50 nm
0
.5-1 1-1.5 1.5-2 2-2.5 2.5-3 3-3.5 3.5-4
d/ nm
Figure 3. Representative TEM images and size distribution associated with (a) Rh/TiO
b) Rh-ReO /TiO
2
and
(
x
2
.
Using energy dispersive X-ray spectroscopy (EDS) analysis over several randomly selected
images of Rh-ReO /TiO both Rh and Re were present with a distribution of Re/Rh atomic
x
2
ratios within 0.10-0.31, i.e. slightly lower than the ratio of 0.4 determined by ICP-OES analysis.
8
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
The difference suggests that some zones are likely to contain very dispersed monometallic Re
nanoparticles that are too small to be detected by TEM. This result is consistent with the
observation of tiny Re nanoparticles on Pd-ReO
x
/TiO
prepared using a similar procedure.[33]
2
In turn, in the case of the Rh-ReO /ZrO catalyst, the Re/Rh ratios measured were mainly in the
x
2
range 0.33-0.45, i.e. closer to the ratio of 0.5 determined by ICP-OES analysis; some rare zones
having a low density of particles contained Rh rich particles. This result rather suggests Re in
close proximity to Rh nanoparticles over ZrO
2
. Moreover, the Rh-ReO
x
particles on the ZrO
2
support were larger than on TiO , which would also indicate that the Re species decorate the
2
Rh particles or that small aggregates of Rh nanoparticles and rhenium oxide species with an
estimated size of 3.1 nm were formed.
Figure 4 shows the profiles of H
2
O associated with the H
/TiO and Rh-ReO
2
-thermogravimetric analysis/mass
spectrometry (TGA/MS) of the Rh-ReO
x
2
x
/ZrO catalysts. The mass spectrum
2
of the evolved gas captured was recorded as a function of temperature. The only signals
observed were attributed to m/z = 17 and 18 which correspond to the formation of water due to
the reduction. In the literature, temperature-programmed reduction (TPR) profiles of a series of
[34]
[35]
Rh catalysts supported on TiO
80°C was high enough to reduce all rhodium oxide to metallic state. The Rh-Re/TiO
catalysts presented different profiles of reduction temperature. The signal for Rh-ReO
2
and ZrO
2
showed that a reduction temperature lower than
1
2
and ZrO
2
2
x
/TiO
showed three main hydrogen consumption peaks. A large reduction peak centered at 100°C can
be ascribed to the reduction of rhodium oxide, but also to the reduction of rhenium oxide in
interaction with Rh; this is consistent with previous studies which report the lowering of the
reduction of Re in the presence of metallic Rh.[28,31] For instance, the temperature-programmed
reduction (TPR) profile of a Rh-ReO
which is close to the values provided here.[31] A wide consumption peak exhibited between 300
and 400°C is attributed to the reduction of ReO species; indeed, the peak temperature of Re
oxidized species in supported monometallic rhenium catalysts usually appears in the range of
00 to 375°C.[28,33,36,37] The broad shoulder up to 500°C corresponds to the reduction of bulk
titania support.[33] These findings suggest that a large fraction of Re species was present as
isolated clusters on the TiO support and reduced without promotion by Rh. On the other hand,
Rh-ReO /ZrO displayed a broad peak associated with hydrogen consumption between 50 and
50°C centered at 150°C, with no discernible reduction peak corresponding to isolated ReO
x
/SiO
2
(Re/Rh = 0.5) showed one reduction peak at 120°C,
x
3
2
x
2
2
x
species. This observation means that in that solid, the contact between Rh and Re has been
established and the presence of metallic Rh promotes the reduction of Re species.
9
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
(
2)
(
1)
100 200 300 400 500 600 700 800 900
Temperature /°C
Figure 4. H
2
O MS signal for (1) Rh-ReO
x
/TiO2, (2) Rh-ReO
x
/ZrO
2
X-ray photoelectron spectroscopy (XPS) analysis was performed on the as-prepared precursors,
the solids reduced in situ in the XPS chamber, and then after passivation with flowing 1vol%
O
2
in N
displayed in Figure 5, and those of the Rh-ReO
compiles the chemical state of the catalyst components and their relative abundance in the
different materials.
2
at room temperature. The Rh 3d and Re 4f spectra of the Rh-ReO
x
/TiO solids are
2
x
/ZrO catalyst are depicted in Figure S3. Table
2
2
1
0
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
Figure 5. XPS spectra for Rh 3d and Re 4f core levels of the TiO
2
-supported Rh/ReO
x
materials: (a) in catalyst precursor, (b) after reduction, and (c) after passivation.
Table 2. Distribution of Rh and Re oxidation states and surface atomic ratio by XPS.
Treatment
Distribution of oxidation states (%)
Rh3+ Rh+ Rh0 Re7+ Re6+ Re4+ Re3+ Re0
Rh-ReO /TiO
(Re/Rh = 0.4)[a]
26
Re/Rh
x
2
Precursor 74
-
73
27
26
25
-
-
-
0.61
In situ reduction -
Exposure to air -
19 81
70 30
-
41
-
33 0.67
70
5
-
-
0.42
Rh-ReO
x
/ZrO
2
(Re/Rh = 0.5)[a]
80 20
Precursor 82
18
-
-
-
0.64
1
1
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
In situ reduction -
Exposure to air -
16 84
57 43
-
33
45
-
-
67
-
-
-
0.75
0.78
55
[
a]
as measured by ICP-OES analysis
In the two catalyst precursors, the Rh 3d profile fitting showed the characteristic doublets of
3
+
+
[38]
Rh (BE = 308.2 eV) and Rh (BE = 309.1 eV).
Further, the Re 4f profile could be
decomposed in two doublets with Re 4f7/2 components appearing at ca. 46.2 eV and 44.2 eV,
indicating the presence of Re7+ and Re6+ with estimated percentages in the two solids of
approximately 75% and 25%, respectively.
After in situ reduction of the bimetallic catalyst precursors, the binding energy of the Rh 3d5/2
0
at 307.6 eV shows that metallic Rh was the major rhodium species (81% and 84% of the Rh
+
was reduced on TiO
2
and ZrO
2
, respectively), although a certain proportion of Rh was still
present. However, there was some disparity in Re oxidation state. While over ZrO
2
all Re
it was
6
+
3+
remained fully oxidized after reduction at 450°C (33% Re and 67% Re ), over TiO
2
0
6+
3+
partially reduced to Re (33%) in addition to Re (26%) and Re (41%), suggesting Rh-Re
interactions in the latter material. Surprisingly, in both materials rhodium could not be reduced
0
entirely to Rh after reduction at 450°C. It looks like that the presence of Re oxides prevented
the total reduction of Rh.
0
After subsequent exposure to ambient air of the reduced Rh-ReO
x
/TiO
2
catalyst, Rh was partly
0
+
0
re-oxidized (30% Rh and 70% Rh instead of 81% Rh ) as well as Re, which was analysed as
7
+
6+
4+
Re (70%), Re (25%) and a small percentage of Re (5%). In the same way, measurements
of the Rh-ReO /ZrO material after reduction and exposure to air revealed that no more than
3% of Rh was present as Rh . Re composition was 55% Re and 45% Re . The re-oxidation
x
2
0
7+
6+
4
0
of Re is consistent with the known oxophilic character of Re.
Finally, no significant change in the Re/Rh atomic ratio of the catalysts was seen by XPS
regardless of the nature of the support (TiO
2
or ZrO ) and whether the catalyst was the precursor
2
or the reduced material. Moreover, while the XPS Re/Rh ratio was close to 0.6-0.8, the actual
global Re/Rh ratio was of 0.4-0.5, which indicates some segregation of Re at the surface of the
two catalysts. In contrast, exposure to air of the TiO
.42, suggesting some redistribution.
2
-supported catalyst decreased this ratio to
0
1
2
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
The extent of irreversible CO chemisorption was measured for the reduced Rh/TiO
Rh-ReO /TiO , and Rh-ReO /ZrO catalysts. The CO adsorption amounts are shown in Table
. The monometallic Rh/ZrO displayed a higher CO/Rh ratio (0.74) than Rh/TiO (0.27),
indicating a much higher Rh dispersion on zirconia. This is consistent with the TEM
characterizations, which showed no detectable Rh nanoparticles on ZrO , while a mean particle
size of 1.6 nm was determined on TiO (Fig. 3 and Fig. S2). After deposition of Re on Rh/TiO
and Rh/ZrO , the CO uptake did not change significantly compared with the monometallic
2
, Rh/ZrO
2
,
x
2
x
2
1
2
2
2
2
2
2
catalysts. The CO/Rh ratio was 0.25 and 0.80, respectively, suggesting that the amount of
surface metallic atoms was almost the same after addition of Re.
Though part of Re was metallic in the Rh-ReO
x
/TiO
2
catalyst after reduction, the CO/Rh ratio
may be influenced by different
was unaffected by Re loading. The CO chemisorption on TiO
2
factors. Rh particles may be partly covered by Re species during the preparation method.
Decoration or partial encapsulation of Rh particles by the support through a strong metal-
support interaction (SMSI) effect known for titania-supported metals exposed to reducing
conditions may occur.[39] These two factors would lower the number of accessible metallic Rh0
0
sites on TiO
2
. On the other hand, the presence of metallic Re , which is able of chemisorbing
CO, will positively affect the amount of CO.
On the other hand, considering that Re oxides do not chemisorb CO and the absence of metallic
Re° in the reduced Rh-ReO
x
/ZrO
2
(Table 2, Fig. S3), this would indicate no covering of Rh
. This observation is not in line with the TGA-MS results
surface with Re species over ZrO
2
(
Fig. 4) and the EDS data which rather indicated an interaction between Rh and Re in the Rh-
ReO /ZrO catalyst. Furthermore, it is to be noted that the ZrO support needs much higher
temperatures than TiO to be reduced, and no SMSI effect is expected Regarding Rh-ReO /ZrO
the reader should keep in mind that there is no metallic Re , that Re oxides do not chemisorb
CO, and that ZrO support needs much higher temperatures than TiO to be reduced, hence no
SMSI effect is expected.[40] Therefrom, considering that (i) the same CO/Rh ratio was obtained
for Rh/ZrO and Rh-ReO /ZrO despite different particle sizes estimated from TEM, (ii) the
x
2
2
2
x
2
°
2
2
2
x
2
TGA-MS results (Fig. 4) and the EDS data indicated an interaction between Rh and Re, it is
proposed that Rh might aggregate with Re oxide species.
Overall, the different characterizations of the materials suggest differences in the distribution
of Rh and Re and in the reducibility of Re species on TiO
could not be proposed.
2
and ZrO , though a detailed structure
2
1
3
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
2
.2. Catalytic results
.2.1. Typical evolution of products as a function of time in the hydrogenolysis of
2
erythritol
First, the effect of the amount of catalyst (expressed as mmol of Rh introduced) on initial
reaction rate associated with erythritol disappearance (based on the first two hours of the
reaction) was studied under 80 bar of hydrogen at 200°C in the presence of Rh-ReO
x
/ZrO .
2
Without catalyst, the reaction did not take place. As the amount of rhodium increased from 0.07
to 0.18 mmol (7.2 mg to 18.5 mg), the initial rate of disappearance of ERY increased from
-1
0
.005 to 0.015 molERY h . The linear relationship (Figure S4) confirms that the reaction rates
were obtained in kinetic region without mass transfer resistance.
Figure 6 shows a typical example of the temporal dependence of product concentrations when
-1
the reaction (0.4 mol L erythritol) was performed under 80 bar of hydrogen at 200°C, over 0.3
g Rh-ReO /TiO catalyst. Erythritol was converted smoothly to reach 81% conversion after 25
h. Butanetriols and butanediols were the main products obtained. They all appeared from the
x
2
-1
-1
initial stage of the reaction and their concentration attained 0.07 mol L and 0.11 mol L ,
respectively; the time course of the reaction suggests that it is not necessary to go through 1,4-
anhydroerythritol to produce the triols and diols. Threitol concentration progressively increased
-1
to reach a plateau at 0.04 mol L from 9.5 h; this indicates a lower reactivity of the threitol
isomer compared with erythritol. Cyclic products (1,4-AE, 3OH-THF, THF), butanol (BuOH),
C3 products (GLY, PG, 1-PrOH), and ethyleneglycol (EG, the latter is not shown) were formed
-1
continuously in very low concentrations (< 0.04 mol L ).
0.5
0.4
0.3
0.2
0.1
0.0
0.16
0.12
0.08
0.04
0.00
0
5
10
15
20
25
Reaction time / h
Figure 6. Evolution of the concentrations of the different categories of products as a function
of time during erythritol hydrogenolysis. () Erythritol; () Cycles;() C3 products; ()
Threitol; () BuOH; ( ) BDO; () BTO.
1
4
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
-1
Reaction conditions: ERY 0.4 mol L , 120 mL H
2
O, 200°C, 80 bar of H
2
, 0.3 g Rh-ReO
x
/TiO .
2
Figure S5 presents the temporal evolution of the TOC values calculated from HPLC analysis in
the liquid phase (TOC calculated), on one hand, and obtained from direct analysis using the
TOC analyzer (TOC measured), on the other hand. At time zero, the calculated and measured
-1
TOC values of prepared solution were 19.0 and 18.8 g L , respectively. During the
C
hydrogenolysis reaction, no difference was observed between the two values indicating that all
the products in aqueous phase were analysed by HPLC. These values decreased slightly during
reaction (4% loss after 25 h), meaning that a small amount of gaseous products was formed due
to over-hydrogenolysis.
2
.2.2. Screening of rhodium catalysts for hydrogenolysis of erythritol at 200°C under 80 bar
H
2
The catalytic performances of the various supported mono- and bimetallic catalysts are
summarized in Table 3.
The bare supports (TiO
2
, ZrO ) showed negligible activity for hydrogenolysis of erythritol (8%
2
and 1% conversion after 24 h, respectively); only cyclic compounds (1,4-AE, 3OH-THF and
THF) were formed presumably because of some intrinsic acidity of the supports. The absence
of metallic rhodium for H dissociation explains that no C-O and C-C bond cleavage occurred.
2
Table 3. Conversion (Conv.) and carbon selectivities observed for the hydrogenolysis of
erythritol over rhodium-based catalysts. Selectivities are given at final conversion.
-1
Reaction conditions: ERY 0.4 mol L , 120 ml H
2
O, T=200°C, 80 bar of H
2
, 0.3 g catalyst, BTO
=
=
butanetriols, BDO = butanediols, BuOH = butanols, Cycles = 1,4-AE+3OH-THF+THF, C3
GLY+PG+PrOH.
Catalyst
Molar ratio Time Conv.
Carbon selectivity (%)
Re/Rh
-
[h]
25
[%] BTO BDO BuOH Threitol Cycles C3
Rh/TiO
2
10
70
67
1
11
11
17
0
2
1
0
3
88
44
3
0
12
7
Rh/ZrO
2
-
24
31
17
28
Rh-ReO
x
/SiO
2
0.4
40
1
5
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
Rh-ReO
x
/TiO
2
0.4
0.5
25
8
80
82
24
28
36
26
5
4
11
8
11
12
8
Rh-ReO
x
/ZrO
2
11
Rh/TiO
analyzed were mainly cyclic compounds (88% selectivity). In contrast, the deposition of Rh on
ZrO was associated with an increase in activity (70% conversion after 24 h); the cyclization
2
exhibited a very low activity, resulting in only 10% conversion after 25 h; the products
2
reaction was still predominant with a 44% selectivity to cyclic compounds at final conversion.
C3 compounds (selectivity of 12%), BTO (selectivity of 17%) and BDO (selectivity of 11%)
products were also generated. The lower conversion over Rh/TiO
explained by the different metallic dispersion of Rh over the supports as shown in Table 1, in
which the CO adsorption amounts on Rh/ZrO was higher than that on Rh/TiO (CO/Rh was
.74 and 0.27, respectively). The difference of dispersion was confirmed by TEM analysis (Fig.
and Fig. S2a).
2
than over Rh/ZrO
2
can be
2
2
0
3
It was reported in the literature that the addition of ReO
Rh/SiO ) enhanced the catalytic activity of glycerol hydrogenolysis in water and promoted the
formation of deoxygenated products (1,2- and 1,3-propanediol).[31] In the present investigation,
erythritol conversion attained 67% over Rh-ReO /SiO catalyst, after 31 h. The epimerization
x
species to supported rhodium metal
(
2
x
2
of erythritol to threitol was observed with a selectivity of 40%. The selectivity to linear
deoxygenated products (BTO + BDO + BuOH) by C-O hydrogenolysis attained 46%. However
the silica support was unstable under the hydrothermal conditions used (200°C); leaching of
2
4% of Si was detected by ICP-OES analysis of the final solution. For that reason we moved
to supports known to be stable in water at high temperature (TiO , ZrO
2
2
).[29,30]
Modification of Rh/TiO
of erythritol; the conversion increased from 10% over Rh/TiO
5 h. The selectivity to deoxygenated products (BTO + BDO + BuOH) was of 65% at final
conversion. The hydrogenolysis activity was also enhanced when working with ZrO as
support; the reaction time needed for attaining 82% conversion was only of 8 h over Rh-
ReO /ZrO and the final selectivity to the deoxygenated products was 58%. These corroborate
the efficiency of the combination of Rh and ReO for activity and selectivity to deoxygenated
products. The results are in agreement with reports on hydrogenolysis of tetrahydropyran-2-
2
with ReO
x
(nRe/nRh=0.4) enhanced the activity for the hydrogenolysis
2
to 80% over Rh-ReO /TiO after
x
2
2
2
x
2
x
1
6
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
methanol to 1,6-hexanediol in which the catalytic activity as well as the selectivity were
enhanced over metal oxide modified supported Rh catalysts.[26–28]
2
.2.3. Temperature effect over Rh-ReO
x
/ZrO
2
The effect of reaction temperature was studied over the range 150-240°C under 80 bar of H
2
in
the presence of Rh-ReO /ZrO . The evolution of erythritol concentration as a function of time
x
2
is shown in Figure 7a, the selectivities to the different products at 80% conversion are shown
in Figure 7b.
0.5
0.4
0.3
0.2
0.1
0.0
a)
0
10
20
30
40
50
Reaction time / h
35
100
24
b)
2
5
10
1
0
12
80
60
40
20
0
6
12
25
8
4
28
13
16
26
8
6
3
46
18
36
28
12
170
200
240
Reaction temperature / °C
Figure 7. Effect of temperature a) on temporal evolution of erythritol concentration b) on
selectivitiy to the reaction products at 80%conversion. () 240°C; () 200°C; ( ) 150°C; (◼)
Cycles; (◼) C3; (◼) Threitol; (◼) BuOH; (◼) BDO; (◼) BTO.
-1
Reaction conditions: ERY 0.4 mol L , 120 ml H
2
O, 80 bar of H
2
, 0.3g Rh-ReO
x
/ZrO catalyst,
2
C-O cleavage: BuOH+BDO+BTO.
1
7
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
As expected, the higher the temperature, the more rapid the reaction. At 240°C erythritol was
sharply converted to attain full conversion after 5 h, while at 200°C erythritol conversion
reached 92% after 10 h and full conversion after 24 h. Finally, at 150°C erythritol conversion
was 27% after 9 h and 80% after 48 h.
Upon decreasing the reaction temperature from 240 to 150°C, the selectivity to linear
deoxygenated products increased from 33 to 79% (Fig. 7b). The results show that the selectivity
towards C-O bond cleavage was enhanced with decreasing temperature. The same trend was
observed in the dehydroxylation of sorbitol over CuO-ZnO catalyst[41] in which the selectivity
to linear deoxygenated products increased from 35 to 63% as temperature decreased from 240
to 180°C. In parallel, the selectivity to C2-C3 compounds, produced from the C-C cleavage in
the middle-carbon chain through decarbonylation or retro-aldol condensation reactions,
increased from 5 to 16% with an increase in temperature from 150°C to 240°C. This tendency
agrees with that found for transformation of sorbitol into liquid alkanes over a bifunctional
catalytic system Pt/ZrO
2
+ TiO
2
-WO
,[42] in which the number of C-C cleavages increased as
x
temperature increased from 200°C to 240°C. In addition, the highest selectivity towards cycles
was observed at 240°C (selectivity = 24%), suggesting that dehydration reactions are favored
at high temperature; the heterocyclic compounds are thermodynamically-stable. A similar
progression was noted by Sato et al,[43] for the dehydration of 1,4-butanediol over conventional
acid catalysts such as zirconia, alumina, and silica-alumina. They observed that the conversion
increased along with the temperature and the dehydration compound (tetrahydrofuran) was the
main product.
The % loss of TOC measured at the end of the reaction over Rh-ReO
x
/ZrO
2
increased from 3 to
1
3% as temperature increased from 150 to 240°C, due to the formation of gaseous products. It
is worth noting that CO
2
has been reported as the main product during the aqueous-phase
/C (245°C, 62 bar of H
);[44] the formation of CO
was due to the decarbonylation of the intermediate aldehyde over the metallic surface, followed
by water-gas shift reaction. In addition to CO , C1-C4 alkanes in gas phase had also been
hydrodeoxygenation of sorbitol over Pt-ReO
x
2
2
2
detected. In the present erythritol hydrogenolysis reaction, the selectivity to C3 products such
as glycerol (formed by decarbonylation reaction) increased from 5 to 12% as temperature
increased from 150 to 240°C, which can explain the increase in % loss of TOC measured. The
temperature was chosen at 200°C for further experiments to minimize the possible formation
of gaseous products and keep a reasonable activity.
2
.2.4. Effect of support (TiO
2
or ZrO
2
) and H pressure during erythritol hydrogenolysis.
2
1
8
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
The effect of H
2
pressure (in the range 30-120 bar) on initial rate (Figure S6) and product
selectivity (Figure 8) was investigated for erythritol hydrogenolysis at 200°C, over Rh-
ReO /ZrO and Rh-ReO /TiO
x
2
x
2
.
At low pressure (30 bar), the two catalysts showed similar initial reaction rates of 220±10
mmolERY g
Rh h-1. However, as pressure increased from 30 to 120 bar, the initial rate over Rh-
-1
Rh h-1, whereas it increased nearly linearly over
Rh h-1. These results show a nearly first order kinetics in
-supported catalyst in the range of pressure considered; which
-1
ReO
x
/TiO
2
increased slowly to 461 mmolERY
g
-1
Rh-ReO
x
/ZrO up to 1583 mmolERY g
2
H
2
pressure over the ZrO
2
indicates that the hydrogen is involved in the rate-determining step and suggesting that one
active hydrogen species is produced from one hydrogen molecule.[17,45] In contrast, the order
was positive but close to zero over the TiO -supported one. This suggests that the attack of
2
hydrogen active species is not the rate-determining step, but the C-O bond cleavage in the
substrate can be the rate-determining reaction.[17] These observations in initial rates over two
catalysts explain the large difference in the time to reach ca. 80% conversion between the two
catalysts under 80 bar H (8 h vs. 25 h, Table 3). Differences in dispersion and in interactions
2
between rhodium, rhenium and the support suggested by the characterizations of the solids (see
above) may explain the different behaviors on the two supports.
The carbon selectivity to the different categories of products and the TOC values measured at
8
0 % conversion are presented in Figure 8 for the two supported-catalysts under different
pressures. As examples, Figure S7 illustrates the evolution of selectivity as a function of
erythritol conversion at 200°C under 80 bar or 120 bar.
1
00
100
80
60
40
20
0
7
7
7
1
1
1
2
9
9
5
8
6
4
2
0
0
0
0
0
8
1
3
14
6
1
1
21
11
7
5
5
5
8
4
10
28
5
29
4
36
26
4
3
27
37
35
2
8
24
1
0
13
3
0
80
120
30
80
120
Hydrogen pressure /bar
Rh-ReO /TiO
Rh-ReO /ZrO
x
2
x 2
1
9
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
Figure 8. Dependence of product selectivities and %TOC measured at 80% conversion of
erythritol. (▲) % TOC measured; (◼) Cycles; (◼) C3; (◼) Threitol; (◼) BuOH; (◼) BDO;
(◼) BTO.
-1
Reaction conditions: 200°C, 0.3g catalyst, ERY 0.4 mol L , 120 ml H
2
O.
First, regarding the carbon balance, there was no difference between TOC measured and TOC
calculated. In the presence of both catalysts, the mass balance was > 90% except under 30 bar
of hydrogen. The loss of carbon in liquid phase is due to the formation of short-chain
compounds which have been transferred from the liquid phase to the gas phase, such as CO
alkanes.
2
or
At the temperature of 200°C, whatever the support (TiO
2
or ZrO ) and the pressure (PH2 in the
2
range 30-120 bar), the selectivity to butanols (in the range 4-6%) and to C3 compounds (in the
range 7-11%) remained roughly constant. Cyclic compounds (mainly 1,4-AE) resulting from
internal dehydration of erythritol in the presence of Rh-ReO
were favored under 30 bar of H (13% and 21%, respectively). These data are consistent with
observations during the hydrogenolysis of sorbitol to hexane over Pt/NbOPO at 250°C, in
x
/TiO
2
and Rh-ReO
x
/ZrO
2
catalysts
2
4
aqueous medium: it was shown that the oxygenated compounds resulting from dehydration of
sorbitol (1,4-sorbitan, isosorbide, and 2-(tetrahydrofuran-2-yl) ethan-1-ol) were the main
products under low hydrogen pressure.[46]
Moreover, the epimerization of erythritol to threitol was higher under 120 bar of H
ReO /TiO (selectivity 14%) and Rh-ReO /ZrO (selectivity of 9%), than under 30 bar
selectivity of 5%). It was also observed in the literature that the presence of ReO species
enhanced the selective 1,3-isomerization of allylic alcohols[47] and isomerisation of sorbitol to
mannitol over Pt-ReO
/C.[44] A mechanism similar to that described for epimerization of xylitol
2
over Rh-
x
2
x
2
(
x
x
to arabitol may be suggested via dehydrogenation-hydrogenation route.[48] It was proposed that
the reaction starts with reversible dehydrogenation of xylitol to xylose on the metal surface; the
isomerization of xylose to xylulose is catalyzed by base[49] or Lewis acid sites[50] and then
xylulose undergoes hydrogenation reaction over the metal surface to arabitol. So, as hydrogen
pressure increased, the dissolved hydrogen concentration in aqueous medium will increase and
result in more hydrogen molecules that access the metal active site; thus, the final
hydrogenation step of unsaturated intermediates will be favored. This behavior is also in
accordance with results for selective hydrogenolysis of xylitol to ethylene glycol and propylene
2
0
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
glycol over Cu/SiO
2
at 200°C, which showed that the selectivity to arabitol increased as
hydrogen pressure increased.[51]
Furthermore, the selectivity to BTO over Rh-ReO
x
/TiO
2
and Rh-ReO
x
/ZrO catalysts increased
2
from 10 and 13% to 35 and 37%, respectively, when the pressure increased from 30 to 120 bar.
This enhancement was accompanied with the decrease in selectivity to BDO (from 43 to 28%)
over Rh-ReO
x
/TiO
2
, while no drastic changes were noticed in the presence of Rh-ReO
x
/ZrO ,
2
i.e. the selectivity to BDO remained constant around 26-29%.
Owing to the multiple hydroxyl groups on the polyols, a number of different products are
obtained, depending on which hydroxyl groups are eliminated. A more detailed examination
of the selectivity to the different triols (1,2,3- and 1,2,4-BTO) and diols (1,2-, 2,3-, 1,3- and 1,4-
BDO) obtained during erythritol hydrogenolysis over both supported catalysts under varied
pressures was undertaken. The temporal concentration of the butanetriols and butanediols
observed over the Rh-ReO
representative example.
x
/TiO
2
catalyst under 80 bar H
2
are presented (Figure 9a) as a
0
0
0
0
0
.08
a)
.06
.04
.02
.00
0
5
10
15
20
25
Reaction time / h
1.8
1.5
1.2
0.9
0.6
0.3
0.0
b)
0
30
60
90
120
Hydrogen Pressure / bar
2
1
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
Figure 9. (a) Concentration of the different butanetriols and butanediols as a function of time.
) 1,2,3-BTO; (▲) 1,2,4-BTO; (◇) 1,2-BDO; (☆) 2,3-BDO; (⁕) 1,4-BDO. Reaction
(
-1
conditions: ERY 0.4 mol.L , 120 mL H
2
O, 200°C, 80 bar of H
2
, 0.3 g Rh-ReO
x
/TiO . (b) n1,2,3-
2
BTO/n1,2,4-BTO ratio as a function of pressure and support; (▲) Rh-ReO
x
/ZrO , () Rh-ReO
2
x
/TiO .
2
The triol and diol compounds were formed from the beginning of the reaction as primary
products; their concentrations kept increasing until the end of the reaction, i.e. 80% conversion
of erythritol. The molar ratio between the two butanetriols (n1,2,3-BTO/n1,2,4-BTO) are shown in
Figure 9(b). They were dependent on the catalyst (Rh-ReO
x
/TiO
2
, Rh-ReO
x
/ZrO
2
) and the
pressure. The ratio was close to 1 over Rh-ReO /TiO ; this suggests that there is no preferential
x
2
removing of the primary or secondary hydroxyl group from erythritol, which yield 1,2,3-BTO
and 1,2,4-BTO, respectively. However the removal of the primary hydroxyl is favored over Rh-
ReO
x
/ZrO at low pressure (ratio ~ 1.6).
2
In regards to the diols, both 1,2-BDO and 2,3-BDO were the main diols formed, while 1,4-BDO
was always formed in lower concentrations, and 1,3-BDO was only detected in traces or not at
all.
The selectivities over TiO
ones observed by the group of Tomishige for erythritol hydrogenolysis over various Ir- and Rh-
based catalysts at lower temperature.[10] They found that at 50% conversion, the Rh-ReO
/SiO
Re/Rh molar ratio 0.5) catalyst formed preferably butanols (selectivity 46%) than butanediols
selectivity 8%). In contrast, Ir-ReO /SiO (Re/Ir molar ratio 1) catalyst promoted the formation
2
- or ZrO
2
-supported Rh-ReO catalysts were then compared to the
x
x
2
(
(
x
2
of butanediols (at 74% conversion, the selectivity to BDO was 48%). However, these authors
observed that 1,4-BDO, and at a lesser extent 1,3-BDO, were by far the major diols formed. On
the contrary, when the present Rh-ReO catalysts were applied to erythritol at 200°C, 1,2-BDO
x
and 2,3-BDO, the diols with two adjacent hydroxyl groups, were obtained as the major diols.
The discrepancy between both studies must result from the nature of the metal (Rh or Ir), and
support (SiO
2
, TiO
2
, ZrO ), and the different temperature used (200°C in the present work
2
instead of 120°C). This point will be discussed later in this paper.
Lastly, 1-BuOH and 2-BuOH were always formed with low selectivity; the trend is towards the
formation of a slightly higher concentration of 1-BuOH.
2
.2.5. Hydrogenolysis of butanetriols over Rh-ReO
x
/TiO
2
and Rh-ReO
x
/ZrO at different
2
pressures
2
2
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
In order to understand the trends of selectivity to BTO, BDO and BuOH as a function of
pressure over the two catalysts during hydrogenolysis of erythritol, some catalytic tests were
performed starting from butanetriols (1,2,3- or 1,2,4-BTO) at the initial concentration of 0.12
-1
mol L . The initial reaction rates are summarized in Table 4.
Table 4. Initial reaction rates as a function of catalyst and pressure for hydrogenolysis of ERY,
1
,2,3-BTO, 1,2,4-BTO, and 1,4-AE.
-1
Reaction conditions: 200°C, BTO 0.122 mol L , 120 mL H
2
O, 0.075g catalyst
-1
-1
Pressure H
2
V [ mmolSubstrate gRh h ]
0
Catalyst
[
bar]
ERY 1,2,3-BTO 1,2,4-BTO 1,4-AE
3
0
210
281
461
232
888
1583
78
87
-
263
247
240
281
738
-
-
-
Rh-ReO
x
/TiO
2
80
20
1
-
3
0
97
398
-
212
845
-
Rh-ReO
x
/ZrO
2
80
1
20
-1
Over Rh-ReO
x
/TiO2, the initial reaction rate of conversion of 1,2,3-BTO (78-87 mmolBTO
g
Rh
-1
-1
Rh h-1) was not significantly modified by the H
h ) and 1,2,4-BTO (247-263 mmolBTO
g
2
pressure applied (30 to 80 or 120 bar). On the contrary, in the presence of Rh-ReO
x
/ZrO
2
, the
initial rate for the conversion of 1,2,3-BTO and 1,2,4-BTO increased from 97 and 281 mmolBTO
Rh h-1, respectively, to 398 and 738 mmolBTO
from 30 to 80 bar. As for erythritol hydrogenolysis, almost a zero order reaction with respect to
pressure was observed for BTOs hydrogenolysis over the TiO supported catalyst, whereas
a first-order kinetics in H pressure was noted for the ZrO supported catalyst (Figure S6). These
results suggest different mechanisms of activation of H on both promoted catalysts. In
-
1
g
Rh h-1, respectively, as H
pressure was varied
-1
g
2
H
2
2
2
2
2
comparison, Chia et al[28] and Tomishige et al[31,52] reported a first-order reaction for cyclic
ethers and glycerol hydrogenolysis for Rh-Re supported on carbon and Rh-Re and Ir-Re
supported on silica, respectively.
One may also note that the reaction of 1,2,4-BTO hydrogenolysis proceeded twice or three
times more rapidly than the reaction of 1,2,3-BTO in the same conditions. This difference in
reactivity was also observed by Amada et al over Ir-ReOx/SiO
2
.[10] The reactivity of 1,2,4-BTO
was approximately equal to that of erythritol, and 1,2,3-BTO showed much lower activity. We
2
3
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
propose that the cause of these differences may rely on the ability to form cyclic ethers. Indeed,
,2,3-BTO cannot be dehydrated to form a cyclic ether, whereas 3-hydroxytetrahydrofuran
3OH-THF) and THF were produced from 1,2,4-BTO with a global selectivity of 11-12% at
1
(
5
0% conversion over Rh-ReO
x
/TiO
2
and 21-25% at 80% conversion over Rh-ReO
x
/ZrO
2
(Figure 10).
Figure S8 gives an example of hydrogenolysis of 1,2,3-BTO and 1,2,4-BTO under 80 bar of H
over the ZrO -supported catalyst. Figure 10 summarizes the influence of the support (TiO
ZrO ) and the pressure applied (30 or 80 bar) on the selectivity to the different products during
hydrogenolysis of both triols. The results are given at 80% conversion for Rh-ReO /ZrO and
at 50% conversion for Rh-ReO /TiO as the latter was less active. The selectivity to BDO for
the hydrogenation of a given triol in the presence of Rh-ReO /TiO was constant with H
2
2
2
or
2
x
2
x
2
x
2
2
pressure; however, the values were drastically different (92% from 1,2,3-BTO and 52-56%
from 1,2,4-BTO). Over-hydrogenolysis to butanol remained low (< 5%) and selectivity to C3
compounds slightly increased as pressure increased (8-11%). As for the results in the presence
of Rh-ReO
x
/ZrO , at 80% conversion, the selectivity to BDO was also not significantly affected
2
by the pressure; again, the selectivity to BDO was higher from 1,2,3-BTO (62-69%) than 1,2,4-
BTO (30-33%). The selectivity to BuOH (3-6%) remained low. The selectivity to C3
compounds was clearly higher from 1,2,4-BTO than 1,2,3-BTO and increased with the pressure
(Figure 10).
O
T
BDO
Cycles
-
B
O
T
,
4
-B
O
T
,
2
4
1
,
-B
O
T
r
,2
3
a
1
r
,
,2
1
-B
,3
,2
b
r
0
a
8
b
O
T
C3
0
a
1
3
b
0
r
-B
,4
,2
O
T
0
a
8
b
-B
,4
,2
1
O
T
BuOH
3
1
-B
,3
,2
1
r
r
O
T
a
b
r
0
a
-B
,3
,2
1
8
b
0
a
3
b
0
r
a
8
b
0
3
2
4
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
Figure 10. Dependence of product selectivities as a function of the pressure and the support in
hydrogenolysis of 1,2,3-BTO and 1,2,4-BTO. The results are given at 50% conversion over Rh-
x
ReO /TiO
2
, 80% conversion over Rh-ReO
x
/ZrO . (◼) BDO; (◼) Cycles; (◼) C3; (◼) BuOH.
2
-1
Reaction conditions: 200°C, BTO 0.122 mol L , 120 mL H
2
O, 0.075g catalyst.
2
.2.6. Selectivity to butanediols and butanols
The diols may be formed from erythritol through different routes: (i) erythritol may be first
hydrogenolysed to the two butanetriols and subsequently to the diols; hydrogenolysis of 1,2,3-
BTO will produce 2,3-BDO, 1,2-BDO, and 1,3-BDO, whereas hydrogenolysis of 1,2,4-BTO
will produce 1,2-BDO, 1,3-BDO, and 1,4-BDO (Scheme 2), (ii) they may also be formed via
the ring opening hydrogenolysis of 1,4-AE via 3OH-THF; however, only 1,2- and 2,3-BDO
will be produced by this route, (iii) elimination of two hydroxyl groups directly from erythritol;
as shown in Fig. 6, BDO are formed from the start of the reaction which suggests the feasibility
of this last route ; moreover this mechanism agrees well with the proposition of Arai et al on
the possibility of the direct hydrogenolysis of 1,4-AE to 2-BuOH without the desorption of
intermediates from the active site over a Rh-MoO
catalyst.[17] As an attempt to identify the
x
preferential mechanism (i, ii, iii), we looked at the selectivity to the different butanediols and
butanols during the hydrogenolysis of the butanetriols and 1,4-anhydroerythritol.
2
5
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
2
2
2
2
2
2
2
C-O bond hydrogenolysis
Scheme 2. Reaction pathways associated with the formation of BDO and BuOH from BTO.
2
.2.7. Hydrogenolysis of butanetriols
As an example, Table 5 includes the detailed selectivity towards butanediols and butanols
obtained during the hydrogenolysis of butanetriols over Rh-ReO
x
/TiO
2
and Rh-ReO
x
/ZrO
2
under 80 bar. The same trends were observed at lower pressure. The difference in selectivity
towards one or another BDO can be linked to the existence of preferential C-O bond cleavage.
(
Scheme 2). It is usually recognized[53] that in acid catalyzed dehydration, the selectivity is
driven by the stability of ions formed in the reaction, leading to dehydration of substituted C
bearing a secondary alcohol first. In contrast, selectivity for hydrogenolysis using a
heterogeneous metallic catalyst is often driven by steric hindrance, leading to cleavage of the
primary alcohol first.
Table 5. Carbon selectivity to BDO or BuOH at 50% conversion over Rh-ReO
conversion over Rh-ReO /ZrO during hydrogenolysis of BTO.
x
/TiO and 80%
2
x
2
2
6
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
-1
Reaction conditions: 200°C, 80 bar of H
2
, BTO 0.122 mol L , 120 mL H
2
O, nBTO/nRh= 530,
0
.075g catalyst.
Carbon selectivity%
Susbtrate 2,3-BDO 1,2-BDO 1,4-BDO 1,3-BDO 1-BuOH 2-BuOH
Catalyst
1
1
1
1
,2,3-BTO
,2,4-BTO
,2,3-BTO
,2,4-BTO
72
n.a
48
20
34
14
23
n.a
15
n.a
4
0
3
1
2
2
4
3
2
4
2
Rh-ReO
Rh-ReO
x
/TiO
2
0
x
/ZrO
2
n.a
< 1
n.a.: not applicable
Starting from 1,2,3-BTO as the reactant, whether the Rh-ReO
x
nanoparticles were supported on
TiO or ZrO , the selectivity towards 2,3-BDO (72%; 48%) was clearly much higher than that
2
2
to 1,2-BDO (20%; 14%) and 1,3-BDO (not detected), indicating that the reactivity of the
terminal primary -OH group in position 1 in 1,2,3-BTO is higher than that of the secondary –
OH groups in positions 2 and 3 (Scheme 2). A ratio 2,3-BDO/1,2-BDO of c.a. 3.5 was
calculated for the two catalysts. In addition, 1,3-BDO was not observed, meaning that removal
of the most hindered secondary hydroxyl group in position 2 did not occur.
On the other hand, starting from 1,2,4-BTO as the reactant, 1,2-BDO was the major product
with 34% and 23% selectivity over Rh-ReO
x
/TiO
2
and Rh-ReO
x
/ZrO , respectively. Lower
2
amounts of 1,4-BDO (15%; 4%) and negligible amounts of 1,3-BDO (3%; < 1%) were detected.
So the reactivity of the terminal primary –OH group at position 4, the one with the less steric
hindrance, is higher than that of the two other OH-groups (positions 1 and 2). Moreover, the
reactivity of the secondary –OH group (position 2) is higher than the one of the adjacent primary
OH-group (position 1); this latter reactivity trend is different from that observed from 1,2,3-
BTO, where the elimination of the primary OH group was favored over the cleavage of either
of the two secondary OH-groups. The ratio 1,2-BDO/1,4-BDO was ca. 2.3 and 5.7 over TiO
2
and ZrO , respectively.
2
It should also be noted that unlike from 1,2,3-BTO, cyclic compounds (3OH-THF and THF)
were formed from 1,2,4-BTO. The amount of C3 products is higher from hydrogenolysis of
1
,2,4-BTO than from 1,2,3-BTO.
2
7
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
These results are different from those analysed by Amada et al over an Ir-ReO
x
/SiO
2
catalyst at
1
00°C.[10] These authors observed the preferential formation of 1,3-BDO from 1,2,3-BTO (47
%
selectivity at 1% conversion only), whereas the main product from 1,2,4-BTO was 1,4-BDO
(
selectivity 73% at 22% conversion), formed by dissociation of the C-O bond neighboring the
CH OH group. 1,4-BDO was also obtained with the highest selectivity (63%) compared to 1,3-
BDO and 1,2-BDO (14% and 9%, respectively) at 13% conversion in 1,2,4-BTO
hydrogenolysis by Chia et al, over a Rh-ReO /C catalyst at 100°C under 34 bar H
.[28] These
2
x
2
studies highlighted the high reactivity of a secondary hydroxyl group adjacent to a primary OH-
group. These results also demonstrate the crucial role of the support.
Finally, the selectivity at 80% conversion towards the butanols formed by elimination of two
hydroxyl groups over both catalysts was examined. In the hydrogenolysis of 1,2,3-BTO, the
selectivity to 2-BuOH (3% and 4%, respectively) was slightly higher than that to 1-BuOH (1%
and 2%; Table 5). This is due to the fact that 2-BuOH can be formed from the major 2,3-BDO
via cleavage of any of the secondary –OH group (positions 2 and 3) and from cleavage of the
primary –OH group of 1,2-BDO (position 1, Scheme 2), while 1-BuOH can be obtained solely
through cleavage of secondary –OH group of 1,2-BDO (the less favored BDO from 1,2,3-
BTO). However the observation of a non-negligible amount of 1-BuOH suggests a high
reactivity of the secondary OH in comparison with the primary OH in 1,2-BDO. In contrast, in
the hydrogenolysis of 1,2,4-BTO, the selectivity to 1-BuOH (2%; 4%) was higher than that to
2
-BuOH (2%; 2%). 1-BuOH was obtained from cleavage of secondary –OH group in the more
available 1,2-BDO or from cleavage of one or the other primary –OH group in 1,4-BDO,
whereas 2-BuOH was obtained only from 1,2-BDO through elimination of the primary –OH
groups. As mentioned above, C-O cleavage from 1,2-BDO occurs preferentially on the
secondary OH-group yielding 1-BuOH.
2
.2.8. Hydrogenolysis of 1,4-anhydroerythritol
Furthermore, some reactions were performed starting from 1,4-anhydroerythritol. The effect of
pressure (30 and 80 bar) on the evolution of product selectivity is presented over Rh-
ReO /ZrO at 200°C (Figure S9). The initial conversion rates of 1,4-AE were 213 and 845
H
2
x
2
Rh h-1 (Table 4), i.e. a positive order of ca. 1 with respect with H
pressure over
-
1
mmol1,4-AE
g
2
this catalyst.
First, the ring opening hydrogenolysis of 1,4-AE via 3OH-THF yielded 1,2,3-BTO, which was
further converted to BDO and butanol. Under 30 bar, the selectivity to 1,2,3-BTO decreased
2
8
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700260
from 25 to 7% as conversion increased from 20 to 65%, while at the same time the selectivity
to BDO increased from 20 to 26%; differently, under 80 bar, the selectivity to 1,2,3-BTO
decreased from 21 to 4% as conversion increased from 32 to 82%, and the selectivity to BDO
increased from 16 to 26%. The butanediols produced were 2,3-BDO and 1,2-BDO and the 2,3-
BDO/1,2-BDO ratio was estimated to be approximately 2 all along the reaction, i.e. a
preferential C-O cleavage of the primary CH
The same trends were observed, when starting from 1,2,3-BTO (Table 5). Differently, over a
Rh-ReO /SiO (Re/Rh = 0.13) operating at 120°C under 80 bar H , 1,2-BDO, 2,3-BDO, and
,3-BDO were obtained with selectivity of 13%, 6%, and 3%, respectively, at 29% conversion
of 1,4-AE.[17] Ir-ReO yielded even higher selectivity to 1,3-BDO.[10,17] These authors
/SiO
2
OH group. Notably, 1,3-BDO was not detected.
x
2
2
1
x
2
suggested that under the reaction conditions they used, 1,3-BDO was formed both by
hydrogenolysis of 1,2,3-BTO by successive C-OH hydrogenolysis of 1,4-AE via the
intermediate 3OH-THF. In contrast, they observed a relatively higher selectivity to 2,3-BDO
over a Rh-MoO
x
/SiO catalyst. The selectivity to the butanediols is thus very dependent on the
2
reaction conditions and the catalyst used.
In addition, 1,4-AE underwent successive C-O bond hydrogenolysis to 3OH-THF and THF.
However, the selectivity to cycles remained low and constant (8-9% under 30 bar and slightly
higher at 14% under 80 bar) as a function of conversion. This suggests that ring opening
hydrogenolysis of 1,4-AE to 1,2,3-BTO is leading over C-O hydrogenolysis to 3OH-THF. An
interesting point is the absence of formation of 1,3-BDO during 1,4-AE hydrogenolysis, which
suggests that no hydrogenolysis of 3OH-THF to 1,3-BDO took place during the reaction. 3OH-
THF has already been found to be poorly reactive.[54]
The C-C cleavage of 1,4-AE yielded C3 products (glycerol which is then converted to 1,2-
propanediol, 1-PrOH, and 2-PrOH). The selectivity to C3 remained constant (32% under 30
bar, slightly lower at ca. 24% under 80 bar) as a function of conversion. Finally, there was no
effect of H
2
pressure on the selectivity to butanol which remained constant and low (0-4%) as
a function of conversion.
2
.2.9. Comparison of evolution of selectivity to BDO during erythritol, and BTO or 1,4-AE
hydrogenolysis
The results from the hydrogenolysis of the butanetriols (Fig. S8, Table 5) and of 1,4-
anhydroerythritol (Fig. S9) were then compared with the evolution of the selectivity to the
2
9
This article is protected by copyright. All rights reserved.