Fuel 357 (2024) 129984

Contents lists available at ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Full Length Article

Liquid-phase sorbitol hydrogenolysis over Co catalysts: Effect of external vs

endogenous H2 on the product distribution
A.J. Reynoso a, c, J.L. Ayastuy a, *, U. Iriarte-Velasco b, L. Vivier c, C. Especel c,
M.A. Gutiérrez–Ortiz a
a
Department of Chemical Engineering, Faculty of Science and Technology, University of the Basque Country UPV/EHU, Sarriena S/N, 48940 Leioa, Spain
b
Department of Chemical Engineering, Faculty of Pharmacy, University of the Basque Country UPV/EHU, Paseo de la Universidad, 7, 01006 Vitoria, Spain
c
Université de Poitiers, CNRS UMR 7285 IC2MP, Institut de Chimie des Milieux et des Matériaux de Poitiers, 4 rue Michel Brunet, TSA 51106, 86073 Poitiers Cedex 9,
France

A R T I C L E I N F O A B S T R A C T

Keywords: Aqueous sorbitol hydrogenolysis (6.25 wt% aq. solution) in batch reactor was studied over bare and Pt-modified
Sorbitol cobalt aluminate-based catalysts at distinct reaction atmospheres (N2 and H2 pressurized) focusing on the cor­
Hydrogenolysis relation between the physico-chemical properties of catalysts and the product distribution. It was found that over
HDO
trimetallic xPt/CoAl catalysts (x = 0.3 and 1 wt%) operating at 240 ◦ C and hydrogen pressure (60 bar) resulted in
APR
Cobalt
a yield of 27–35 molC % C5-6 hydrocarbons. Under inert atmosphere, the highest yield (57–70 molC %) was to
monofunctional products with only one oxygen functionality, for all catalysts. A preferential retro-aldol
condensation route is manifest. Post-reaction analyses showed less structural and chemical modification in the
Pt-containing catalysts and a slight positive difference in those used under hydrogen pressure over those used
under inert atmosphere.

1. Introduction APR comprises two cascade reactions: firstly, the oxygenate mole­
cule is dehydrogenated through C–H, C–C and O–H bond cleavages to
The quest for sustainable and environmentally benign sources of form CO and H2, and subsequently Water-Gas Shift (WGS) converts CO
energy has become urgent. Practically, all the chemicals manufactured into CO2 and H2. HDO is a hydrogenolytic reaction that involves the
from fossil fuels can be produced from biomass feedstocks in the so- removal of an oxygen atom (deoxygenation) with the simultaneous or
called bio-refineries [1]. Biomass derived from non-edible sources is subsequent addition of H2 (hydrogenation). It was found that the
considered as the only sustainable feedstock to produce fuels and removal of oxygen-containing functionalities can occur through direct
chemicals. Sorbitol, a sugar alcohol derived from lignocellulosic waste, hydrogenolysis (C–O bond cleaved with H2), dehydration (C–O bond
is one of the 12 bio-based building blocks for production of fuels and cleaved through removal of water), decarbonylation (removal of CO),
chemicals [2], which is currently produced by hydrogenation of glucose and decarboxylation (removal of CO2) [10].
or directly from cellulose [3]. There exists a growing interest in the The catalyst formulation and structure play a critical role in the
research of cost-effective catalytic conversion processes for sorbitol [4]. resulting compounds. Most of the studies have focused on catalytic
Liquid-phase processing of sorbitol can be tailored to selectively systems based on noble metals such as Re, Pd, Rh, Ir, and especially, Ru
produce gasoline pool [5]. Actually, aqueous-phase catalytic processes and Pt. Platinum supported on different acidic materials and bimetallic
are promising and reliable methods for the production of advanced Pt/metal-oxide catalysts have been extensively investigated for the
sustainable biofuels, fine chemicals and pharmaceuticals from biomass- catalytic conversion of sorbitol, and high yields to C5/C6 alkanes were
derived oxygenates [6,7]. These processes can be used to obtain H2 or reported [11–13]. Bifunctional Ru-Mo catalysts yielded, mainly, poly-
value added chemicals either taking advantage of the in-situ produced and mono-oxygenated liquids and hydrocarbons in gaseous phase
H2 (Aqueous-Phase Reforming, APR) or by supplying H2 from an [14,15]. Among the non-precious metal catalysts, Ni-based catalysts
external source (Aqueous-phase Hydrodeoxygenation, HDO) [8,9]. have been the most widely used in the hydrothermal processing of

* Corresponding author.
E-mail address: joseluis.ayastuy@ehu.eus (J.L. Ayastuy).

https://doi.org/10.1016/j.fuel.2023.129984
Received 25 May 2023; Received in revised form 8 September 2023; Accepted 27 September 2023
Available online 1 October 2023
0016-2361/© 2023 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
A.J. Reynoso et al. Fuel 357 (2024) 129984

sorbitol [16,17]. It is more scarce the research on Cu- and Co-catalytic with Varian GC-450 and GC-430 system chromatography, while liquid
systems [18,19]. Eagan et al. reported the effectiveness of Co-based samples were manually collected in vials and analysed offline by HPLC
catalysts for sorbitol HDO, and concluded that C5+ mono-alcohols and (BIO-RAD Aminex HPX 87H column). The total quantity of carbon in the
heterocycles were majority, while ketones and aldehydes were readily aqueous phase was measured by TOC (Shimadzu Carbon-L Series). The
hydrogenated. Due to the leaching and sintering, the Co catalyst was quantification of both gaseous and liquid products was carried out by
irreversibly deactivated [20]. external calibration.
The novelty of the current study is the investigation of both the effect Sorbitol conversion (XSorb), carbon conversion to gas (Xgas) and
of the composition of cobalt-based catalysts and the source of hydrogen, product yield (Yi) are defined on a carbon basis, as follows:
on the product distribution and yield to value-added chemicals by the
nt=0 t=t
Sorb − nSorb
aqueous-phase hydrogenolysis of sorbitol. For this purpose, bare Co3O4, XSorb (%) = 100 × t=0
(1)
nSorb
cobalt aluminate with Co/Al = 0.25 and 0.625 atomic ratios, and Pt-
impregnated cobalt aluminate (0.3 and 1 wt% Pt) were synthesized. In
nt=0 t=t
Sorb and nSorb are the moles of sorbitol initially charged to the reactor
previous works, the suitability of these catalysts for glycerol APR was
and those remaining at time t, respectively.
studied [21,22]. To investigate the effect of endogenous or exogenous
source or hydrogen, the reactor was pressurized in N2 (APR conditions) nt=0 t=t
C − nC,TOC
or H2 (HDO conditions), respectively. Thorough characterizations were Xgas (%) = 100 × t=0
(2)
nC
performed on the synthesized catalysts (either in fresh and reduced
forms) to determine their morphology, structure, and chemical proper­ nt=0 and nt=t
C C,TOC are the moles of carbon initially charged to the reactor
ties, which help us to correlate with the reaction performance. In and moles of carbon measured in liquid phase as TOC at time t,
addition, spent catalysts were characterized to unveil the changed un­ respectively.
dergone during sorbitol transformation. The novelty of the present work
lies on the comparative approach to study the production of both ntC,i
Yi (%) = 100 × (3)
gaseous and liquid hydrocarbons under different atmospheres: APR nt=0
C
(inert atmosphere) and HDO (hydrogen pressure) reaction conditions.
ntC,i is the moles of carbon in a specific product i. Hydrogen yield (YH2)
2. Experimental section was calculated as follows:
ntH2 1
2.1. Catalyst synthesis YH2 (%) = 100 × × (4)
nt=0
Sorb 13
Cobalt aluminate bimetallic catalysts (xCoAl, with x = 0.25 and
The factor 1/13 is the sorbitol/H2 reforming ratio (C6 H14 O6 +
0.625) were prepared by coprecipitation (x being the Co/Al molar ratio).
6H2 O→6CO2 + 13H2 ). The carbon balance fluctuated between 91 and
Likewise, Pt-modified CoAl trimetallic catalysts (yPt/CoAl, where y =
112 % for sorbitol HDO and 82–87 % for sorbitol APR. The difference
0.3, 1) were obtained by wet impregnation according to the procedure
may arise from the loss of volatile products during liquid sampling and/
described elsewhere [22]. As a reference, bare monometallic Co3O4 was
or formation of insoluble products of degradation [23].
also synthesized following the same procedure as xCoAl.
3. Results and discussion
2.2. Characterization techniques
3.1. Materials characterization
The bulk composition of the catalysts was measured by ICP-AES.
Textural properties of catalysts were determined by N2 physisorption.
The N2 adsorption/desorption isotherms (Fig. S1) presented type IV
The crystalline structure of the catalysts was characterized by XRD, and
isotherms characteristic of mesoporous solids, and H2 type hysteresis
the crystallite average size was calculated by Scherrer equation. Catalyst
loop at high relative pressures. For all catalysts, both the Co/Al ratio and
reducibility in the range 50–950 ◦ C was studied by H2-TPR. The catalysts
the Pt loading were close to the nominal values (Table 1). In the calcined
acidity was evaluated by ammonia chemisorption and NH3-TPD.
form, the specific surface area (SBET) of the bimetallic xCoAl solids
Carbonaceous material deposits on spent catalysts were quantified by
decreased with Co content. However, these values were notably higher
Temperature-Programmed Hydrogenation (TPH). Metallic surface area
than those measured for bare Co3O4. After Pt loading, SBET increased,
was estimated by H2 chemisorption at 40 ◦ C. The activity of the metal in
along with a notable increase (by a factor of 2) on pore volume and pore
the model reaction of cyclohexane dehydrogenation was evaluated in a
size (Table S2), indicating a significant modification of the textural
fixed bed reactor at atmospheric pressure, at 250 ◦ C, feeding cyclo­
properties. After reduction at 600 ◦ C for 1 h (heating rate 5 ◦ C/min), the
hexane/H2 mixture. The oxidation state of the elements and surface
SBET decreased due to the exsolution of metal to the surface [24]. The
composition was evaluated by XPS. Morphology of the particles was
decrease in SBET upon reduction was less pronounced as Co content
studied by STEM-HAADF. More details on analytical methods are given
increased in the xCoAl solids. This behaviour was further intensified
in ESI.
upon Pt loading. The pore size distribution (PSD) (Table S2, Fig. S1b)
clearly illustrates this behaviour where the PSD remained similar in the
2.3. Aqueous-phase transformation of sorbitol
Pt-containing solids whereas it shifted to larger pore size values in xCoAl
catalysts. Thus, it could be concluded that Pt loading notably enhanced
Aqueous-phase sorbitol transformation was performed in a 300 mL
the structural stability of the cobalt aluminate spinel in reductive
stainless steel reactor (Autoclave Engineers) at 240 ◦ C and 60 bar,
environments.
pressurized under H2 (for HDO) or N2 (for APR). The gas–liquid, liq­
XRD pattern of calcined xCoAl and xPt/CoAl catalysts (Fig. 1)
uid–solid and internal mass transfer resistances were insignificant in
matched with both cobalt aluminate (PDF 00-044-0160) and cobalt
both HDO and APR conditions (Table S1), what allowed kinetic studies.
oxide (PDF 00-042-1467) spinels (being difficult to distinguish between
The reactor was loaded with 1 g of ex-situ reduced catalyst (reduc­
them). The diffractogram of the xPt/CoAl samples showed no peak due
tion temperature 600 ◦ C, 1 h duration) and an aqueous solution of
to Pt and/or PtOx phases, suggesting high Pt dispersion. The crystallite
sorbitol (10 g of sorbitol in 150 mL of H2O). After sealing, the autoclave
size of the spinel phase (Table 1) increased with Co/Al; and for the same
was pressurized with H2 (for HDO) or N2 (for APR) to the required
Co/Al ratio, it increased with the incorporation of Pt. The lattice
pressure. At specific intervals, gaseous products were online analysed

2
A.J. Reynoso et al. Fuel 357 (2024) 129984

Table 1
Chemical, textural and structural properties of the catalysts.
Catalyst Co/Al Pt SbBET (m2/g) dcspinel Lattice ddCo Total H2 Accessible metal atomsf AgS TOFg Surface acid
(at./ loading (nm) parameterc (nm) Uptakee (atMetal/g) (mol/ (s− 1) sites densityh
at.)a (wt. %)a (nm) (mmolH2/g) h⋅gMe) (μmolNH3/m2)
Calc. Red. Co Pt
18 18
(×10− ) (×10− )

Co3O4 n.a. n.a. 37.7 7.4 19.7 0.8059 ± 28.3 17.4 1.22 n.a. 0.74 1.61 0.06
0.0009
0.25CoAl 0.267 n.a. 136.2 96.6 3.8 0.7993 ± n.d. 3.81 2.28 n.a. 1.02 0.34 0.78
0.0058
0.625CoAl 0.634 n.a. 125.3 101.7 5.0 0.8042 ± 16.5 7.04 23.0 n.a. 1.78 0.59 0.51
0.0014
0.3Pt/ 0.633 0.29 146.3 131.0 6.3 0.8052 ± 6.9 7.29 30.4 5.2 8.35 2.29 0.56
CoAl 0.0011
1Pt/CoAl 0.623 1.09 138.6 126.1 6.3 0.8063 ± 6.5 7.36 25.5 20.7 7.92 2.20 0.68
0.0006

n.a.: not analysed; n.d.: not detected; a from ICP-AES; b from nitrogen isotherms; c from XRD of the calcined samples; d from XRD of the reduced samples; e H2-TPR; f H2
pulse chemisorption; g Specific activity (As) and turnover frequency (TOF) for cyclohexane dehydrogenation model reaction; h Temperature-programmed desorption of
ammonia (NH3-TPD).

Fig. 1. XRD patterns for the (a) calcined and (b) reduced samples.

parameter followed a similar trend to the crystallite size and could be Co3+ to Co2+, and the second broad peak (at about 425 ◦ C) were
explained by the difference in the ionic radii between Al3+, Co2+ and attributed to the reduction of Co2+ species to metallic cobalt [21]. For
Pt2+. Upon reduction, bare Co3O4 showed exclusively peaks from x5CoAl catalysts, the reduction profile could be decomposed into four
metallic Co in both hexagonal closed packed hcp (PDF 00-005-0727) and H2 consumption events. The first consumption at around 300 ◦ C was
face-centered cubic fcc (PDF 00-015-0806) phases, suggesting complete ascribed to the reduction of surface Co3+ with weak or without inter­
reduction. Reduced forms of the xCoAl solids still exhibited peaks from action with the alumina or cobalt aluminate phase. The peak at 400 ◦ C
spinel phase, indicating the irreducibility (at 600 ◦ C) of the Co species corresponded to the reduction of surface Co3+ cations in close interac­
linked to alumina. Also, diffraction peaks characteristic of hcp and fcc tion with the surrounding species. The largest peak at around 580 ◦ C was
metallic cobalt could be identified. Catalyst 0.25CoAl showed peaks assigned to Co2+ → Co0 reduction, and the latest peak (c.a. 780 ◦ C) was
attributable to γ-alumina (PDF 01-079-1558) and gibbsite (PDF 00-033- attributed to the reduction of cobalt ions as stoichiometric cobalt
0018). Both series of xCoAl and xPt/CoAl catalysts exhibited low in­ aluminate. xPt/CoAl catalysts, in addition to the peaks already described
tensity peaks from both forms of metallic cobalt revealing that Co spe­ for xCoAl catalysts, showed a low temperature peak (<200 ◦ C) that
cies were better dispersed as compared to bare Co3O4 catalyst. could be ascribed to the simultaneous reduction of PtOx and the above
The reduction profile of bare Co3O4 showed two well-defined peaks mentioned Co3+ species (to Co2+) with weak interaction with their co­
(Fig. 2). The first peak at 300 ◦ C corresponded to the reduction of surface ordination environment. The reduction of the later species would be

3
A.J. Reynoso et al. Fuel 357 (2024) 129984

with a mean size of 1.62 nm for 0.3Pt/CoAl, and 1.88 nm for 1Pt/CoAl,
which would explain the absence of diffraction peaks in XRD.
Accessible metallic atoms of the catalysts were characterized by
means of both H2 chemisorption and by activity in cyclohexane dehy­
drogenation model reaction. The accessible atoms exponentially
increased with cobalt content of binary oxides (Table 1). A ten-fold in­
crease was observed from 0.25CoAl to 0.625CoAl samples (2.28 ×
1018–23.0 × 1018 Co atoms per gram, respectively). Contrarily, bare
Co3O4 showed the least accessible cobalt atoms, in line with its larger
metallic Co nanoparticles (from XRD). Upon reduction of the spinel-type
solids, Co exsolution from the spinel matrix generates metallic cobalt on
the surface, which can agglomerate at high temperature. Interestingly,
Pt-containing catalysts contained a larger amount of accessible cobalt
atoms as compared to their counterpart (0.625CoAl). As expected, the
available metallic Pt sites increased with Pt loading.
The specific and intrinsic activity obtained in the dehydrogenation of
cyclohexane to benzene (structure-insensitive), are shown in Table 1.
The AS values followed the general trend of increasing with the
measured amount of accessible metal atoms. The specific activity of bare
cobalt was low, consistent with other results for cobalt-based catalysts
[26]. The aluminate-based catalyst 0.625CoAl, with a ten-fold increased
metallic atoms, showed slightly larger AS value. After Pt doping, the
specific activity was markedly increased. The intrinsic activity (TOF)
decreased in the xCoAl assays as compared to Co3O4 catalyst. The latter
fell out of the general trend (Fig. S2). Note that the cobalt aluminate
spinel catalysts showed a lower than expected intrinsic activity, how­
ever, after Pt doping, it was significantly enhanced (TOF increased from
0.59 s− 1 for 0.625CoAl to 2.20–2.29 s− 1 for Pt-containing catalysts). The
low activity of cobalt aluminate-derived catalysts could be due to the
presence of Coδ+ species that hinder the adequate adsorption of cyclo­
Fig. 2. H2-TPR profiles for the fresh calcined solids. hexane [27]. The obtained results confirmed the synergistic effect of Pt-
Co alloying, promoting the reducibility and increasing the amount of
promoted by hydrogen spillover from the Pt0 surfaces [25], as deduced available metallic Co species. Moreover, these trimetallic ensembles are
from the associated hydrogen consumption (Table 1). more active for the cyclohexane dehydrogenation than those on the
The metal distributions were resolved by EDX mapping. Bare Co3O4 xCoAl catalysts. The modification of the electronic properties of the Co
sample (Fig. 3a and 3f) contained stacked nearly spherical particles that particles by Pt doping was confirmed by XPS measurements, as dis­
were difficult to differentiate. xCoAl catalysts showed cuboid shape cussed below.
partially dispersed cobalt nanoparticles of around 20 nm, in line with The surface acid-sites density after reduction at 600 ◦ C is presented
XRD data. Homogeneous well-dispersed Pt metallic particles were found in Table 1. Bulk Co3O4 presented a remarkably low acidity (0.06
in both Pt-containing assays. These catalysts presented near-spherical Pt µmolNH3/m2) in spite of its very low surface area. Among bimetallic
nanoparticles supported over large Co and Al nanoparticles (Fig. 3d–e). solids, acidity decreased as Co/Al molar ratio increased (e.g. 0.78 and
Overall, the size of Pt nanoparticles ranged between 0.5 and 3.5 nm, 0.51 µmolNH3/m2 for 0.25CoAl and 0.625CoAl, respectively) due to the

Fig. 3. EDX maps of catalysts and their corresponding HAADF-STEM images below: (a,f) Co3O4, (b,g) 0.25CoAl, (c,h) 0.625CoAl, (d,i) 0.3Pt/CoAl, and (e,j)
1Pt/CoAl.

4
A.J. Reynoso et al. Fuel 357 (2024) 129984

formation, after reduction, of gibbsite [28]. Doping with Pt caused a than the bulk ratio, due to the lower surface free energy of Al than Co
modest increase of acid site density (0.56 vs 0.68 µmolNH3/m2 for 0.3 [34]. It was not possible to calculate the Pt composition at the surface
and 1Pt/CoAl, respectively). due to the low signal intensity. Noteworthy, Pt-doped catalysts exhibited
Fig. 4 shows the high-resolution Co 2p spectrum with the 2p3/2 and higher surface Co/Al atomic ratio compared to 0.625 CoAl sample.
2p1/2 splitting while the binding energies (BE) and the surface Co/Al
atomic ratios are summarized in Table S3. The Co 2p3/2 and Co 2p1/2 3.2. Sorbitol transformation in aqueous phase
could be deconvoluted in the corresponding Coδ+ (indistinguishable
between Co2+ and Co3+) [29], Co0 and shake up satellite peaks. Aqueous-phase sorbitol transformation via HDO or APR is a complex
For Co3O4 catalyst, the Co 2p3/2 peak was deconvoluted into BE process with simultaneous reactions (Scheme 1). This reaction scheme
values of 777.7 and 779.9 eV, corresponding to Co0 and Coδ+, respec­ was suggested on the basis of the obtained products, either in the gas and
tively, and Co 2p1/2 peaks at 792.7 and 794.8 eV, attributed to Co0 and liquid phase. Sorbitol can be dehydrated (via acid catalysis) to form
Coδ+, respectively, along with weak satellite peaks at 782.4 and 798.2 sorbitan (by cyclization), isosorbide and other oxygenated C5+ com­
eV [30]. Compared with bare Co3O4, the binding energy of Co 2p for pounds. These intermediates can then undergo successive dehydration/
cobalt aluminates blueshifted, indicating a more electronegative coor­ hydrogenation reactions to produce C5+ monofunctional compounds
dination environment. The BE of Co 2p3/2 further increased after Pt and subsequently, heavy gases (C5+). Retro-aldol condensation route,
addition, what reflected a decrease in the electron density at the Co sites however, can occur on metallic sites under basic conditions, which leads
after Co-Pt atoms exchange [31]. This bimetallic interaction, thus, to the formation of lighter oxygenated compounds (glycerol, propylene
induced an electron transfer from Co to Pt, and the consequent elec­ glycol) [35]. These compounds can also undergo further hydrogenolysis
tronic excess on Pt would lead to a lower interaction with the benzene reactions to yield monofunctionals and C2-C4 alkanes. Both routes can
formed during the cyclohexane dehydrogenation, and consequently to a produce, via decarbonylation or decarboxylation, shorter compounds,
higher TOF [11], in agreement with values shown in Table 1. and the formation of CO2 and/or CH4 either by WGS or by methanation.
In the reduced Co3O4 catalyst, 76 % of the surface cobalt was in the It is obvious that the product distribution will depend on the relative
metallic form, and remarkably lowered in the cobalt aluminate catalysts rates of C–C vs C–O bond cleavage. The C–O bond cleavage was
(35–47 % of the total Co). Interestingly, the proportion of metallic cobalt suggested to occur either through dehydrogenation to form adsorbed
in the Pt-doped assays was higher as compared to parent 0.625CoAl, in species followed by direct cleavage on metallic sites or by dehydration
line with previous H2-TPR data. on surface acid sites [36]. Since for Pt and Co the formation of metal­
As depicted in Fig. S3, the Al 2p peak at 74.2 eV can be assigned to Al –carbon bonds is more stable than metal–oxygen bonds, they had low
cations located in the octahedron sites of the spinel-phase [32]. The C–O bond breaking activity [20,37], so it is, preferably, in the acid sites
presence of surface Pt species was confirmed for both xPt/CoAl catalysts that dehydroxylation occurs. Low rates in C–C cleavage would led to
as a very weak Pt 4d5/2 peak at 314.0–314.2 eV, corresponding to the Pt0 large chain products while the selectivity to alkanes/oxygenates de­
species [33]. pends on the rates in C–O cleavage.
The surface Co/Al atomic ratio (Table S3) was considerably lower For the sake of simplicity, the myriad of products obtained were
divided into four main categories: Light Gas (LG), composed by CO, CO2,
and C1-C4 alkanes; Heavy Gas (HG), composed by C5-C6 alkanes; Mon­
Co
+
Co
0
Co
+
Co
0 ofunctionals (MF), involving all species that contain exactly one oxygen
Satellite functionality (e.g., mono-alcohols, ketones, aldehydes, heterocycles);
Satellite
and Higher Oxygenates (HO), which are all species with two or more
oxygen atoms (e.g., polyols, mixed alcohol–carbonyl compounds,
2p1/2
1Pt/CoAl 2p3/2 hydroxy-heterocycles, carboxylic acids).

3.3. Sorbitol hydrodeoxygenation (HDO)

All the catalysts were very active for sorbitol transformation under
H2 atmosphere, and achieved complete conversion of sorbitol after 4 h of
0.3Pt/CoAl
reaction, except bare Co3O4 (Fig. S4). A first-order reaction kinetic was
deduced by using the integral method (Fig. S5). Both xCoAl catalysts
Intensity (a.u.)

showed similar initial rates. Both Pt-doped catalysts showed high ac­
tivity in both HDO and APR conditions, due to the enhanced hydroge­
nation/dehydrogenation capability of the metal function (as shown in
0.625CoAl cyclohexane dehydrogenation). These results were in line with others
[38]. The good activity shown by 0.25CoAl catalyst could be attributed
to its greater acidity (Table 1) in relation to the 0.625CoAl catalyst.
The carbon distribution of the HDO reaction products according to
chain length (Fig. S6) indicated that all catalysts favoured the formation
of C1 compounds (C–C cleavage at the end of the carbon chain). The
0.25CoAl production of C2 to C4 compounds indicated multiple end-chain C–C
cleavage. The highest yield to C1 products was achieved by 0.625CoAl
(60 %) which showed low yields to C ≥ C2 products. These results
suggested that cobalt aluminate spinel based catalysts have potential to
be developed into active and tuneable catalysts for HDO reaction. For
Co3O4 instance, decarbonylation reaction mainly leads to C5 products along
with CO2 and others C1 compounds [39]. Bare Co3O4 seemed to favour
815 810 805 800 795 790 785 780 775 C5 compounds. Cobalt aluminate notably enhanced C–C scission to
Binding Energy (eV) yield light gaseous species (e.g. CO2, CH4). Instead, 0.25CoAl and xPt/
CoAl presented a better disposition for C3 compounds.
Fig. 4. High-resolution XPS spectra of Co 2p for the reduced solids. The carbon distribution values, measured at close sorbitol

5
A.J. Reynoso et al. Fuel 357 (2024) 129984

Scheme 1. Reaction pathways for sorbitol transformation over cobalt aluminate-derived catalysts, based on the obtained products.

conversions (Xsorb = 78–90 %), are presented in Table 2. In general glycerol) were taken as key products for retro-aldol condensation route.
terms, all catalysts produced a high proportion of liquid products (above Sorbitan was only detected for bare Co3O4, and was related to the
78 %), and the gas phase was mainly composed of hydrocarbons. particularly low surface acid site density of this catalyst (Table 1), which
Oxygenated liquid compounds (MF + HO) represented between 78 % limited the purely acidic-catalysed reaction route. The trimetallic 0.3Pt/
and 88 % of the products, while gaseous hydrocarbons accounted for CoAl showed a marginal 0.3 % yield of sorbitan, showing similar yields
between 11 % and 20 %. The higher CO2 production by bare Co3O4 and of isosorbide than 0.625CoAl. Isosorbide, instead, was detected for all
Pt-doped catalysts was due to the metal function with high dehydroge­ catalysts, with a maximum yield of 5.8 % for 1Pt/CoAl assay. The yield
nation activity which promotes C–C bond cleavage [40]. Among liquid to hexane was low (Yhexane < 1 %) by all the investigated catalysts as due
products, MF products exceeded HO products for bare Co3O4 and to the low reactivity of isosorbide, unfavourable for obtaining C6 alkanes
0.625CoAl assays, and were similar for the others. [11]. Among C2-C3 HO liquid compounds, 1,2-propylene glycol was, by
Fig. 5(a) depicts the yield to the four main product categories at HDO far, the most abundant. Compared to 0.625CoAl, Pt-containing catalysts
conditions. Bare Co3O4 produced mostly liquid products (54 % MF, 29 % boosted the yield to these products, which could be ascribed to their
HO). 0.625CoAl catalyst produced the highest yield to MF (66 %). As higher available metal, of high activity in cyclohexane dehydrogenation.
compared to bare Co3O4, production of HG compounds considerably Ethylene glycol was, overall, the second most abundant product what is
decreased for the cobalt aluminate spinel catalysts. A remarkable boost indicative of the effective conversion of glycerol. Assuming glycerol as
in yield to HO products was observed for trimetallic 0.3Pt/CoAl and the unique precursor of ethylene glycol and hydroxyacetone, the ratio
1Pt/CoAl catalysts if compared to its counterpart 0.625CoAl, due to the between yields to ethylene glycol/hydroxyacetone might be used as an
higher number of metallic centres of Pt-containing catalyst, of high indicator of the C–C/C–O scission activity. The trend was: 0.25CoAl
dehydrogenation activity (Table 1), as they promote retro-aldol (7.4) > 1Pt/CoAl (4.4) ≈ Co3O4 (4.1) > 0.3Pt/CoAl (3.1) > 0.625CoAl
condensation route [41]. (1.5). This trend reveals the metal/acid balance impact on product
To better elucidate the pathway favoured by each of our catalyst, the selectivity. Indeed, dehydrogenation may be favoured rather than
yields to key intermediate products were analysed in Fig. 5(b). Sorbitan dehydration depending on the distance between the acidic and metallic
and isosorbide (early products) and hexane (final product) were taken as sites of the catalyst [42]. Interestingly, 0.25CoAl and 1Pt/CoAl catalysts
key compounds of dehydration route; C2-C3 HO compounds (mainly exhibited the highest density of surface acid sites. These functional
composed by ethylene glycol, propylene glycol, hydroxyacetone and metal/acid dual sites could promote hydrogenolysis instead of high C–C

Table 2
Catalytic results at isoconversion for the HDO (78–90.5 %) and APR (49–57.4 %) reaction. Reaction conditions: 10 g of sorbitol, 150 mL of H2O, 240 ◦ C, 60 bar.
Catalyst Xgas (%) Carbon yield (molC %) YH2 (%)a

Gas Liquid

C1 C2-C4 C5+ C1-C4 MF C5+ MF C1-C4 HO C5+ HO

HDO conditions Co3O4 17.3 6.2 (2.4) 0.8 10.4 34.2 19.4 13.5 15.6 n.a.
0.25CoAl 11.7 7.2 (0.6) 0.8 3.6 28.7 11.7 39.6 8.4 n.a.
0.625CoAl 13.7 8.8 (0.2) 1.3 3.6 49.6 15.9 8.7 12.1 n.a.
0.3Pt/CoA1 21.8 15.3 (2.2) 0.8 5.7 24.3 13.4 27.4 13.1 n.a.
1Pt/CoAl 14.3 8.5 (2.5) 0.6 5.2 31.6 9.3 33.9 10.8 n.a.
APR conditions Co3O4 15.5 9.1 (0.6) 2.6 3.7 43.4 13.9 4.4 22.7 4.3
0.25CoAl 9.4 6.4 (0.7) 0.5 2.5 39.5 26.5 6.8 17.8 5.7
0.625CoAl 14.0 7.5 (0.6) 0.9 5.5 70.3 0.0 7.7 8.0 5.5
0.3Pt/CoA1 18.9 13.9 (2.0) 0.8 4.2 35.6 22.0 9.2 14.4 17.2
1Pt/CoAl 13.7 10.1 (1.2) 0.5 3.1 38.7 21.9 9.1 16.7 13.7

In parenthesis, yield to CO2; n.a.: not analysed; a Maximum hydrogen yield obtained during sorbitol conversion under N2 atmosphere.

6
A.J. Reynoso et al. Fuel 357 (2024) 129984

Fig. 5. Yield of products by categories (a) HDO, (c) APR; yield of selected products at sorbitol isoconversion (c) HDO, (d) APR.

scission activity producing CO2 and other C1 compounds [43]. In short, lesser availability of H2. Fig. S4 showed the time evolution of sorbitol
due to the low acidic character and neatly metallic character of our conversion over time. Unlike in sorbitol HDO, none of the studied cat­
Co–containing catalysts, they promoted the retro-aldol condensation alysts reached the complete sorbitol conversion in 8 h of reaction. Tri­
route for sorbitol HDO [20]. metallic Pt-Co catalysts were the most active assays for sorbitol
transformation. Around 93 % conversion was achieved by Pt-doped
3.4. Sorbitol aqueous-phase reforming (APR) catalysts at 8 h of reaction. On the contrary, bimetallic xCoAl and bare
Co3O4 were less active. The apparent kinetic constants (calculated
Under the so-called APR conditions (240 ◦ C/60 bar of N2), a positive assuming first order kinetics, Fig. S5) were 3–10 times lower than those
reaction order with respect to sorbitol concentration was obtained for all measured under HDO conditions. Activity results indicated that part of
the catalysts. The lower sorbitol conversion could be associated to the sorbitol produced hydrogen by the APR on the active sites of the

7
A.J. Reynoso et al. Fuel 357 (2024) 129984

catalyst, which was then used to the hydrogenolysis [44]. (HDO: Xsorb = 88.4 %, Xgas = 21.8 %; APR: Xsorb = 86.7 %, Xgas = 29.9
The yield to organic compounds during sorbitol APR (categorized by %).
the number of carbon atoms, Fig. S6) exposed some differences between According to Fig. 6(a), APR conditions led to a higher yield to hy­
cobalt-containing catalysts. Bare Co3O4 catalyst showed a high yield to drocarbons, thus decreasing the conversion to oxygenated liquids. For
the C1 and C6 compounds (57 and 16 %, respectively). Similar to HDO this catalyst, the production of CO2 two-fold increased by switching
conditions, 0.625CoAl catalyst showed low yields to C2+ products, being from APR to HDO conditions. Concerning the product categorization,
the catalyst with the lowest yields to C5-C6 products (3–4 %). On the APR yielded more MF species in contrast to the high HO yield observed
contrary, the production of C1 compounds was highest by this catalyst. in the HDO reaction. These manifest differences show that the absence of
The increased yield to C1 products as compared to HDO conditions could an external source of hydrogen limits dehydration-hydrogenation re­
be related to the H2 concentration in the reactor [39], which favour actions. Therefore, an inert atmosphere and the high metal site avail­
decarbonylation. For its part, 0.25CoAl and xPt/CoAl catalysts showed ability of 0.3Pt/CoAl catalyst favoured the cleavage of the C–C bond by
the highest yield to C5 products. At APR conditions, the yield to C2 and retro-aldol reaction.
C4 compounds was very low (<3.7 %) by all the investigated catalysts. Sorbitol HDO yielded more hydrogenated liquids (higher H/C ratio)
The carbon distribution at sorbitol isoconversion condition (49–57 % than APR (low H/C and O/C ratios) (Fig. S7). Overall, the obtained O/C
range) is shown in Table 2. Under inert atmosphere, even more carbon ratios were 0.72 and 0.76 for APR and HDO conditions. Some authors
was released as oxygenated liquids, as compared to HDO conditions. As indicated that more hydrogenated intermediates are favoured when
a general trend, slightly higher oxygenated liquids/hydrocarbons ratios C–O bond cleavage reactions take place [48]. Instead, the high occur­
were obtained at APR conditions (5–30 % higher than at HDO condi­ rence of C–C bonds scission in our catalytic system may be associated
tions). As compared to HDO conditions, at APR conditions all the cata­ with a high decarbonylation activity, which is also associated to the
lysts yielded higher MF/HO ratios (2.5-fold increase for Pt-containing production of CO2 and CH4, and deoxygenation reactions [49]. Thus,
catalysts), which suggested that the efficiency of in-situ hydrogenation catalytic behaviour observed in sorbitol APR would be explained.
using in-situ produced hydrogen was higher than that using external Sorbitol HDO over 0.3Pt/CoAl exhibited high yields to polyols and
hydrogen, in concordance with others [45]. As at HDO conditions, Pt- diols, as propylene glycol (11.8 % vs 7.4 % for APR), ethylene glycol
containing catalysts were the most active in CO2 (final reforming (6.0 % vs 2.9 %), glycerol (5.5 % vs 1.2 %), and 1,2-butanediol (4.1 % vs
product) production. 1.6 %). At APR conditions, the most relevant products were
At this conversion range, CO2 and CH4 (LG) were the main gaseous tetrahydropyran-2-methanol (THP-2-MeOH) and hydroxyacetone with
species obtained (joint yield 6–15 %), while HG yields did not exceed yields 7 and 3 times higher (respectively) than those obtained under
5.5 % (Fig. 5(c)). Given the carbon atom distribution (C1 = 42.3–81.1 %, HDO. Other products such as xylitol (2.7 %), erythrytol (1.5 %) and
Fig. S6) it can be asserted that the C–C bond cleavage activity of the
investigated catalysts was high. Among liquid compounds, MF were the
most abundant, with higher yields than those observed under HDO
conditions.
Both isosorbide and sorbitan were detected for all catalysts (Fig. 5d).
The yield to isosorbide was very similar for all catalysts. In general,
CoAl-containing catalysts were selective towards C3 liquid compounds,
which indicates that these catalysts, regardless of the reaction atmo­
sphere, favoured retro-aldol reaction route and subsequent dehydration
and hydrogenation reactions. Under inert atmosphere, hydroxyacetone
was the most abundant liquid product, as the low availability of
hydrogen limited the hydrogenation of hydroxyacetone. Hexane pro­
duction was insignificant (Yhexane ranged from 0.2 to 0.4 %). The highest
yield to H2 was obtained by 0.3Pt/CoAl catalyst with YH2 = 17.2 %.
Overall, hydrogen yield varied as follows: 0.3Pt/CoAl > 1Pt/CoAl ≫
0.25CoAl ≈ 0.625CoAl ≈ Co3O4. These data are consistent with the
performance of these catalysts in the glycerol reforming [21,22]. Despite
less amount of accessible Pt sites for 0.3Pt/CoAl catalysts if compared to
1Pt/CoAl, it achieved higher hydrogen yield, which could be due to the
stronger Pt-Co interaction [46], observed by H2-TPR and XPS, that affect
the adsorption energies [47].
Table S4 shows a comparison between the results of 0.625CoAl and
0.3Pt/CoAl catalysts with other works in the literature regarding the
production of value-added products from sorbitol, in batch reactor. Our
catalysts showed a sorbitol conversion comparable to that of catalysts
based on noble metals (metal loadings above 4 wt%). In spite of the high
sorbitol/catalyst mass ratio (gsorbitol/gcat = 10/1) used, our catalysts
obtained comparable selectivities at value-added product such as 1,2-
propylene glycol and C5-C6 alkanes, especially at HDO conditions. At
APR conditions, our catalysts were less active and selective than those
reported in the literature. The much severe reaction conditions for our
catalysts, and the not addition of a base additive could explain their
lower performance.

3.5. HDO vs APR over 0.3Pt/CoAl catalyst

The performance of 0.3Pt/CoAl catalyst for sorbitol transformation Fig. 6. (a) Comparative of carbon distribution, and (b) yield of categories of
under H2 and N2 atmosphere was compared at isoconversion conditions products over 0.3Pt/CoAl catalyst.

8
A.J. Reynoso et al. Fuel 357 (2024) 129984

tetrahydrofuran-2-methanol (THF-2-MeOH) (6.0 %) were produced at a H2-TPR of the exhausted catalysts (Fig. S11) showed H2 consumption
similar yield regardless of the pressurizing gas. Other liquid compounds peaks below 600 ◦ C. The uptake corresponding to the hydrogenation of
obtained at significant amounts in both processes were methanol, 2- carbonaceous deposits (as shown by TPH) was far below the total
methyltetrahydrofuran, 2-butanol, and 2-propanol. hydrogen uptake up to 600 ◦ C. Thus, the main uptake corresponded to
It is worth noting that produced tetrahydrofuran- and the reduction of oxidized Pt species and Coδ+ entities, generated during
tetrahydropyran-like chemicals are important solvent and building the reaction due to the oxidant reaction medium (Table 3), in concor­
blocks in the chemical industry. For instance, 2-methyltetrahydrofuran dance with others [53]. Results revealed that Co oxidation was lower
is an ecologically-friendly alternative to tetrahydrofuran to be utilized during HDO as compared to APR conditions, indicating H2 atmosphere
as a specialty solvent [50]. Tetrahydrofuran-2-methanol is a green sol­ could hinder its oxidation. Nevertheless, it is presumed that the oxida­
vent and a chemical intermediate for the pharmaceutical and chemical tion was constrained to the outmost layer of cobalt particles, since
specialities sector [51]. metallic Co could be detected by XRD. Noticeably, it occurred at a lower
extent in the Pt-containing catalysts, suggesting that the presence of Pt
could prevent Co from oxidation. This brought out the favourable effect
3.6. Characterization of spent catalysts of the noble metal, on which H2 could be dissociated, and subsequently
spill over onto cobalt [54].
Catalysts used in the APR and HDO of sorbitol (8 h of reaction) were The textural, structural and oxidation state changes undergone by
characterized by several techniques, and the obtained results are sum­ the catalysts would decrease their activity performance, therefore
marized in Table 3. The N2 adsorption–desorption isotherms of the regenerative treatment such as oxidation/reduction [55,56] could be
exhausted solids were similar to those of the reduced assays (Fig. S8) applied.
suggesting minor changes in the initial textural structure. However, Sorbitol forms a carbonaceous layer on the catalyst surface [57],
some quantitative changes could be observed. For example, SBET of which may protect the catalyst against leaching and phase trans­
xCoAl and Pt-containing catalysts increased during both HDO and APR formations. However, leaching of metals was observed for our catalysts.
experiments with concomitant decrease of the pore volume and the ICP-AES analysis of the post-reaction liquid, in line with previous TPH
average pore size (Table S2). Among xCoAl catalysts, the increase in SBET results, revealed that leaching was higher in inert atmosphere than
was less acute as the Co/Al ratio increased. Pt-containing catalysts under H2 pressure. The leached Co ranged in the 4.8–7.7 % interval for
showed smaller increase in SBET after usage, especially 0.3Pt/CoAl Co3O4 and xCoAl catalysts whereas it decreased to around 1.8–2.4 %
(10–16 % increase). Bare Co3O4 showed a notable development of sur­ after Pt doping, which evidenced that Pt prevents Co oxidation during
face area (from 7.4 to 21–22 m2/g, Table 3) with a substantial decrease the reaction [58]. Leaching was not exclusive to cobalt, and around 6 to
(by 50 %) in the average pore size. As a general trend, APR caused more 9 % of the total Al in xCoAl catalysts leached-off. Again, leaching of Al
impact than HDO conditions in SBET. The hydration of γ-alumina to was notably decreased in the xPt/CoAl catalysts. In the worst scenario,
boehmite and the presence of significant amounts of carbon deposits on up to 0.20 % of the available Pt was leached (e.g. 1Pt/CoAl, APR reac­
the catalyst surface (discussed below) could contribute to such variation tion). Despite the occurrence of the unwanted leaching of the metals, it
in the textural properties. is still unclear the role of this phenomenon in the catalytic activity, since
All the Al-containing catalysts preserved the characteristic XRD tests carried out in batch reactor showed that the leached species, as
peaks from spinel (Fig. S9). Characteristic peaks from metallic cobalt they remain in the reaction medium, could still act as active species
were still detected for all catalysts, which suggested that bulk cobalt through homogeneous catalysis [59]. Overcoming metal leaching in the
remained in metallic form. Well-defined peaks of metallic cobalt aqueous-phase biomass treatment processes remains a challenge. Stra­
remained unaltered for Co3O4 catalyst. For the Al-containing assays, the tegies such as coating, promotion of metal-support interaction, embed­
crystallite sizes of metallic Co increased for all the spent catalysts after ding metal particles, or alloying are feasible methods since they can
reaction, indicating their coalescence, in line with others [20]. The prevent both sintering and leaching of metals [60,61].
absence of Pt peaks indicated it still was highly dispersed. It is suggested
that Co atoms are removed from the catalyst and redeposited to form 4. Conclusions
larger, more thermodynamically stable particles catalysts [52], which
was a critical problem for these catalysts as decreased the accessible Liquid-phase sorbitol hydrogenolysis at different reaction conditions
metal sites. New sharp peaks due to boehmite (PDF 01-088-2112) was studied over bare and promoted cobalt aluminate-derived catalysts.
emerged for the Al-containing. From Fig. S9 it seems to be favoured It was validated that the reaction atmosphere is determinant for tuning
under HDO as compared to APR conditions. Note that among the xCoAl the relative rates of C–C versus C–O bond cleavage, modifying the
series, the alumina hydration to boehmite was inhibited in the Co-rich product selectivity. HDO conditions favoured production of higher
specimens (0.625CoAl and Pt-doped catalysts). No peak from carbon oxygenated compounds whereas decarbonylation was limited. APR
was observed, indicating it was highly dispersed and/or amorphous. conditions favoured dehydrogenation/deoxygenation products. 1,2-pro­
The carbonaceous deposits were quantified by TPH coupled to MS pylene glycol and ethylene glycol were leading products under HDO
(Table 3, Fig. S10). Methane was released in the interval 400-700 ◦ C in conditions, while under APR conditions, hydroxyacetone top the list.
the range 20–77 µmolC/gcat, with no significant differences among HDO Platinum doping, in addition to considerably increase H2 production
and APR conditions.

Table 3
Characterization of the spent catalysts.
Catalyst SaBET (m2/g) dCo0b (nm) Ccdeposits (μmolC/gcat) Co oxidizedd (%) Leached @ HDO conditions (%)e Leached @ APR conditions (%)e

HDO APR HDO APR HDO APR HDO APR Co Al Pt Co Al Pt

Co3O4 21.0 22.3 54.3 55.7 62.3 65.4 0.7 0.6 5.2 n.a. n.a. 5.6 n.a. n.a.
0.25CoAl 167.6 205.9 54.0 33.3 20.9 20.3 9.8 11.3 4.8 7.8 n.a. 5.5 9.0 n.a.
0.625CoAl 127.2 132.3 28.0 n.a. 39.1 40.4 8.0 11.8 7.1 5.9 n.a. 7.7 7.4 n.a.
0.3Pt/CoAl 151.8 143.8 12.2 n.a. 29.5 31.6 5.2 5.6 1.8 1.4 0.05 1.7 1.8 0.04
1Pt/CoAl 152.8 169.5 18.0 n.a. 74.6 76.9 3.0 3.2 2.0 1.1 0.14 2.4 1.3 0.20

n.a.: not analysed; a from nitrogen isotherms; b

from XRD; c from TPH; d
from H2 consumption up to 600 ◦ C of spent catalysts with respect to calcined forms of the
catalysts; e from ICP-AES.

9
A.J. Reynoso et al. Fuel 357 (2024) 129984

under APR conditions, enhanced the production of monofunctionals, [9] Shahbudin MI, Jacob DM, Ameen M, Aqsha A, Azizan MT, Yusoff MHM, et al.
Liquid value-added chemicals production from aqueous phase reforming of sorbitol
which was attributed to the strong Pt-Co interaction. The physico-
and glycerol over sonosynthesized Ni-based catalyst. J Environ Chem Eng 2021;9
chemical characterization of the spent catalysts indicated that Pt could (4):105766.
stabilize the catalyst in both HDO and APR conditions, reducing both the [10] He Z, Wang X. Hydrodeoxygenation of model compounds and catalytic systems for
leaching and the degree of sintering of cobalt, and hindering the pyrolysis bio-oils upgrading. Catal. Sustainable Energy 2012;1(2013):28–52.
[11] Messou D, Vivier L, Especel C. Sorbitol transformation over bimetallic Ru-Pt/SiO2-
oxidation of cobalt. Al2O3 catalysts: effect of the preparation method. Energy Convers Manag 2016;
Boehmite formation, cobalt oxidation and sintering, deposition of 127:55–65.
small amount of carbonaceous material, and Co and Al leaching were [12] Messou D, Vivier L, Especel C. Sorbitol transformation into biofuels over bimetallic
platinum based catalysts supported on SiO2-Al2O3-Effect of the nature of the
identified the main deactivation phenomena. Thus, the main challenge second metal. Fuel Process Technol 2018;177:159–69.
remains to overcome oxidation and subsequent leaching which could [13] Duarte HA, Sad ME, Apesteguía CR. Aqueous phase reforming of sorbitol on Pt/
deactivate the catalysts. Al2O3: effect of metal loading and reaction conditions on H2 productivity. Int J
Hydrog Energy 2016;41(39):17290–6.
[14] Sun F, Chen L, Weng Y, Wang T, Qiu S, Li Q, et al. Transformation of biomass
CRediT authorship contribution statement polyol into hydrocarbon fuels in aqueous medium over Ru-Mo/CNT catalyst. Catal
Commun 2017;99:30–3.
[15] Messou D, Vivier L, Canaff C, Especel C. Biofuel synthesis from sorbitol by aqueous
A.J. Reynoso: Conceptualization, Investigation, Writing – original phase hydrodeoxygenation over bifunctional catalysts: in-depth study of the
draft, Data curation. J.L. Ayastuy: Funding acquisition, Conceptuali­ Ru–Pt/SiO2–Al2O3 catalytic system. Catalysts 2019;9(2):146.
zation, Writing – review & editing, Data curation. U. Iriarte-Velasco: [16] Weng Y, Qiu S, Xu Y, Ding M, Chen L, Zhang Q, et al. One-pot aqueous phase
catalytic conversion of sorbitol to gasoline over nickel catalyst. Energy Convers
Writing – review & editing, Formal analysis. L. Vivier: Supervision, Manag 2015;94:95–102.
Resources, Methodology. C. Especel: Supervision, Resources, Method­ [17] Zan Y, Miao G, Wang H, Kong L, Ding Y, Sun Y. Revealing the roles of components
ology. M.A. Gutiérrez–Ortiz: Resources, Funding acquisition, in glucose selective hydrogenation into 1,2-propanediol and ethylene glycol over
Ni-MnOx-ZnO catalysts. J Energy Chem 2019;38:15–9.
Supervision. [18] Jia Y, Liu H. Selective hydrogenolysis of sorbitol to ethylene glycol and propylene
glycol on ZrO2-supported bimetallic Pd-Cu catalysts. Chin J Catal 2015;36(9):
1552–9.
Declaration of Competing Interest [19] Kong Y, Zhu Y, An Z, Zhang J, Shu X, Song H, et al. Efficient selective activation of
sorbitol C-O bonds over C–C bonds on CoGa (221) generated by lattice-induction
The authors declare that they have no known competing financial strategy. Nano Res 2022.
[20] Eagan NM, Chada JP, Wittrig AM, Buchanan JS, Dumesic JA, Huber GW.
interests or personal relationships that could have appeared to influence Hydrodeoxygenation of Sorbitol to Monofunctional Fuel Precursors over Co/TiO2.
the work reported in this paper. Joule 2017;1(1):178–99.
[21] Reynoso AJ, Ayastuy JL, Iriarte-Velasco U, Gutiérrez-Ortiz MA. Cobalt aluminate
spinel-derived catalysts for glycerol aqueous phase reforming. Appl Catal B-
Data availability Environ 2018;239:86–101.
[22] Reynoso AJ, Iriarte-Velasco U, Gutiérrez-Ortiz MA, Ayastuy JL. Highly stable Pt/
Data will be made available on request. CoAl2O4 catalysts in aqueous-phase reforming of glycerol. Catal Today 2021;367:
278–89.
[23] Maris EP, Ketchie WC, Murayama M, Davis RJ. Glycerol hydrogenolysis on carbon-
Acknowledgments supported PtRu and AuRu bimetallic catalysts. J Catal 2007;251(2):281–94.
[24] Li G, Hu L, Hill JM. Comparison of reducibility and stability of alumina-supported
Ni catalysts prepared by impregnation and co-precipitation. Appl Catal A-Gen
This research was supported by grant PID2019-106692RB-I00 fun­ 2006;301(1):16–24.
ded by MCIN/AEI/10.13039/501100011033. A.J. Reynoso would like [25] Nabaho D, Niemantsverdriet JW, Claeys M, van Steen E. Hydrogen spillover in the
to thank Dominican Republic MESCYT for the financial support. Like­ Fischer-Tropsch synthesis: an analysis of platinum as a promoter for
cobalt–alumina catalysts. Catal Today 2016;261:17–27.
wise, the authors thank for technical support provided by SGIker of
[26] Choojun K, Worathanaseth A, Kuhatasanadeekul S, Kurato T, Ketaniruj S,
UPV/EHU and European funding (ERDF and ESF). Phichitsurathaworn P, et al. Effect of cobalt complex precursors on reactivity of
cationic cobalt catalysts: cyclohexane dehydrogenation. Catal Commun 2019;125:
108–13.
Appendix A. Supplementary material [27] Guczi L, Sharma KV, Koppány ZS, Sundararajan R, Zsoldos Z. Comparative study
on low temperature methane activation over cobalt and ruthenium supported on
Supplementary data to this article can be found online at https://doi. Al2O3, SiO2 and NaY. In: de Pontes M, Espinoza RL, Nicolaides CP, Scholtz JH,
Scurrell MS, editors. Studies in surface science and catalysis, Vol. 107. Elsevier;
org/10.1016/j.fuel.2023.129984.
1997. p. 333–8.
[28] Yang X, Sun Z, Wang D, Forsling W. Surface acid–base properties and hydration/
References dehydration mechanisms of aluminum (hydr)oxides. J Colloid Interface Sci 2007;
308(2):395–404.
[29] Yang J, Liu H, Martens WN, Frost RL. Synthesis and characterization of cobalt
[1] Cherubini F. The biorefinery concept: using biomass instead of oil for producing
hydroxide, cobalt oxyhydroxide, and cobalt oxide nanodiscs. J Phys Chem C 2010;
energy and chemicals. Energy Convers Manag 2010;51(7):1412–21.
114(1):111–9.
[2] Bozell JJ, Petersen GR. Technology development for the production of biobased
[30] Biesinger MC, Payne BP, Grosvenor AP, Lau LWM, Gerson AR, Smart RSC.
products from biorefinery carbohydrates-the US Department of Energy‫׳‬s “Top 10”
Resolving surface chemical states in XPS analysis of first row transition metals,
revisited. Green Chem 2010;12:539–54.
oxides and hydroxides: Cr, Mn, Fe, Co and Ni. Appl Surf Sci 2011;257(7):2717–30.
[3] Silva JGAB, Santos RC, Rodríguez-Castellón E, Teixeira LSG, Pontes LAM. Catalytic
[31] Zhang JM, Sun SN, Li Y, Zhang XJ, Zhang PY, Fan YJ. A strategy in deep eutectic
conversion of glucose into sorbitol over niobium oxide supported Ru catalysts. Mol
solvents for carbon nanotube-supported PtCo nanocatalysts with enhanced
Catal 2021;507:111567.
performance toward methanol electrooxidation. Int J Hydrogen Energ 2017;42
[4] Jofre FM, Bordini FW, Bianchini IA, Queiroz SS, Boaes TS, Hernández-Pérez AF,
(43):26744–51.
Felipe MGA. 8 - Xylitol and sorbitol: production routes, challenges and
[32] Zhu MP, Yang JCE, Sun DD, Yuan B, Fu ML. Deciphering the simultaneous removal
opportunities in biorefineries integration. In: Chandel AK, Segato F, editors.
of carbamazepine and metronidazole by monolithic Co2AlO4@Al2O3 activated
Production of Top 12 Biochemicals Selected by USDOE from Renewable Resources.
peroxymonosulfate. Chem Eng J 2022;436:135201.
Elsevier; 2022. p. 233–68.
[33] Valekar AH, Oh KR, Lee SK, Hwang YK. Simultaneous production of lactate and
[5] Qiu S, Huang Z, Sang X, Liu Y, Zhang Q, Chen Q, et al. Regulating isomerization of
formate from glycerol and carbonates over supported Pt catalyst. J Ind Eng Chem
gasoline-alkanes in aqueous-phase hydrodeoxygenation of sorbitol using
2021;101:66–77.
regenerable Ni@MoOx catalysts. Fuel Process Technol 2023;242:107647.
[34] Kannan S, Swamy CS. Catalytic decomposition of nitrous oxide over calcined cobalt
[6] Li X, Yan L, Liu Z, Liao Y, Ma L. Low-temperature hydrodeoxygenation of phenol
aluminum hydrotalcites. Catal Today 1999;53(4):725–37.
using modified TiO2(B) nanosheets supported highly dispersed Pt catalyst. Fuel
[35] Vilcocq L, Cabiac A, Especel C, Guillon E, Duprez D. Transformation of sorbitol to
2023;338:127314.
biofuels by heterogeneous catalysis: chemical and industrial considerations. Oil
[7] Chumachenko YA, Shkurenok VA, Smolikov MD, Gulyaeva TI, Buluchevskiy EA,
Gas Sci Technol-Rev IFP Energies nouvelles 2013;68(5):841–60.
Prosvirin IP, et al. Continuous one-pot hydrodeoxygenation of sorbitol to fuel
[36] Peng B, Zhao C, Mejía-Centeno I, Fuentes GA, Jentys A, Lercher JA. Comparison of
components over Pd/WOx–ZrO2 catalysts. Mol Catal 2021;513:111730.
kinetics and reaction pathways for hydrodeoxygenation of C3 alcohols on Pt/
[8] Xu H, Ju J, Li H. Toward efficient heterogeneous catalysts for in situ
Al2O3. Catal Today 2012;183(1):3–9.
hydrodeoxygenation of biomass. Fuel 2022;320:123891.

10
A.J. Reynoso et al. Fuel 357 (2024) 129984

[37] Zhao H, Zheng L, Li X, Chen P, Hou Z. Hydrogenolysis of glycerol to 1,2-pro­ [49] Annuar NHR, Alexzman ZA, Daud ARM, Alias AFN, Hairi HM, Setiabudi HD.
panediol over Cu-based catalysts: a short review. Catal Today 2020;355:84–95. A review on hydrogenolysis of sorbitol over heterogeneous catalysts. J Environ
[38] Do PTM, Foster AJ, Chen J, Lobo RF. Bimetallic effects in the hydrodeoxygenation Chem Eng 2022;10(2):107229.
of meta-cresol on γ-Al2O3 supported Pt–Ni and Pt–Co catalysts. Green Chem 2012; [50] Aycock DF. Solvent applications of 2-methyltetrahydrofuran in organometallic and
14:1388–97. biphasic reactions. Org Process Res Dev 2007;11(1):156–9.
[39] Moreno BM, Li N, Lee J, Huber GW, Klein MT. Modeling aqueous-phase [51] Englezou G, Kortsen K, Pacheco AAC, Cavanagh R, Lentz JC, Krumins E, et al. 2-
hydrodeoxygenation of sorbitol over Pt/SiO2–Al2O3. RSC Adv 2013;3:23769–84. Methyltetrahydrofuran (2-MeTHF) as a versatile green solvent for the synthesis of
[40] Kim S, Kwon EE, Kim YT, Jung S, Kim HJ, Huber GW, et al. Recent advances in amphiphilic copolymers via ROP, FRP, and RAFT tandem polymerizations. J Polym
hydrodeoxygenation of biomass-derived oxygenates over heterogeneous catalysts. Sci 2020;58(11):1571–81.
Green Chem 2019;21:3715–43. [52] Besson M, Gallezot P. Deactivation of metal catalysts in liquid phase organic
[41] Kandasamy T, Banu M, Shanthi RV, Sivasanker S. Suitability of different supported reactions. Catal Today 2003;81(4):547–59.
Ru, Pt and Ni catalysts for the hydrogenolysis of sorbitol. Results Eng 2022;15: [53] Putrakumar B, Nagaraju N, Kumar VP, Chary KVR. Hydrogenation of levulinic acid
100594. to γ-valerolactone over copper catalysts supported on γ-Al2O3. Catal Today 2015;
[42] Vilcocq L, Cabiac A, Especel C, Lacombe S, Duprez D. New insights into the 250:209–17.
mechanism of sorbitol transformation over an original bifunctional catalytic [54] Prins R. Hydrogen spillover. Facts and fiction. Chem Rev 2012;112(5):2714–38.
system. J Catal 2014;320:16–25. [55] Qing S, Hou X, Li L, Feng G, Wang X, Gao Z, et al. Deactivation feature of Cu/SiO2
[43] Vikla AKK, Koichumanova K, He S, Seshan K. Aqueous-phase reforming of catalyst in methanol decomposition. Int J Hydrogen Energ 2019;44(31):16667–74.
hydroxyacetone solution to bio-based H2 over supported Pt catalysts. Green Energy [56] Argyle MD, Bartholomew CH. Heterogeneous catalyst deactivation and
Environ 2022. https://doi.org/10.1016/j.gee.2022.09.002 (in press). regeneration: a review. Catalysts 2015;5(1):145–269.
[44] Zhang J, Lu F, Yu W, Chen J, Chen S, Gao J, et al. Selective hydrogenative cleavage [57] Ravenelle RM, Copeland JR, Van Pelt AH, Crittenden JC, Sievers C. Stability of Pt/
of C-C bonds in sorbitol using Ni–Re/C catalyst under nitrogen atmosphere. Catal γ-Al2O3 catalysts in model biomass solutions. Top Catal 2012;55:162–74.
Today 2014;234:107–12. [58] Zhang L, Mao J, Li S, Yin J, Sun X, Guo X, et al. Hydrogenation of levulinic acid into
[45] Zhang S, Yang X, Zheng K, Xiao R, Hou Q, Liu B, et al. In situ hydrogenation of gamma-valerolactone over in situ reduced CuAg bimetallic catalyst: Strategy and
furfural conversion to furfuryl alcohol via aqueous-phase reforming of methanol. mechanism of preventing Cu leaching. Appl Catal B-Environ 2018;232:1–10.
Appl Catal A-Gen 2019;581:103–10. [59] Sievers C, Noda Y, Qi L, Albuquerque EM, Rioux RM, Scott SL. Phenomena
[46] Dietrich PJ, Akatay MC, Sollberger FG, Stach EA, Miller JT, Delgass WN, et al. affecting catalytic reactions at solid-liquid interfaces. ACS Catal 2016;6(12):
Effect of Co loading on the activity and selectivity of PtCo aqueous phase reforming 8286–307.
catalysts. ACS Catal 2014;4(2):480–91. [60] Huo J, Tessonnier JP, Shanks BH. Improving hydrothermal stability of supported
[47] Huber GW, Shabaker JW, Evans ST, Dumesic JA. Aqueous-phase reforming of metal catalysts for biomass conversions: a review. ACS Catal 2021;11(9):5248–70.
ethylene glycol over supported Pt and Pd bimetallic catalysts. Appl Catal B:Environ [61] Xiong H, Schwartz TJ, Andersen NI, Dumesic JA, Datye AK. Graphitic-carbon
2006;62(3–4):226–35. layers on oxides: toward stable heterogeneous catalysts for biomass conversion
[48] West RM, Kunkes EL, Simonetti DA, Dumesic JA. Catalytic conversion of biomass- reactions. Angew Chem Int Ed 2015;54:7939–43.
derived carbohydrates to fuels and chemicals by formation and upgrading of mono-
functional hydrocarbon intermediates. Catal Today 2009;147(2):115–25.

11