The kinetics of glycerol hydrodeoxygenation to 1,2-propanediol over Cu/ZrO2 in the aqueous phase

Article abstract of DOI:10.1016/j.apcata.2019.03.001

The kinetics of glycerol hydrodeoxygenation to 1,2-propanediol via the selective cleavage of the primary C-O bond was systematically studied in the aqueous phase over a co-precipitated Cu/ZrO₂ catalyst. Unsupported pure metallic Cu was used as reference catalyst. Batch experiments were performed in an autoclave by varying the reaction temperature (175–225 °C), H₂ partial pressure (25–35 bar) and initial glycerol concentration (2–8 wt%). The Cu/ZrO₂ catalyst was found to be highly selective to 1,2propanediol (up to 95%), and ethylene glycol was obtained as major by-product from parallel C–]C bond hydrogenolysis. The apparent activation energies amounting to 106 and 105 kJ mol^-1 for Cu/ZrO₂ and pure metallic Cu, respectively, of the hydrodeoxygenation pathway provide further evidence for metallic Cu acting as the active site. Kinetic analysis of the rate of glycerol consumption yielded a zero-order dependence on the concentration of glycerol suggesting an essentially almost full coverage of adsorbed glycerol as most strongly bound organic adsorbate. In contrast, a first-order dependence on hydrogen concentration was observed. Hydrogen is assumed to be not only required for the fast hydrogenation of the intermediate acetol, but also for the removal of adsorbed atomic oxygen originating from water dissociation to create empty sites for dissociative glycerol adsorption. Thus, the active Cu sites are assumed to be fully adsorbate-covered under reaction conditions.

Full text of DOI:10.1016/j.apcata.2019.03.001

Accepted Manuscript
Title: The kinetics of glycerol hydrodeoxygenation to
1,2-propanediol over Cu/ZrO2 in the aqueous phase
Authors: Thomas Gabrysch, Martin Muhler, Baoxiang Peng
PII:
S0926-860X(19)30100-0
DOI:
Reference:
https://doi.org/10.1016/j.apcata.2019.03.001
APCATA 17006
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
12 October 2018
25 February 2019
2 March 2019
Please cite this article as: Gabrysch T, Muhler M, Peng B, The kinetics of glycerol
hydrodeoxygenation to 1,2-propanediol over Cu/ZrO2 in the aqueous phase, Applied
Catalysis A, General (2019), https://doi.org/10.1016/j.apcata.2019.03.001
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

The kinetics of glycerol hydrodeoxygenation to 1,2-propanediol
over Cu/ZrO2 in the aqueous phase
Thomas Gabrysch1, Martin Muhler1,2, and Baoxiang Peng*1,2
1Laboratory of Industrial Chemistry, Ruhr-University Bochum, Universitätsstr. 150, 44780 Bochum, Germany
2Max Planck Institute for Chemical Energy Conversion, Stiftstraße 34-36, 45470 Mülheim an der Ruhr, Germany
*Corresponding author, Tel.: +49 234 32 24102, Email address: baoxiang.peng@techem.rub.de (B. Peng)
Graphical abstract

Highlights


The kinetics of glycerol hydrodeoxygenation to 1,2-propanediol over Cu/ZrO2 has
been systematically studied.
The kinetic study yielded a zero-order dependence on glycerol concentration and a
first-order dependence on H2 concentration


H2 removes adsorbed atomic oxygen originating from water dissociation
The similar apparent activation energies for the hydrodeoxygenation over Cu/ZrO2
and Cu powder further identify the active sites as fully adsorbate-covered metallic
Cu.
Abstract
The kinetics of glycerol hydrodeoxygenation to 1,2-propanediol via the selective cleavage of the
primary C-O bond was systematically studied in the aqueous phase over a co-precipitated Cu/ZrO2
catalyst. Unsupported pure metallic Cu was used as reference catalyst. Batch experiments were
performed in an autoclave by varying the reaction temperature (175 – 225 °C), H2 partial pressure
(25 – 35 bar) and initial glycerol concentration (2 – 8 wt%). The Cu/ZrO2 catalyst was found to be
highly selective to 1,2-propanediol (up to 95%), and ethylene glycol was obtained as major by-
product from parallel C–C bond hydrogenolysis. The apparent activation energies amounting to
106 and 105 kJ mol-1 for Cu/ZrO2 and pure metallic Cu, respectively, of the hydrodeoxygenation
pathway provide further evidence for metallic Cu acting as the active site. Kinetic analysis of the rate
of glycerol consumption yielded a zero-order dependence on the concentration of glycerol
suggesting an essentially almost full coverage of adsorbed glycerol as most strongly bound organic
adsorbate. In contrast, a first-order dependence on hydrogen concentration was observed. Hydrogen
is assumed to be not only required for the fast hydrogenation of the intermediate acetol, but also for
the removal of adsorbed atomic oxygen originating from water dissociation to create empty sites for
dissociative glycerol adsorption. Thus, the active Cu sites are assumed to be fully adsorbate-covered
under reaction conditions.
Keywords: hydrogenolysis; ethylene glycol; metallic copper; kinetic study

1. Introduction
In recent years increasing bio-diesel production resulted in an excess production of glycerol, which is
the main by-product of this process.[1-3] The further utilization of biomass-based aqueous glycerol
solutions to value-added products such as ethers,[4,5] esters,[6-8] propanediols[9-11] and
acrolein[12-14] is of great interest to increase the economic efficiency and environmental
sustainability of bio-refineries. A promising utilization route is the selective hydrodeoxygenation
(HDO) of glycerol to 1,2-propanediol (1,2 PDO). HDO requires the selective cleavage of a terminal C-O
bond while preserving C-C bonds, whereas hydrogenolysis refers in particular to the cleavage of C-C
bonds. Most previous studies mainly focused on developing suitable HDO catalysts such as supported
noble (e.g., Pt, Ru)[15-17] and less active, but selective non-noble (e.g., Cu, Ni, Co)[18-20] metals and
characterising their relevant properties such as acidity, surface areas and metal dispersion.[21-23]
Among these two classes of catalysts, Cu-based catalysts were found to be optimum, combining high
selectivity to 1,2-PDO and moderate activity.[24-26] Owing to the mechanistic complexity of HDO,
the number of studies focussing on reaction kinetics is rather low.
A kinetic analysis of glycerol HDO over Cu:Zn:Cr:Zr mixed metal oxide catalysts was conducted by
Sharma et al.[27] using high initial glycerol concentrations up to 100%. The addition of Zn and Zr to
the catalyst matrix significantly increased glycerol conversion and the selectivity to 1,2-PDO, which
was ascribed to enhanced bifunctionality. Assuming a Langmuir-Hinshelwood-Hougen-Watson
reaction mechanism following the dehydration-hydrogenation pathway, pseudo-first order kinetics
for glycerol consumption and an activation energy of 132 kJ mol-1 were derived for the HDO route.
Vasiliadou et al.[28,29] studied the reaction kinetics of glycerol HDO over a 18 wt% Cu/SiO2 catalyst.
Two parallel routes leading to the formation of 1,2-PDO (95% selectivity) and 1,3-PDO (main by-
product) were considered, and the kinetic parameters were derived assuming a power-law kinetic
model. A reaction order of 0.17 was found for the overall glycerol consumption, and a reaction order
of 1.06 was derived considering the H2 concentration in the liquid phase. The activation energies for
the 1,2-PDO and 1,3-PDO routes were found to be 94.3 and 135.3 kJ mol-1, respectively. However,
this work was performed using 1-butanol as solvent. The consideration of possible solvent effects is
an important factor for liquid-phase reactions, and water is usually preferred to any organic solvents.
Bienholz et al.[30] reported that glycerol conversion increased from 5% to 55% over a Cu/ZnO
catalyst when water was replaced by 1,2-butanediol as solvent. N2O reactive frontal chromatography
measurements using the catalyst before and after reaction revealed that the use of 1,2-butanediol
led to a lower decrease of the Cu surface area due to inhibited growth of the Cu particles. Wang et

al.[31] studied solvent effects with a Cu/ZnO catalyst using water, methanol and ethanol. It was
found that solvents with low surface tension such as methanol and ethanol tend to inhibit the
agglomeration of Cu particles, thus resulting in higher degrees of Cu dispersion and higher glycerol
conversion, while highly polar solvents such as water tend to increase the product selectivity to 1,2-
PDO.
In our previous study, HDO of glycerol over Cu/ZrO2 in the aqueous phase was found to proceed via a
two-step dehydration-hydrogenation reaction pathway, yielding 1,2-PDO as main product with a high
selectivity of 95% at 31% conversion under moderate reaction conditions (200 °C, 25 bar H2, 8 h
reaction time).[32] A linear dependence of glycerol conversion on the specific Cu surface area was
observed. Thus, metallic Cu was identified as the active site for both dehydration and hydrogenation,
whereas the acidic sites of the ZrO2 support had no detectable effect on the catalytic activity under
the applied aqueous reaction conditions.
The aim of the present work is to investigate glycerol HDO kinetics under aqueous conditions in the
presence of an active, highly selective and rather stable Cu/ZrO2 catalyst. In contrast to other kinetic
studies claiming acid sites of the support to catalyse dehydration followed by hydrogenation on
metallic sites, here a kinetic study over Cu-based catalysts is reported, which provides no evidence
for bifunctionality. The reaction kinetics of glycerol HDO was first studied over Cu/ZrO2 with 18 wt%
CuO nominal loading by varying the reaction temperature (175 – 225 °C), H2 partial pressure
(25 - 35 bar) as well as the initial glycerol concentration (2 – 8 wt%). Unsupported pure metallic Cu,
derived from the reduction of commercial CuO nanopowder, was used as reference catalyst in
comparison to the Cu/ZrO2 catalyst to prove the dominant role of metallic Cu in the HDO of glycerol
kinetically.
2. Experimental
2.1. Materials
The Cu/ZrO2 catalyst was prepared using copper nitrate (Cu(NO3)2 ∙ 3 H2O, Sigma Aldrich, 99.999%),
zirconium oxynitrate (ZrO(NO3)2 ∙ 3 H2O, Sigma Aldrich, 99.0%) and sodium hydroxide (NaOH, VWR
Chemicals, 98.5-100.5%). Glycerol (VWR Chemicals, 99.5%), 1,2-propenediol (Fluka, 99.5%), ethylene
glycol (Alfa Aesar, 99%), hydroxyacetone (Alfa Aesar, 95%) and methanol (Fischer Chemicals, 99.8%)
were used for experiments and GC calibration. CuO powder (Sigma Aldrich, < 50 nm) was used as
reference catalyst precursor.

2.2. Catalyst preparation
The 18 wt% CuO/ZrO2 precursor was prepared by co-precipitation at room temperature.[32] Copper
nitrate and zirconium oxynitrate were dissolved in 100 mL deionized water (HPLC purity). 25 wt%
aqueous solution of sodium hydroxide was used as precipitating agent. The pH value of the solution
was maintained constant at 10.5 during precipitation in an autotitrator. After precipitation the
suspension was aged for 15 min at room temperature under continuously stirring. The obtained
precipitate was filtered and washed several times with deionized water until the filtrate was free of
nitrate. After drying at 105 °C overnight, the precursor was calcined at 490 °C for 3 h in synthetic air
(20% O2/N2) and ground to particle sizes < 25 µm.
2.3. Kinetic experiments
The kinetic experiments were carried out in a Parr 5050 stainless-steel autoclave reactor (100 mL)
equipped with temperature controller, mechanical stirrer and sampling tube for taking liquid
products. Prior to each experiment, CuO/ZrO2 and CuO nanopowder were pre-reduced in the reactor
in the absence of the aqueous glycerol solution. For each reaction 400 mg of CuO/ZrO2 or a defined
amount of CuO was loaded in the reactor. Subsequently, the reactor was pressurized and purged
several times with H2 (99.999%) to remove ambient air. The pre-reduction was carried out at 200 °C
and 20 bar H2 pressure for 10 h to reduce CuO/ZrO2 and CuO powder to metallic Cu. After cooling
and depressurization, 20 mL of 4 wt% aqueous glycerol solution was dosed into the autoclave
through the sampling tube without opening the reactor to minimize the exposure of the reduced
catalyst to ambient air. After flushing the reactor several times with H2 again, the reactor was
pressurized with H2. For each temperature variation experiment the total pressure in the reactor was
kept constant at 40 bar. The reaction was performed in the temperature range from 175 to 225 °C,
using 2 to 8 wt% initial glycerol concentration and 25 to 35 bar H2 pressure for 8 h with a stirring
speed of 750 rpm. During reaction liquid samples of approximately 1.5 mL were taken from the
reactor after 1, 3, 5 and 8 h. The liquid samples were filtered using membrane filters to remove
residual catalyst particles and analysed by gas chromatography (GC).
2.4. Quantitative analysis of liquid products

Liquid-phase reaction products were analysed using an Agilent 7820A gas chromatography equipped
with a capillary column (Phenomenex Zebron ZB-WAXplus, 30 m × 0.32 mm × 0.25 µm) and a flame
ionization detector (FID). The GC was calibrated by using an external standard resulting in relative
standard deviations (RSD%) of 2-4% depending on the different compounds. The conversion of
glycerol and the selectivity of liquid-phase products were calculated using the following two
equations:
푚표ꢃꢆꢅ 표ꢈ ꢇ표ꢂ푣ꢆ푟푡ꢆ푑 푔ꢃ푦ꢇꢆ푟표ꢃ
[ ]
퐶표푛푣푒푟푠ꢀ표푛 표푓 푔푙푦푐푒푟표푙 % =
∙ 100
(1)
(2)
푚표ꢃꢆꢅ 표ꢈ ꢁꢂꢁ푡ꢁ푎ꢃ 푔ꢃ푦ꢇꢆ푟표ꢃ
휈 ∙ 푚표ꢃꢆꢅ 표ꢈ ꢅ푝ꢆꢇꢁꢈꢁꢇ 푝푟표푑푢ꢇ푡
푖
[ ]
푆푒푙푒푐ꢄꢀ푣ꢀꢄ푦 % =
∙ 100
푚표ꢃꢆꢅ 표ꢈ ꢇ표ꢂ푣ꢆ푟푡ꢆ푑 푔ꢃ푦ꢇꢆ푟표ꢃ
where 휈ꢁ is the stoichiometric factor, amounting to 1 for glycerol, acetol, 1,2-PDO and propanol, 2/3
for ethylene glycol and ethanol, and 1/3 for methanol.
The turnover frequency (TOF) for the HDO pathway was calculated as moles of produced 1,2-PDO
per moles of surface metallic Cu atoms after 8 h (Equation 3). An average number of 1.47×1019 atoms
per m2 Cu surface area was used.[33]
푚표ꢃꢆꢅ 표ꢈ 푝푟표푑푢ꢇꢆ푑 1,2−푃퐷푂
−1
[
]
=
푇푂퐹 ℎ
(3)
푚표ꢃꢆꢅ 표ꢈ ꢅ푢푟ꢈ푎ꢇꢆ 퐶푢 푎푡표푚ꢅ ∙ 8 ℎ
3. Results and discussion
The HDO of glycerol to 1,2-PDO over Cu/ZrO2 proceeds via a two-step dehydration-hydrogenation
reaction pathway forming acetol as dehydration intermediate, which is further hydrogenated to
1,2-PDO with approximately 95% selectivity.[32] The parallel direct C-C bonds hydrogenolysis leading
to ethylene glycol and methanol is the main side reaction with about 5% selectivity. The formation of
1,3-PDO resulting from HDO of the secondary C-O bond has not been observed. Further degradation
products such as propanols and ethanol resulting from overhydrogenolysis are present only in trace
amounts (<2% selectivity). Therefore, only the formation of 1,2-PDO and ethylene glycol are
considered in the following. The simplified reaction scheme for the hydrodeoxygenation and
hydrogenolysis of glycerol is shown in Scheme 1, where 푘 is the overall rate constant of glycerol
consumption, 푘1 is the rate constant of HDO and 푘2 is the rate constant of hydrogenolysis.
In order to assess the impact of reaction parameters like glycerol concentration, hydrogen pressure
and reaction temperature on glycerol HDO as well as to obtain reliable kinetic data, the external

mass transfer limitation has to be excluded. Control experiments were carried out by varying the
stirring speed using 500, 750, and 1000 rpm to investigate the influence of film mass transfer
between the bulk liquid phase and the catalyst surface. It was observed that there is no further
increase of the initial reaction rate at 750 and 1000 rpm, suggesting the absence of external mass
transfer resistance on the overall reaction rate. This is consistent with reports in literature.[29]
Therefore, a stirring speed of 750 rpm was chosen for further detailed kinetic study.
3.1 Effect of the initial glycerol concentration
The impact of glycerol concentration on the hydrodeoxygenation kinetics over Cu/ZrO2 was
investigated by varying the initial glycerol concentration between 2 and 8 wt% with 400 mg of
catalyst at 200 °C and 25 bar H2 pressure after 8 h. As shown in Figure 1, glycerol conversion after 8 h
decreased from 73.1% to 20.4% with increasing initial glycerol concentration. The roughly inverse
relation between the achieved degrees of conversion and the initial glycerol concentration points to
zero-order kinetics with respect to glycerol. Interestingly, the selectivities to 1,2-PDO, ethylene glycol
and acetol were hardly affected by the initial glycerol concentration and remained rather constant
during reaction. 1,2-PDO was found as main product with a selectivity between 91.0% and 94.8%.
Ethylene glycol originating from the parallel hydrogenolysis of C-C bonds was found as main by-
product with a selectivity varying between 7.2% and 4.3%. The selectivity to acetol showed a slight
increase with increasing glycerol concentration but remained below 1% for each concentration
variation experiment, indicating that the hydrogenation of acetol to 1,2-PDO is a fast consecutive
reaction. It has been previously reported that a 4 wt% aqueous acetol was almost completely
hydrogenated to 1,2-PDO (97% yield) in less than 1 h.[32]
Furthermore, the rate of glycerol conversion was studied by taking liquid samples after 1, 3, 5 and
8 h. The concentration of glycerol was found to decrease linearly as a function of reaction time
(Figure 2), suggesting that glycerol HDO does not depend on the applied initial glycerol concentration
indicating zero-order kinetics. Correspondingly, the slopes are almost the same. A reaction order of
zero in glycerol is consistent with some literature reports,[28,34] however, first-order kinetics has
been also reported.[35-37] The glycerol reaction order mainly depends on the employed catalysts,
reaction conditions and H2 supply (external or in situ production by aqueous phase reforming). It has
to be mentioned that the derived zero-order dependence is not due to the lower H2 concentration in
solution compared with the glycerol concentration, as the dehydration is the rate-determining step

and the concentration of dissolved H2 is steadily in excess relating to the concentration of acetol.[32]
Instead, it suggests that the metallic Cu sites are almost fully covered by adsorbed glycerol.[29,34]
Based on the assumption of a Langmuir-Hinshelwood mechanism for glycerol HDO, power-law
kinetics can be applied to the overall glycerol consumption and the formation of the two main
products 1,2-PDO and ethylene glycol. Equation 4 is the specific reaction rate for glycerol
consumption:
r = − dc = k ∙ cglyn1 ∙ pH
(4)
gly
n2
2
dt
where r, k, cgly, pH , n1 and n2 are reaction rate, rate constant, glycerol concentration, H2 pressure,
2
reaction order for glycerol and H2, respectively. Since a zero-order reaction for glycerol conversion
(
)
푛1 = 0 was derived from the concentration variation experiments, the reaction rate depends
exclusively on the H2 partial pressure. Consequently, Equation 4 can be simplified to
r = − dc = k ∙ pH
(5)
gly
n2
2
dt
As the amount of gaseous H2 is in 7-8 fold excess to the stoichiometric demand (i.e., 64 mmol H2 vs
8.7 mmol glycerol), the pressure of H2 inside the reactor can be considered constant. Thus, the rate
of glycerol conversion depends exclusively on the rate constant. The averaged reaction rate of the
overall glycerol consumption at 200 °C and 25 bar H2 pressure for three different initial glycerol
concentrations (Figure 2) is derived as
r = − dc = k = 4.74 ∙ 10−2 ± 1.9 ∙ 10−3 mol ∙ L−1 ∙ h−1 ∙ gcat
(6)
gly
−1
dt
with respect to the mass of the catalyst applied in the kinetic measurements.
3.2 Effect of the hydrogen pressure
Hydrogen pressure variation experiments were carried out in a pressure range of 25 to 35 bar H2 with
an increment of 5 bar at 200 °C for 8 h with 400 mg catalyst precursor and 4 wt% glycerol aqueous
solution. The H2 concentrations in water at 200 °C were estimated to amount to 0.024, 0.034 and
0.047 mol L-1 at 25, 30 and 35 bar H2 partial pressure, respectively.[38] The impact of the H2 pressure
variation on the rate of glycerol HDO is shown in Figure 3. The H2 pressure predominantly affects the
conversion of glycerol. When increasing the H2 pressure from 25 to 35 bar, glycerol conversion after
8 h almost doubled from 31.4% to 57.2%. The selectivity to acetol decreased from 1.0% to 0.4%,

which is caused by the equilibrium shift from acetol to 1,2-PDO at higher H2 pressure. It is worth
mentioning that the H2 pressure has marginal impacts on the selectivity to the main products 1,2-
PDO and ethylene glycol. The selectivity to 1,2-PDO slightly decreased by 1.1%, while the selectivity
to ethylene glycol increased by 2.2% with increasing H2 pressure. This observation indicates that the
direct C-C bond hydrogenolysis leading to ethylene glycol is slightly favoured at higher H2 pressure.
Furthermore, only methanol and 2-propanol were detected as mono-alcoholic by-products
originating from overhydrogenolysis of products and intermediates. These by-products are produced
in very low amounts with approximately 1% selectivity (not shown in detail), suggesting that the
readsorption of the primary HDO products 1,2-PDO and ethylene glycol on the active sites is unlikely
to occur.[32]
Liquid-phase hydrogenation reactions using external H2 supply usually require high H2 pressure,
because the conversion of reactants is limited by the solubility of H2 in the liquid phase. When the H2
pressure is used in the rate expressions Equations 4 and 5, it has to be assumed that the absorption
and diffusion of H2 into the aqueous phase is fast establishing gas-liquid equilibrium. The importance
of using H2 concentration instead of H2 partial pressure for liquid-phase hydrogenation has been
emphasized by the Boudart group.[39] Therefore, the H2 concentration in the aqueous phase instead
of the H2 partial pressure in the reactor is used for the reaction rate calculation in the present study,
and Equation 5 can be expressed as
r = − dc = k ∙ cH
(7)
gly
n2
2
dt
In contrast to the glycerol concentration variation experiments, the variation of the H2 pressure does
not indicate a reaction order of zero with respect to H2 (Figure 4). Instead, the log-log plot of glycerol
conversion vs. H2 concentration in the aqueous phase results in a reaction order of 1.0 for H2. This
result leads to the question of the origin of the positive effect of the H2 concentration on the rate of
glycerol conversion. The dehydration of glycerol to acetol and water was identified as the rate-
determining step, whereas the subsequent hydrogenation of the intermediate acetol to
1,2-propanediol was observed to be a fast reaction. Thus, the surface dehydration of glycerol to
acetol is the slowest step among dissociative glycerol adsorption, dehydration, acetol hydrogenation
and product desorption. Furthermore, the metallic Cu surface sites are assumed to be covered
mostly by adsorbed glycerol, presumably as glycerolate, and marginally by hydrogen atoms resulting
in the observed zero-order dependence on glycerol concentration. The adsorption of glycerol
presumably does not affect H2 adsorption. It has to be mentioned that water is subcritical under the

present reaction conditions, resulting in high levels of H+ and OH-, which are beneficial for the
dehydration of glycerol to acetol.[40]
It has to be considered that the reaction occurs in the aqueous phase, and that also H2O can adsorb
dissociatively on Cu in the investigated temperature range according to Equation 8. Water
dissociation was verified by H2 TPD experiments over Cu/ZrO2 with an onset temperature of 180 °C in
our previous reports.[41,42] In addition, similar metallic Cu surface oxidation by H2O during the
water gas shift reaction over Cu-based catalysts has been proposed based on both kinetic
experiments and DFT calculations.[43-45] Thus, the positive effect of H2 is ascribed to the formation
of empty sites required for dissociative glycerol adsorption. Further studies using operando XPS to
determine the Auger parameter are considered helpful to verify the oxidation state of Cu.
Cusurf + H2O ⇌ Cusurf − Oads + H2
(8)
3.3 Effect of the reaction temperature
Temperature variation experiments over Cu/ZrO2 were conducted to investigate the influence on
glycerol conversion and product selectivities. These kinetic data were used to derive the apparent
activation energies for the HDO and hydrogenolysis routes. As previously reported, metallic Cu is
identified to be the active site for the consecutive dehydration-hydrogenation pathway.[32] In order
to further verify this assumption, the reaction kinetics was additionally studied over pure metallic Cu
obtained by the reduction of commercial CuO nanopowder and compared with the kinetics over
Cu/ZrO2.
3.3.1 Cu/ZrO2 catalyst
The results of the temperature variation experiments using Cu/ZrO2 are summarised in Table 1.
Without catalyst, glycerol was not converted at 200 oC even after 20 h. An increase of the reaction
temperature significantly increased glycerol conversion from 9.9% at 175 °C to 93.0% at 225 °C.
Similar to the pressure variation experiments, the selectivity of the direct C-C bond hydrogenolysis of
glycerol to ethylene glycol marginally increased from 4.5% to 7.2%. Consequently, the glycerol HDO
pathway is favoured over the C-C hydrogenolysis at lower temperature. The selectivity to acetol
gradually increased from 0.4% to 5.4% with increasing temperature, indicating that the consecutive
hydrogenation to 1,2-PDO is limited at higher temperature. This incomplete hydrogenation of acetol
to 1,2-PDO can be attributed to a lower H2 partial pressure at higher reaction temperature at 40 bar
total pressure. The selectivity to 1,2-PDO remained rather constant at around 95% up to 200 °C, and

a further temperature increase to 225 °C led to a slight decrease by 7%. This decrease is caused by
the incomplete hydrogenation of acetol and by the increasing selectivity of ethylene glycol at 225 °C.
The TOF for the HDO pathway increased from 2.5 to 21.5 h-1 with increasing temperature over
Cu/ZrO2. These values are larger compared with literature reports under similar reaction conditions
in the aqueous phase, e.g., 2.1 h-1 for Cu/ZnO at 200 °C and 50 bar H2 pressure[30], 2.6 h-1 for
Cu/Al2O3 at 220 °C and 70 bar H2 pressure[46], pointing to the relatively high activity of Cu/ZrO2.
Vasiliadou et al.[28] reported higher TOF values of 30 and 50 h-1 after 5 h using two types of Cu/SiO2
in n-butanol, which can be ascribed to the employed harsher reaction conditions (240 °C and 80 bar
H2 pressure) and the possible solvent effect. The TOF values for 1,2-PDO formation over Cu powder
are comparable with that over Cu/ZrO2.
3.3.2 Comparison of Cu/ZrO2 with metallic Cu powder
Corresponding to Cu/ZrO2, the conversion of glycerol over metallic Cu powder (Table 1) was also
significantly enhanced from 2.7% to 56.7% with increasing temperature. However, the selectivity to
1,2-PDO was much lower (56 to 63%), whereas the selectivity to ethylene glycol was much higher (30
to 42%). The detailed kinetic results at 225 °C using Cu/ZrO2 and Cu powder are shown in Figures 5a
and 5b, respectively. It has to be pointed out that this specific comparison was conducted using equal
numbers of active sites. Since the specific Cu surface areas of Cu/ZrO2 and metallic Cu powder
determined by N2O frontal chromatography measurements amount to 4.2 and 0.73 m² g-1 (ratio of
5.75), respectively, 0.4 g Cu/ZrO2 and 2.3 g CuO were used for this study accordingly. For both
catalysts the conversion of glycerol increased linearly with increasing reaction time, again pointing to
zero-order kinetics for the overall glycerol consumption. More interestingly, the linear correlation in
case of pure Cu powder further supports the dominant role of metallic Cu as the active site in the
HDO of glycerol. Similar degrees of conversion of 93% (Cu/ZrO2) and 95% (Cu powder) were achieved
after 8 h, but the selectivity to 1,2-PDO and ethylene glycol showed a strong dependence on the
catalyst. Using Cu/ZrO2 resulted in approximately 90% selectivity to 1,2-PDO and 6% selectivity to
ethylene glycol, whereas metallic Cu led to around 60% and 40% selectivity to 1,2-PDO and ethylene
glycol, respectively. This observation indicates that the unsupported metallic Cu powder facilitates
both HDO and the competing C-C bond hydrogenolysis, whereas Cu/ZrO2 mainly favours the selective
HDO to 1,2-PDO.

The shift from HDO to C-C hydrogenolysis may be due to the different sizes of the metallic Cu
particles. The Cu particle size of Cu/ZrO2 is 5 nm, whereas the particle size of the Cu powder was
derived as 42 nm from the corresponding XRD pattern using the Scherrer equation.[32, 40] It is
obvious that the HDO pathway is favoured over smaller Cu particles, whereas C-C bond
hydrogenolysis is significantly enhanced over larger Cu particles. Glycerol HDO to 1,2-PDO requires
the adsorption of one OH-group on the active site. This mode of adsorption is probably less sterically
hindered than the coordinative adsorption of two carbon atoms, which is required to initiate C-C
bond hydrogenolysis. Consequently, the higher HDO selectivity on smaller Cu particles is presumably
attributed to the suppression of C-C hydrogenolysis due to sterical hindrance. A similar correlation
between Cu particle size and 1,2-PDO selectivity was observed for Cu/ZnO catalysts.[47] With
increasing Cu particle size from 26 to 33 nm the selectivity to 1,2-PDO was found to decrease from
78% to 28%. In another study using Cu/MgO catalysts, the selectivity to 1,2-PDO was also found to
decrease by 5%, when the Cu particle size increased from 12 to 30 nm.[48]
Figure 6 shows the 1,2-PDO yields as a function of reaction time at three different temperatures
using the same amount of Cu/ZrO2 (a) and metallic Cu powder (b). Linear correlations between the
yields of 1,2-PDO and reaction time are observed, suggesting a zero-order dependence for the HDO
route with respect to glycerol concentration. The apparent activation energies Ea1 for glycerol HDO
over both catalysts were derived from the Arrhenius plots (Figure 7). The apparent activation
energies of glycerol HDO over Cu/ZrO2 and metallic Cu powder were found to be 106 and
105 kJ mol-1, respectively. The rather similar activation energies further indicate that the sequential
dehydration and hydrogenation (HDO) is catalysed by the same active site on both catalysts, which is
identified as fully adsorbate-covered metallic Cu. In addition, the temperature dependence for C-C
bond hydrogenolysis was analysed over both catalysts. The kinetic parameters for both catalysts and
the two different reaction pathways are summarized in Table 2. Although the apparent activation
energy of glycerol hydrogenolysis over metallic Cu powder (109 kJ mol-1) is larger than that over
Cu/ZrO2 (97 kJ mol-1), the selectivity to ethylene glycol is significantly higher for Cu powder. This is
ascribed to the fact that the pre-exponential factor for metallic Cu powder is roughly two orders of
magnitude larger, thus resulting in a higher rate constant.

4. Conclusions
The reaction kinetics of glycerol hydrodeoxygenation and the competing C-C hydrogenolysis was
studied over a co-precipitated Cu/ZrO2 catalyst under aqueous conditions. The kinetic experiments
were performed by varying the reaction parameters such as temperature (175 – 225°C), H2 partial
pressure (25 – 35 bar) and initial glycerol concentration (2 – 8 wt%). Glycerol conversion was strongly
influenced by temperature and H2 pressure leading up to 93% conversion at 225 °C and 25 bar H2
pressure. The selectivity to 1,2-PDO slightly decreased from 95% to 88% with increasing temperature
and pressure, while the selectivity to ethylene glycol produced from the competing C-C bond
hydrogenolysis increased from 4% to 7%.
These two parallel routes were considered for the kinetic study and a power-law model was applied.
The apparent activation energy for the HDO pathway was determined to be 106 kJ mol-1 over
Cu/ZrO2. For unsupported pure metallic Cu as catalyst, the apparent activation energy was found to
be 105 kJ mol-1 for the HDO route. The similar HDO activation energies over both catalysts indicate
the dominant role of fully adsorbate-covered metallic Cu as the active site for the consecutive
dehydration-hydrogenation of glycerol to 1,2-PDO. The hydrogenolysis of glycerol to ethylene glycol
is found to be structure-sensitive, being favoured for larger Cu particles.
The overall glycerol consumption shows a zero-order dependence on the initial glycerol
concentration and a nearly first-order dependence on H2 concentration. Furthermore, the formation
rate of 1,2-propanediol is also independent of the glycerol concentration. These results imply that
the high selectivity to the HDO product 1,2-PDO is due to the high coverage of glycerol inhibiting
product readsorption. The role of H2 is assumed to be two-fold: it is required for fast hydrogenation
of the formed intermediate acetol and to create empty sites for dissociative glycerol adsorption by
removing adsorbed atomic oxygen originating from water dissociation.
References
[1]
[2]
[3]
Y. Nakagawa, X. Ning, Y. Amada, K. Tomishige, Appl. Catal. A: Gen. 433 (2012) 128-134.
https://doi.org/10.1016/j.apcata.2012.05.009
D.T. Johnson, K.A. Taconi, Environ. Prog. 26 (2007) 338-348.
https://doi.org/10.1002/ep.10225
Y. Nakagawa, K. Tomishige, Catal. Sci. Technol. 1 (2011) 179-190.
https://doi.org/10.1039/C0CY00054J

[4]
A.M. Ruppert, J.D. Meeldijk, B.M.W. Kuipers, B.H. Erne, B.M. Weckhuysen, Chem. Eur. J. 14
(2018) 2016-2024. https://doi.org/10.1002/chem.200701757
[5]
W. Fang, S. Wang, A. Liebens, F. De Campo, H. Xu, W. Shen, M. Pera-Titus, J.-M. Clacens,
Catal. Sci. Technol. 5 (2015) 3980–3990. https://doi.org/10.1039/C5CY00534E
[6]
J.-M. Clacens, Y. Pouilloux, J. Barrault, Appl. Catal. A: Gen. 227 (2002) 181-190.
https://doi.org/10.1016/S0926-860X(01)00920-6
[7]
M. Trejda, K. Stawicka, M. Ziolek, Appl. Catal. B: Environ. 103 (2011) 404-412.
https://doi.org/10.1016/j.apcatb.2011.02.003
[8]
W.D. Bossaert, D.E. De Vos, W.M. Van Rhijn, J. Bullen, P.J. Grobet, P.A. Jacobs, J. Catal. 182
(1999) 156-164. https://doi.org/10.1006/jcat.1998.2353
[9]
F. Cai, D. Pan, J.J. Ibrahim, J. Zhang, G. Xiao, Appl. Catal. A: Gen. 564 (2018) 172-182.
https://doi.org/10.1016/j.apcata.2018.07.029
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
E.S. Vasiliadou, A.A. Lemonidou, WIREs Energy Environ. 4 (2015) 486-520.
https://doi.org/10.1002/wene.159
Y. Wang, J. Zhou, X. Guo, RSC Adv. 5 (2015) 74611-74628.
https://doi.org/10.1039/C5RA11957J
A. Corma, G.W. Huber, L. Sauvanauda, P.O’Connor, J. Catal., 257 (2008) 163-171.
https://doi.org/10.1016/j.jcat.2008.04.016
E. Tsukuda, S. Sato, R. Takahashi, T. Sodesawa, Catal. Comm. 8 (2007) 1349-1353.
https://doi.org/10.1016/j.catcom.2006.12.006
N.P. Rajan, G. Srinivasa Rao, V. Pavankumar, K.V.R. Chary, Catal. Sci. Technol. 4 (2014) 81-92.
https://doi.org/10.1039/C3CY00430A
Y. Shinmi, S. Koso, T. Kubota, Y. Nakagawa, K. Tomishige, Appl. Catal. B: Environ. 94 (2010)
318-326. https://doi.org/10.1016/j.apcatb.2009.11.021
M. Checa, F. Auneau, J. Hidalgo-Carrillo, A. Marinas, J.M. Marinas, C. Pinel, F.J. Urbano, Catal.
Today 196 (2012) 91-100. https://doi.org/10.1016/j.cattod.2012.02.036
E.S. Vasiliadou, A.A. Lemonidou, Org. Process Res. Dev. 15 (2011) 925-931.
https://doi.org/10.1021/op2000173
D. Sun, Y. Yamada, S. Sato, W. Ueda, Appl. Catal. B: Environ. 193 (2016) 75-92.
https://doi.org/10.1016/j.apcatb.2016.04.013
W. Yu, J. Xu, H. Ma, C. Chen, J. Zhao, H. Miao, Q. Song, Catal. Comm. 11 (2010) 493-497.
https://doi.org/10.1016/j.catcom.2009.12.009
X. Guo, Y. Li, R. Shi, Q Liu, E. Zhan, W. Shen, Appl. Catal. A: Gen. 371 (2009) 108-113.
https://doi.org/10.1016/j.apcata.2009.09.037

[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
I. Gandarias, P.L. Arias, J. Requies, M.B. Güemez, J.L.G. Fierro, Appl. Catal. B: Environ. 97
(2010) 248-256. https://doi.org/10.1016/j.apcatb.2010.04.008
M. Balaraju, V. Rekha, P.S. Sai Prasad, B.L.A. Prabhavathi Devi, N. Lingaiah, Appl. Catal. A:
Gen. 354 (2009) 82-87. https://doi.org/10.1016/j.apcata.2008.11.010
F. Cai, W. Zhu, G. Xiao, Catal. Sci. Technol. 6 (2016) 4889–4900.
https://doi.org/10.1039/C6CY00085A
M. Balaraju, V. Rekha, P.S. Sai Prasad, R.B.N. Prasad, N. Lingaiah, Catal. Lett. 126 (2008) 119–
124. https://doi.org/10.1007/s10562-008-9590-6
D. Durán-Martín, M. Ojeda, M. López Granados, J.L.G. Fierro, R. Mariscal, Catal. Today 210
(2013) 98-105. https://doi.org/10.1016/j.cattod.2012.11.013
Z. Xiao, J. Xiu, X. Wang, B. Zhang, C.T. Williams, D. Su, C. Liang, Catal. Sci. Technol. 3 (2013)
1108–1115. https://doi.org/10.1039/C2CY20788E
R.V. Sharma, P. Kumar, A.K. Dalai, Appl. Catal. A: Gen. 477 (2014) 147-156.
https://doi.org/10.1016/j.apcata.2014.03.007
E.S. Vasiliadou, T.M. Eggenhuisen, P. Munnik, P.E. de Jongh, K.P. de Jong, A.A. Lemonidou,
Appl. Catal. B: Environ. 145 (2014) 108-119. https://doi.org/10.1016/j.apcatb.2012.12.044
E.S. Vasiliadou, A.A. Lemonidou, Chem. Eng. J. 231 (2013) 103-112.
https://doi.org/10.1016/j.cej.2013.06.096
A. Bienholz, F. Schwab, P. Claus, Green Chem. 12 (2010) 290-295.
https://doi.org/10.1039/B914523K
C. Wang, H. Jiang, C. Chen, R. Chen, W. Xing, Chem. Eng. J., 264 (2015) 344-350.
https://doi.org/10.1016/j.cej.2014.11.113
T. Gabrysch, B. Peng, S. Bunea, G. Dyker, M. Muhler, ChemCatChem 10 (2018) 1344-1350.
https://doi.org/10.1002/cctc.201701748
O.Hinrichsen, T. Genger, M. Muhler, Chem. Eng. Technol. 23 (2000) 956–959.
https://doi.org/10.1002/1521-4125(200011)23:11<956::AID-CEAT956>3.0.CO;2-L
Y. Amada, Y. Shinmi, S. Koso, T. Kubota, Y. Nakagawa, K. Tomishige, Appl. Catal. B: Environ.
105 (2011) 117-127. https://doi.org/10.1016/j.apcatb.2011.04.001
D.G. Lahr, B.H. Shanks, Ind. Eng. Chem. Res. 42 (2003) 5467−5472.
https://doi.org/10.1021/ie030468l
X. Jin, B. Subramaniam, R.V. Chaudhari, P.S. Thapa, AIChE J. 62 (2016) 1162-1173.
https://doi.org/10.1002/aic.15114
N.N. Pandhare, S.M. Pudi, S. Mondal, K. Pareta, M. Kumar, P. Biswas, Ind. Eng. Chem. Res. 57
(2018) 101-110. https://doi.org/10.1021/acs.iecr.7b03684

[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
H.A. Pray, C.E. Schweickert, B.H. Minnich, Ind. Eng. Chem. 44 (1952) 1146-1151.
https://doi.org/10.1021/ie50509a058
E.E. Gonzo, M. Boudart, J. Catal. 52 (1978) 462–471. https://doi.org/10.1016/0021-
9517(78)90352-4
S. S. Toor, L. Rosendahl, A. Rudolf, Energy 36 (2011) 2328-2342.
https://10.1016/j.energy.2011.03.013
J. Schittkowski, K. Tölle, S. Anke, S. Stürmer, M. Muhler, J. Catal. 352 (2017) 120-129.
https://doi.org/10.1016/j.jcat.2017.05.009
J. Schittkowski, D. Buesen, K. Tölle, M. Muhler, Catal. Lett. 146 (2016) 1011-1017.
https://doi.org/10.1007/s10562-016-1712-y
N.A. Koryabkina, A.A. Phatak, W.F. Ruettinger, R.J. Farrauto, F.H. Ribeiro, J. Catal. 217 (2003)
233–239. https://doi.org/10.1016/S0021-9517(03)00050-2
C.-S. Chen, T.-W. Lai, C.-C. Chen, J. Catal. 273 (2010) 18-28.
https://doi.org/10.1016/j.jcat.2010.04.019
C.-H. Lin, C.-L. Chen, J.-H. Wang, J. Phys. Chem. C. 115 (2011) 18582-18588.
https://doi.org/10.1021/jp2034467
R.B. Mane, A.M. Hengne, A.A. Ghalwadkar, S. Vijayanand, P.H.Mohite, H.S.Potdar, C.V. Rode,
Catal. Lett. 135 (2010) 141-147. https://doi.org/10.1007/s10562-010-0276-5
[47]
[48]
S. Wang, H. Liu, Catal. Lett. 117 (2007) 62-67. https://doi.org/10.1007/s10562-007-9106-9
M. Balaraju, K. Jagadeeswaraiah, P.S. Sai Prasad, N. Lingaiah, Catal. Sci. Technol. 2 (2012)
1967-1976. https://doi.org/10.1039/C2CY20059G

Figure 1. Effect of initial glycerol concentration on the degree of conversion after 8 h reaction time and liquid
product selectivities. Reaction conditions: 20 mL of 2-8 wt% aqueous glycerol solution, 400 mg of 18 wt%
CuO/ZrO2 (pre-reduced), 200 °C, 25 bar H2 pressure, 750 rpm stirring speed.

Figure 2. Glycerol concentration as a function of reaction time with varying initial glycerol concentrations (2-
8 wt%). Reaction conditions: 20 mL of 2-8 wt% aqueous glycerol solution, 400 mg of 18 wt% CuO/ZrO2 (pre-
reduced), 25 bar H2 pressure, 200 °C, 750 rpm stirring speed.

Figure 3. Effect of H2 pressure on glycerol conversion and liquid product selectivities. Reaction conditions:
20 mL of 4 wt% aqueous glycerol solution, 400 mg of 18 wt% CuO/ZrO2 (pre-reduced), 200 °C, 750 rpm stirring
speed and 8 h.

Figure 4. Log-log plot of glycerol conversion with hydrogen concentration in the aqueous phase at varying H2
pressure (25 to 35 bar); liquid samples were taken after 3, 5, and 8 h reaction time. Reaction conditions: 20 mL
of 4 wt% aqueous glycerol solution, 400 mg of 18 wt% CuO/ZrO2 (pre-reduced), 200 °C and 750 rpm stirring
speed.

Figure 5. Glycerol conversion and liquid product selectivities as a function of time over a) 18 wt% CuO/ZrO2
(pre-reduced) and b) metallic Cu powder. Reaction conditions: 20 mL of 4 wt% aqueous glycerol solution,
400 mg of 18 wt% CuO/ZrO2 and 2.3 g of CuO (both reduced prior to reaction), 225 °C, 750 rpm stirring speed
and 8 h.

Figure 6. Plots of the 1,2-PDO yield against reaction time for varying reaction temperatures over a) 18 wt%
CuO/ZrO2 and b) metallic Cu powder. Reaction conditions: 20 mL of 4 wt% aqueous glycerol solution, 400 mg of
unreduced precursor (reduced prior to the reaction), 25 bar H2 pressure and 750 rpm stirring speed.

Figure 7. Arrhenius plots for the selective HDO of glycerol to 1,2-PDO over Cu/ZrO2 (green line) and metallic Cu
powder (orange line). Reaction conditions: 20 mL of 4 wt% aqueous glycerol solution, 400 mg of unreduced
precursor (reduced prior to the reaction), 25 bar H2 pressure and 750 rpm stirring speed.

Scheme 1: Simplified reaction pathways of hydrodeoxygenation and hydrogenolysis of glycerol over Cu-based
catalysts.

Table 1: Effect of reaction temperature on glycerol conversion, TOF for HDO and liquid product selectivities
after 8 h over Cu/ZrO2 and metallic Cu powder. Reaction conditions: 20 mL of 4 wt% aqueous glycerol solution,
400 mg of unreduced catalyst (reduced prior to the reaction), 40 bar total pressure, 750 rpm stirring speed.
Cu/ZrO2
Selectivity / %
Cu powder
Selectivity / %
T /
°C
Conv. TOF*
Conv.
/ %
TOF*
/ h-1
Ethylene
glycol
1,2-
Ethylene
glycol
/ %
/ h-1
Acetol 1,2-PDO
Acetol
PDO
2.5
7.8
21.5
2.5
10.1
41.7
175 9.9
200 31.4
225 93.0
0.4
1.0
5.4
94.2
93.9
87.0
4.5
4.0
7.2
2.7
12.2
56.7
1.4
1.0
2.3
63.0
56.6
56.0
30.6
41.6
29.9
* The TOF for the HDO pathway was calculated after 8 h based on the molar amount of exposed metallic
copper atoms

Table 2: Derived kinetic parameters for glycerol HDO and hydrogenolysis over Cu/ZrO2 and metallic Cu powder.
Hydrodeoxygenation
A1
Hydrogenolysis
A2
Catalyst
Ea1 /
Ea2 /
kJ mol-1
106 ± 2.1
105 ± 2.1
kJ mol-1
97 ± 2.9
109 ±2.2
Cu/ZrO2
Cu
2.3 x 1010
4.0 x 109
9.1 x 107
6.3 x 109