Hydrogen and chemicals from alcohols through electrochemical reforming by Pd-CeO2/C electrocatalyst

Article abstract of DOI:10.1016/j.ica.2021.120245

The development of low-cost and sustainable hydrogen production is of primary importance for a future transition to sustainable energy. In this work, the selective and simultaneous production of pure hydrogen and chemicals from renewable alcohols is achieved using an anion exchange membrane electrolysis cell (electrochemical reforming) employing a nanostructured Pd-CeO₂/C anode. The catalyst exhibits high activity for alcohol electrooxidation (e.g. 474 mA cm^?2 with EtOH at 60 °C) and the electrolysis cell produces high volumes of hydrogen (1.73 l min^?1 m^?2) at low electrical energy input (E_cost = 6 kWh kg_H2^?1 with formate as substrate). A complete analysis of the alcohol oxidation products from several alcohols (methanol, ethanol, 1,2-propandiol, ethylene glycol, glycerol and 1,4-butanediol) shows high selectivity in the formation of valuable chemicals such as acetate from ethanol (100%) and lactate from 1,2-propandiol (84%). Importantly for industrial application, in batch experiments the Pd-CeO₂/C catalyst achieves conversion efficiencies above 80% for both formate and methanol, and 95% for ethanol.

Full text of DOI:10.1016/j.ica.2021.120245

Inorganica Chimica Acta 518 (2021) 120245
Contents lists available at ScienceDirect
Inorganica Chimica Acta
journal homepage: www.elsevier.com/locate/ica
Research paper
Hydrogen and chemicals from alcohols through electrochemical reforming
by Pd-CeO2/C electrocatalyst
Marco Bellinia, Maria V. Pagliaroa, Andrea Marchionnia, Jonathan Filippia,
,
Hamish A. Millera *, Manuela Bevilacquaa, Alessandro Lavacchia, Werner Oberhausera,
,
,
b
,
,*
Jafar Mahmoudiana, Massimo Innocentia c, Paolo Fornasieroa *, Francesco Vizzaa
a Istituto di Chimica dei Composti Organometallici (CNR-ICCOM), Via Madonna del Piano 10, 50019, Sesto Fiorentino, Italy
b Department of Chemical and Pharmaceutical Sciences, University of Trieste, Via L. Giorgieri 1, 34127, Trieste, Italy
c
`
Dipartimento di Chimica ’Ugo Schiff’, Universita degli Studi Firenze, Via della Lastruccia, 3-13, 50019 Sesto Fiorentino, Italy
A R T I C L E I N F O
A B S T R A C T
Keywords:
The development of low-cost and sustainable hydrogen production is of primary importance for a future tran-
sition to sustainable energy. In this work, the selective and simultaneous production of pure hydrogen and
chemicals from renewable alcohols is achieved using an anion exchange membrane electrolysis cell (electro-
chemical reforming) employing a nanostructured Pd-CeO2/C anode. The catalyst exhibits high activity for
alcohol electrooxidation (e.g. 474 mA cmꢀ 2 with EtOH at 60 ◦C) and the electrolysis cell produces high volumes
Electrocatalysis
Hydrogen
Alcohol electroreforming
Renewable feedstock
Valuable chemicals
of hydrogen (1.73 l minꢀ 1 mꢀ 2) at low electrical energy input (Ecost = 6 kWh kgH ꢀ 1 with formate as substrate). A
2
complete analysis of the alcohol oxidation products from several alcohols (methanol, ethanol, 1,2-propandiol,
ethylene glycol, glycerol and 1,4-butanediol) shows high selectivity in the formation of valuable chemicals
such as acetate from ethanol (100%) and lactate from 1,2-propandiol (84%). Importantly for industrial appli-
cation, in batch experiments the Pd-CeO2/C catalyst achieves conversion efficiencies above 80% for both formate
and methanol, and 95% for ethanol.
1. Introduction
electric energy input, up to 50 kWh kgH ꢀ 1, indeed they need precious
2
metal based catalyst at high loadings and an expensive polymer ex-
Modern human society is reliant on fossil fuels. Since the oil crises of
the early 1970s it has become clear this system is not sustainable. This
has led to a huge effort to find alternative energy resources based on
renewables [1–3]. Molecular hydrogen has a key role as high-quality
energy carrier for replacing fossil fuels [4]. Despite intensive research,
hydrogen as energy carrier has not yet reached worldwide utilization
due to several issues. Firstly, no market scale technology has been
developed yet for H2 production, storage and conversion to electricity
matching the necessary sustainable and low-cost criteria [5–9]. A well-
known and consolidate process for hydrogen production is the elec-
trolysis of water; it is the only route able to convert renewable energy
sources (e.g., photovoltaic, wind, biomass, geothermal) into 99.999%
pure hydrogen [10,11]. Polymer Exchange Membrane Electrolysers
(PEM-Electrolysers) are the most performing devices ever reported,
nevertheless they don’t have a significant commercial impact owing to
high operative costs [5]. The best PEM electrolysers still require high
change membrane [12]. The anodic Oxygen Evolution Reaction (OER)
contributes mostly to the high PEM-electrolyser energy input. Replacing
it with the more efficient oxidation of a sacrificial molecule, e.g. an
alcohol [13–15] or ammonia [16,17], is a strategy to lower the high
electrolysis energy cost. “Electrochemical Reforming” or “electro-
reforming” matches the anodic oxidation of aliphatic alcohols with the
cathodic water reduction to hydrogen and the resulting electrolysis
process occurs at cell potentials lower than 1 V, dropping the energy
ꢀ 1
input to c.a. 20 kWh kgH
[18–21]. A scheme of anion exchange
2
membrane based electroreforming is reported in Fig. 1. The strong
alkaline environment (pH greater than 12) allows the partial oxidation
of alcohols to valuable chemicals, such as carboxylic compounds, which
are raw materials for fine chemicals industry. So alcohol electrore-
formers are the only electrochemical devices able to store electric energy
in molecular hydrogen and convert a biomass derived alcohol into high
added value chemicals [19]. In addition, alcohol oxidation reactions
* Corresponding authors.
E-mail addresses: hamish.miller@iccom.cnr.it (H.A. Miller), pfornasiero@units.it (P. Fornasiero), francesco.vizza@iccom.cnr.it (F. Vizza).
https://doi.org/10.1016/j.ica.2021.120245
Received 2 December 2020; Received in revised form 30 December 2020; Accepted 30 December 2020
Available online 12 January 2021
0020-1693/© 2021 Elsevier B.V. All rights reserved.

M. Bellini et al.
Inorganica Chimica Acta 518 (2021) 120245
Fig. 1. A simplified scheme of a state of the art alkaline electroreforming cell.
have faster kinetics in alkaline media, improving the electroreformer
efficiency [22]. Several renewable source derived alcohols have been
investigated as fuel for alkaline electroreforming cells, such as meth-
anol, ethanol, 1,2-propandiol, ethylene glycol, glycerol, 1,4-butanediol
[23–26]. Formate is an attractive sacrificial molecule as well, since it
is a derivative of formic acid, which can be produced from renewable
sources such as the electrochemical reduction or catalytic hydrogena-
tion of CO2. To make formic acid (and formate) a sustainable fuel, the
electric energy or hydrogen employed in such processes must be derived
from neutral sources [27]. Formate salts are more appealing than vol-
atile formic acid solutions because as solids they can be easily stored,
transported and safely handled and fuels can be obtained just mixing the
salts with water [28].
2. Experimental
2.1. Materials and methods
2.1.1. Synthesis of Pd-CeO2/C
The anode catalyst Pd-CeO2/C was synthesized by using a two step
method as described previously [38]. In brief, the C-CeO2 support was
prepared suspending 4 g of carbon black (Vulcan XC-72, Cabot corp.) in
250 ml of distilled water. 10.1 g of Ce(NO3)3⋅6H2O were dissolved in the
suspension and the resulting solution was stirred for 60 min and further
homogenized by ultrasonic treatment (30 min, 59 Hz, 100 W). The pH
was adjusted to 12 adding dropwise a 0.45 M KOH aqueous solution and
subsequently the suspension was stirred for 2 h. The solid product was
separated by filtration and washed several times with distilled water
until neutral pH was obtained. The product was dried at 65 ◦C and then
heated under air in a tube furnace at 250 ◦C for 2 h. The furnace was
cooled down to room temperature under Ar atmosphere.
Palladium nanoparticles supported on carbon black (Pd/C) and their
derivatives represent the state of the art anodic catalysts for alcohols and
formate oxidation in alkaline electroreforming [29,30]. In previous
papers [31,32] we have investigated palladium nanoparticles on a ceria
decorated carbon black support (50% wt C and 50% wt CeO2) obtaining
a high efficiency anode, Pd-CeO2/C, for low aliphatic alcohols and
formate oxidation in alkaline direct fuel cells. Cerium (IV) oxide is a
cheap and not toxic material with a relative high natural availability.
Moreover, CeO2 redox properties makes this material suitable for several
catalytic applications, spanning from electrocatalysis (e.g. fuel cells,
water electrolysis and CO2 reduction to fuels) to homogeneous catalysis
(e.g for water gas shift reaction) [33–35]. Unfortunately, ceria is a semi-
conductive material, so for electrocatalytic applications it is often
combined with a conductive support, such as carbon based materials
(carbon black, carbon nanotubes or graphene) [33]. CeO2 can easily
change the oxidation state switching between Ce3+ and Ce4+ and acting
as an oxygen buffer in the catalytic material. In particular, in palladium-
ceria electrocatalysts for alcohol oxidation, CeO2 promotes the OHꢀ ions
spillover to the metal nanoparticles [35]. So in the Pd-CeO2/C catalyst,
Pd(I)-OHads formation happens at lower potentials compared to Pd/C
analogue catalysts, enhancing the electrocatalytic activity [36]. In a
similar manner, ceria enhances Pd activity in formate electrooxidation
[27]. Most of the technology developed for direct alcohol and formate
fuel cells can be transferred to electroreforming [15]. In this article, the
Pd-CeO2/C anode is assembled in a complete electroreformer, together
with a commercial AEM membrane (Tokuyama A201) and a commercial
Pt/C cathode. The cell was fed with formate and several alcohols,
(methanol, ethanol, ethylene glycol, 1,2-propandiol, 1,4-butanediol and
glycerol), demonstrating the high versatility of Pd-CeO2/C catalyst
given that the availability of biomass derived alcohols is strongly
dependent on seasonality and on geographic areas of production [37].
The electroreformer reported here is able to produce a number of
important chemicals including acetate, lactate, succinate and glycolate.
Palladium nanoparticles were so grown onto C-CeO2 support. 4 g of
C-CeO2 were suspended in water (500 ml) by a vigorous magnetic stir-
ring (30 min) and a subsequent ultrasonic treatment (20 min, 59 Hz,
100 W). Then a 60 ml aqueous solution of K2PdCl4 (1.38 g) was added
dropwise (approximately 1 h) under vigorous stirring. Palladium salt
was reduced by adding dropwise a 2.5 M KOH (8.4 ml) aqueous solution
and 50 ml of ethanol (99.9%, Sigma-Aldrich-Merck); the resulting
◦
mixture was heated at 80 C for 60 min. The so obtained Pd-CeO2/C
catalyst was filtered off, washed several times with distilled water to
◦
neutrality, and finally dried under vacuum at 65 C. The yield of Pd-
CeO2/C was of 97%.
2.1.2. Electron microscopy
Transmission electron microscopy (TEM) was performed on a Philips
CM12 microscope using an accelerating voltage of 100 kV. Samples were
prepared by suspending the catalyst in isopropanol applying ultrasounds
for 20 min (59 Hz, 100 W). The suspension was then dropped onto the
carbon coated copper TEM grids and dried under air.
2.1.3. Membrane electrode assemblies
Membrane electrode assemblies (MEAs) were prepared using the Pd-
CeO2/C anode electrocatalyst, a commercial anion exchange membrane,
and a carbon cloth cathode containing a commercial 40 wt% Pt/C
catalyst (Aldrich). The anion exchange solid polymer membrane used is
a Tokuyama A201 obtained from Tokuyama Corp. (Japan). Prior to use,
membrane counter ions were exchanged with OHꢀ groups by soaking
the membrane in 1 M KOH for 24 h. The membrane was then washed
thoroughly in deionized water. The cathode electrode was prepared by
spreading a catalyst ink onto a carbon cloth W1S1005 (CeTech Co., Ltd.)
gas diffusion layer, with a Meyer rod (n◦150) in order to obtain a 0.4 mg
cmꢀ 2Pt loading. The cathodic ink was prepared in a 5 ml high density
polyethylene vial, mixing 200 mg of the commercial Pt (40 wt%)/C in
2

M. Bellini et al.
Inorganica Chimica Acta 518 (2021) 120245
Fig. 2. TEM micrographs of Pd-CeO2/C. Scale bar is 100 nm.
450 mg of distilled water, 790 mg of 1-propanol, and 1.56 g of the
ionomer Nafion (5 wt% in 2-propanol). The mixture was suspended by
ultrasound treatment. The anode catalyst ink was prepared by mixing
the Pd-CeO2/C catalyst (40 mg) with water (400 mg) and 40 mg of a 10
wt% PTFE aqueous suspension (10% PTFE final loading of the dried
electrode). The resulting thick paste was spread evenly onto a 4 cm2
nickel foam support (Heze Tianyu Technology Development Co, China)
to obtain a Pd loading of ca. 1 mg cmꢀ 2. The active electroreformer cell
was purchased from Scribner-Associates (USA) (25 cm2 fuel cell fixture).
The MEA were fabricated by mechanically pressing the anode, the Pt/C
cathode and the membrane within the cell hardware (6 Nm torque).
2.1.4. Electrolysis cell testing
The fuel solution used was 30 ml of a 2 M KOH and 2 M of alcohol or
formate aqueous solution. Potassium formate, methanol, ethanol, 1,2-
propandiol, ethylene glycol, glycerol or 1,4-butanediol purchased
from Sigma-Aldrich-Merck were tested. The cell anode was fed recir-
culating the fuel in the anode compartment by means of a peristaltic
pump with a 1 ml minꢀ 1 flow rate.
Voltage scans and galvanostatic curves were acquired using an
ARBIN BT-2000 5A-4 channels potentiostat/galvanostat. Chro-
nopotentiometry experiments were performed by applying a constant
electrolysis current of 30 mA cmꢀ 2 until the cell voltage reached the
value of 1.0 V.
The fuel solutions were recovered after each experiment and were
quantitatively and qualitatively analyzed by 13C{1H} NMR spectroscopy
and HPLC. NMR spectra were acquired with a with a Bruker Avance DRX
400 spectrometer. Chemical shifts (δ) are reported in ppm relative to
TMS (1H and 13C NMR spectra). Deuterated solvents (Sigma-Aldrich)
used for NMR measurements were dried with activated molecular sieves;
1,4-dioxane was used as internal standard for product quantification. A
UFLC Shimadzu chromatograph equipped with refraction index detector
(RID) was used; the column is a GRACE-Alltech OA-1000 organic acid
(300 mm × 6.5 mm), thermostated at 65 ◦C. The eluent is 0.01 N H2SO4,
and the eluent flow is 0.4 ml minꢀ 1
.
The hydrogen generated during the electroreforming was recorded
by a Bronkhorst EL-FLOW mass flow meter model F-101C-002-AGD-11-
V with a maximum H2 flow rate of 3 ml minꢀ 1
.
3. Results and discussions
The Pd-CeO2/C catalyst was synthesized following a well-established
procedure [38] and the resulting catalyst morphology was characterized
by TEM microscopy. Fig. 2 shows a TEM micrograph of the catalyst at
low magnification (Fig. 2A) and high magnification (Fig. 2B). CeO2
nanoparticle clusters are heterogeneously distributed over the carbon
support. Pd NPs are mixed intimately with the CeO2 clusters.
Fig. 3. Scan voltage curves of the electroreformer fed with an aqueous 2 M
KOH and 2 M formate (A) or 2 M methanol (B) or 2 M ethanol (C) solutions; 10
mV sꢀ 1 scan rate. Each cell was subjected to three consecutive experiments:
cycle 1 (blue curve), cycle 2 (red curve) and cycle 3 (green curve). A fresh fuel
batch was used for any experiment.
A comprehensive STEM-EDX and XAS analysis of this catalyst
described in previous publications demonstrates a preferential Pd
deposition in the support regions where ceria accumulates during the
3

M. Bellini et al.
Inorganica Chimica Acta 518 (2021) 120245
Table 1
Catalytic data of electroreforming with Pd-CeO2/C at the anode at 60 ◦C. 1,2-P = 1,2-propandiol, EG = ethylene glycol, Gly = glycerol, 1,4-B = 1,4-butanediol.
Entry
Fuel
Cycle
Jmax @ 1 V (mA cmꢀ 2
)
Eonset
H2
Energy Consumption
Conversion
(V)
(mmol)
(kWh kgꢀ 1H2)
(%)
1
HCOOꢀ
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
228
414
424
422
437
415
474
448
411
529
424
307
249
312
446
196
496
425
495
434
255
0.05
0.10
0.10
0.36
0.43
0.43
0.35
0.35
0.40
0.41
0.39
0.54
0.41
0.45
0.44
0.48
0.53
0.55
0.25
0.34
0.58
41
5.1
69
83
80
84
78
80
95
95
97
75
79
78
70
70
72
49
53
55
64
67
70
50
6.1
48
6.7
2
3
4
5
6
7
MeOH
EtOH
1,2-P
EG
88
15.6
17.1
17.6
11.7
14.3
15.2
17.5
17.6
18.1
16.3
16.9
17.4
15.2
15.9
15.7
12.8
18.2
21.4
83
85
65
70
70
107
110
104
91
97
99
Gly
89
94
94
1,4-B
99
103
105
synthesis. This architecture generates an intimate contact between ceria
and palladium nanoparticles promoting the hydroxide ions spillover to
Pd NPs and so improving the catalyst activity toward alcohol electro-
oxidation reaction as previously reported in literature [32,36].
60 ◦C. Higher temperatures were not explored to avoid membrane
damage. The cell activity was investigated by means of scan voltage
experiments from 0 to 1.2 V. Fig. 3 shows the curves acquired for the
electroreformers fed with aqueous 2 M KOH and 2 M HCOOꢀ or MeOH
or EtOH. The maximum current density of 424 mA cmꢀ 2 (Fig. 3A and
table 1, entry 1), of 437 mA cmꢀ 2 (Fig. 3B and table 1, entry 2) and of
474 mA cmꢀ 2 (Fig. 3C and table 1, entry 3) was recorded respectively for
formate, methanol and ethanol based feeding solution at a cell voltage of
1 V. All the tested electroreformers performance was stable over the
The Pd-CeO2/C catalyst applied to a nickel foam with PTFE was
assembled in the electroreforming cell. The device was equipped with a
commercial Pt/C cathode spread onto a carbon cloth gas diffusion layer
and the commercial Tokuyama A201 anion exchange membrane. Elec-
troreforming experiments were performed at the cell temperature of
Fig. 4. Scan voltage characterization of the electroreformers fed with an aqueous 2 M KOH and 2 M 1,2-propandiol (A) or 2 M ethylene glycol (B) or 2 M 1,4-butan-
diol (C) or 2 M glycerol (D); 10 mV sꢀ 1 scan rate. Each cell was subjected to three consecutive experiments: cycle 1 (blue curve), cycle 2 (red curve) and cycle 3 (green
curve). A fresh fuel batch was used for any experiment.
4

M. Bellini et al.
Inorganica Chimica Acta 518 (2021) 120245
[36]. Furthermore, formate oxidation is an easier reaction respect al-
cohols and polyols oxidation [27,30], so the ceria booster effect is more
efficient and leads to a negligible overpotential reaction [22,30].
Despite the higher onset potential for MeOH and EtOH oxidation in
the electroreformer, results are in line with the activity of ceria free cells
reported in literature based on palladium or other precious metal anodes
such as Rh/C [18,21].
The Pd-CeO2/C based electroreformer is able to produce hydrogen
and chemicals not only by the oxidation of small molecules such as
EtOH, MeOH and HCOOꢀ , but also when it is fed with biomass-derived
polyols, such as diols or glycerol. Fig. 4 summarizes the scan voltage
curves acquired for 1,2-propandiol (Fig. 4A), ethylene glycol (Fig. 4B),
1,4-butanediol (Fig. 4C) and glycerol (Fig. 4D) fed cell (2 M KOH 2 M
polyol aqueous fuel) and the electroreformer stability was again inves-
tigated by repeating three times each experiment with a batch of fresh
fuel. The performance with the C3 and C4 diols dropped over the three
consecutive batches during the three repeated experiments, for example
for 1,2-propandiol the maximum current density recorded at 1 V drops
from 529 mA cmꢀ 2 of the first cycle to 307 mA cmꢀ 2 of the third one
(Fig. 4A and table 1, entry 4). Contemporaneously the onset potential
increases from 0.41 V to 0.54 V. In a similar manner, the cell fed with
1,4-butanediol has a current drop from 495 to 255 mA cmꢀ 2 and an
onset potential increase from 0.25 to 0.58 V (Fig. 5C and table 1, entry
7). By comparison, ethylene glycol ad glycerol fed electroreformers
show a reverse trend, where the current density at 1 V increases from
249 to 446 mA cmꢀ 2 for ethylene glycol oxidation (Fig. 4B and table 1,
entry 5) and from 196 to 425 mA cmꢀ 2 for glycerol oxidation (Fig. 4D
and table 1, entry 6).
Constant current experiments with fixed volume fuel batches were
undertaken to study the selectivity of oxidation for each alcohol as well
as the stability over various batches. A constant current load of 120 mA
(30 mA cmꢀ 2) was applied to the cell and the voltage was recorded until
the cut off potential of 1 V was reached. A flow meter was used to
measure the hydrogen produced at the cathode electrode. Each galva-
nostatic experiment, named also potentiodynamic curve, was repeated
three times for each cell, using a fresh fuel batch each time.
Fig. 5 shows the potentiodynamic curves for the 2 M KOH 2 M
formate (Fig. 5A), 2 M methanol (Fig. 5B) and 2 M ethanol (Fig. 5C)
solutions, while In Fig. 6 the experiments performed on the cells fed with
polyols: 1,2-propandiol (Fig. 6A), ethylene glycol (Fig. 6B), 1,4-butane-
diol (Fig. 6C) and glycerol (Fig. 6D) are shown. The cell voltage recorded
during electrolysis increases over time since the fuel is constantly con-
verted into the products. Towards the end of each curve the potential
rises sharply to 1 V. This occurs as the fuel concentration drops to a low
level. The conversion efficiency was calculated from HPLC analysis of
the fuel solution after each test. All the cells evolve c.a. 1.73 l minꢀ 1 mꢀ 2
Fig. 5. Electroreforming of aqueous 2 M KOH 2 M formate (A) or 2 M methanol
(B) or 2 M ethanol (C) at 30 mA cmꢀ 2 current load. Each cell was subjected to
three consecutive experiments using a fresh fuel solution for every galvano-
static: cycle 1 (blue curve), cycle 2 (red curve) and cycle 3 (green curve).
of pure hydrogen with a Fradaic efficiency of c.a. 86–90%. The electrical
ꢀ 1
energy consumed in each test is expressed in kWh kgH
and was
evaluated for each experiment by integrating the instantane2ous charging
power over the experimental duration time. A summary of the data is
shown in Table 1.
three batch experiments indicating no loss in performance during elec-
trocatalysis from batch to batch.
In the case of HCOOꢀ fed cell, the first experiment shows the low Pd-
CeO2/C activity of 228 mA cmꢀ 2 @ 1 V for fuel oxidation. The maximum
current density at 1 V then increases in the following tests, reaching the
values of 414 and 424 mA cmꢀ 2 respectively in the second and in the
third scan. Probably a catalyst stabilization occurs during the first scan,
leading to a Pd-CeO2/C reactivation. In fact, the reversible formation of
poisoning palladium oxides onto the nanoparticles surface is a typical
phenomenon of palladium-ceria catalysts due to the strong oxophilic
CeO2 behavior [38]. Scan voltages acquired for formate oxidation looks
very different respect the other curves of the series. Formate oxidation
onset potential is very low, c.a. 0.1 V (table 1, entry 1), while for MeOH
and EtOH oxidation starts at c.a. 0.4 V (see table 1 entries 2 and 3). In
general, formate and low aliphatic alcohols oxidation is enhanced by the
OHꢀ ions spillover from ceria to Pd nanoparticles which leads to the
formation of Pd(I)-OHads species at lower potential respect Pd/C catalyst
Formate is the substrate that requires the lowest energy input among
the various alcohols studied, c.a 6 kWh kgH ꢀ 1 (table 1, entry 1), due to
2
the very low working potential, less than 0.25 V for c.a. 15 h of elec-
trolysis (Fig. 5A). Methanol and ethanol fed electroreformers require a
higher mean energy input, respectively 16.8 and 13.7 kWh kgH ꢀ 1. The
2
higher mean percentage of fuel converted to products among three ex-
periments, 81% for MeOH and 95% for EtOH, reward the higher energy
consumption of these cells. The low energy input and the higher con-
version, coupled with a negligible toxicity, makes ethanol a more
promising fuel than methanol
Diols and glycerol fed cells have a lower fuel conversion to products
compared to ethanol and methanol fed devices (table 1, entries 4–7).
Conversions span from 77.3% (mean value on three experiments) for
1,2-propandiol oxidation to 52.3% (mean value on three experiments)
for the glycerol. In general, polyol oxidation promoted by palladium
5

M. Bellini et al.
Inorganica Chimica Acta 518 (2021) 120245
Fig. 6. Electroreforming of aqueous 2 M KOH 2 M 1,2-propandiol (A), 2 M ethylene glycol (B), 2 M 1,4-butanediol (C) and 2 M glycerol (D) at 30 mA cmꢀ 2 current
load. Each cell was subjected to three consecutive experiments using a fresh fuel solution for every galvanostatic: cycle 1 (blue curve), cycle 2 (red curve) and cycle 3
(green curve).
based catalysts is less efficient than small alcohols oxidation, such as
Table 2
MeOH and EtOH [18]. This leads to a cell energy input increase, up to
Fuel exhausts analysis by HPLC and 13C NMR. 1,2-P = 1,2-propandiol, EG =
17.7 kWh kgH ꢀ 1 for the 1,2-propandiol fed reformer. For each cell, the
three galvano2statics experiments are reproducible within the experi-
ethylene glycol, Gly = glycerol, 1,4-B = 1,4-butanediol.
Entry
Fuel
Selectivity (%)
mental variability, showing a high stability of the Pd-CeO2/C anode for
long term electrolysis at 60 ◦C and mild current loads (120 mA).
The main goal of the electroreforming technique is the very low
energy input required for pure hydrogen production. The cell here
1
2
HCOOꢀ
Carbonate (100)
Carbonate (95)
Formate (5)
MeOH
3
4
EtOH
1,2-P
Acetate (100)
Lactate (84)
studied is one of the most energy efficient electroreformer ever reported,
ꢀ 1
with an electricity consumption spanning from 6 kWh kgH
for the
Acetate (7)
for 1,4-butanediol fed 2device. All
ꢀ 1
Carbonate (9)
Glycolate (81)
Formate (13)
Oxalate (4)
formate fed cell to 17.7 kWh kgH
2
5
6
7
EG
these values are lower respect the 20 kWh kgH ꢀ 1 energy consumption of
2
the most performing electroreformes reported in literature [12,18].
From an overall energetic point of view, is the exploitation of
renewable alcohols, such as bioethanol, convenient for hydrogen pro-
duction by electroreforming with respect to water electrolysis?
Considering the overall ethanol electroreforming reaction (eq. (1)),
the production of 1 kg of hydrogen (corresponding to 495 mol) requires
the consumption of 11.4 kg of ethanol (247.5 mol) and 9.9 kg of sodium
hydroxide (247.5 mol). NaOH was considered instead of KOH since no
data on potassium hydroxide Life Cycle Assessment (LCA) are reported
in literature.
Acetate (1)
Carbonate (1)
Glycerate (48)
Tartronate (28)
Glycolate (4)
Oxalate (2)
Gly
Carbonate (19)
4-hydroxybutanate (58)
Succinate (42)
1,4-B
CH3CH2OH + OHꢀ →CH3COOꢀ + 2H2
(1)
and the energy input required for its production. Ethanol obtained from
first generation biomasses, such as corn, have an EROEI lower than 1, so
more energy is spent for producing this fuel, compared to the energy
contained in it [40,41]. Only a bioethanol EROEI larger than 5.1 will
lead to a net energy saving in hydrogen production with respect to
current water electrolysis technology. In conclusion, H2 production by
the electroreforming of bioethanol will be energetically convenient
depending on the source of the alcohol, for example, by using sugarcane,
which is commonly reported to have an EROEI of 8, and by using cel-
lulose, which has an EROEI potentially up to 35, but not by using corn
which have an EROEI lower than 1 [42–45].
ꢀ 1
Sodium hydroxide production has an energy cost of 0.97 kWh kg
NaOH
[39] and one mole of alkali is consumed for producing one mole of
dihydrogen (eq.1). So to evaluate the electroreformer overall energy
consumption, 9.7 kWh kgH ꢀ 1 per kilogram of produced hydrogen must
2
be added to the electric energy input of the cell. The ethanol fed
reformer here reported has a mean electricity consumption of 13.7 kWh
kgH ꢀ 1 (average value calculated over three cycles). In contrast, the best
2
PEM water electrolyzers require c.a. 47 kWh kgH ꢀ 1 of electricity to split
2
water into hydrogen and oxygen, so ethanol electroreforming will gain a
net energy saving for H2 production only if the fuel (alcohol) energy cost
The composition of the fuel solution after each test was quantita-
tively and qualitatively characterized using HPLC analysis and 13C{1H}
NMR spectroscopy. Table 2 summarizes the oxidation product distri-
bution and Fig. 7 shows the 13C NMR spectra of the solutions recovered
after the first cycle of each galvanostatic experiment. For formate the
is lower than 23.6 kWh kgH ꢀ 1. This will be achieved only if the energy
2
spent to produce the 1 kg of EtOH used in the electroreformer is less than
2.1 kWh kgꢀEt1OH. Bioethanol net energy cost is usually expressed in terms
of the EROEI, namely, the ratio between the output energy of ethanol
6

M. Bellini et al.
Inorganica Chimica Acta 518 (2021) 120245
Fig. 7. 13C NMR spectra of fuel exhaust solutions. (A) methanol, (B) ethanol, (C) 1,2-propandiol, (D) ethylene glycol, (E) glycerol and (F) 1,4-butanediol spent fuels.
spectra are not reported since carbonate is the only possible reaction
product (100% selectivity, see table 2 entry 1), obtained following a 2e-
oxidative pathway. Methanol oxidation leads to 95% of carbonate
through a 4e- electrochemical process and only 5% of formate, obtained
through a 2e- pathway (Fig. 7A and table 2, entry 2). In contrast, for
EtOH and the other polyols studied the oxidation reaction does not
proceed to CO23ꢀ but follows generally a four electron pathway leading
to the formation of monocarboxylic compounds as principal products.
This is a well-known feature of Pd and Pt based electrocatalysts when
alcohol oxidation happens in strong alkaline media [5]. So in the case of
ethanol, the only product obtained is acetate (100% selectivity), as
described in Fig. 7B and table 2, entry 3. The main oxidation product of
1,2-propandiol is lactate, with 84% selectivity. The minor products ac-
etate (7%) and carbonate (9%) are also present (Fig. 7C and table 2,
entry 4) [29]. In a similar manner, the cell fed with ethylene glycol
yields 81% of glycolate together with 13% of formate and 1% of acetate
and carbonate (Fig. 7D and table 2, entry 5). Glycerol oxidation results
in a mixture of products including glycerate (48%), tartronate (28%)
–
and minor products, obtained from C C chemical bond scission, such as
oxalate, glycolate, formate and carbonate. No trace of the secondary
alcohol oxidation products of glycerol such as dihydroxyacetone,
hydroxypiruvate, or mesoxalate is observed (Fig. 7E and table 2, entry
6).
1,4-butanediol oxidation, leads to the formation a mixture of 58% of
4-hydroxybutanate and 42% of succinate (Fig. 7F and table 2, entry 7).
The selectivity of polyol electrooxidation to the mono carboxylate is
lower with increasing the aliphatic chain length because the oxidation of
the residual alcohol groups in the obtained monocarboxylic product also
occurs. A longer aliphatic chain will reduce the steric and electronic
effects which may hinder oxidation of the monocarboxylic compound
7

M. Bellini et al.
Inorganica Chimica Acta 518 (2021) 120245
Scheme 1. A simplified oxidation mechanism of: (a) formate, (b) methanol, (c) ethanol; (d) 1,2-propandiol; (e) ethylene glycol; (f) glycerol and (g) 1,4-butanediol.
Main products are evidenced in solid boxes, secondary products in dashed boxes.
[19]. So, in the case of ethylene glycol, a small molecule, the oxidation
of glycolate is not favored and only a small amount of 4% of oxalate was
obtained. In contrast, 1,4-butanediol oxidation leads to the formation of
58% of the monocarboxylate (4-hydroxybutanate) and 42% of the
dicarboxylic compound (succinate). In a similar manner, regarding
glycerol oxidation, the intermediate glycerate is readily oxidized and
yields 28% of tartronate.
thoroughly described in the literature [46–48] and are summarized in
scheme 1 (reactions d, e, f and g). Solid boxes emphasize the primary
products, while dashed boxes underline the minor products obtained
from monocarboxylate oxidation to dicarboxylic compounds. The main
products of reactions c-g where obtained oxidizing the alcohol to the
corresponding aldehyde and two electrons. Such a reaction is considered
the overall reaction rate determining step. The aldehyde is then quickly
oxidized to the corresponding carboxylic compound and two electrons
are released to the anode. So the overall reaction involves four electrons
The oxidation mechanisms of 1,2-propandiol, ethylene glycol, glyc-
erol and 1,4 butanediol promoted by Pd based catalysts has been
8

M. Bellini et al.
Inorganica Chimica Acta 518 (2021) 120245
[22]. In a similar manner, the main products are further oxidized to
dicarboxylic products, since the oxidation involve the residual alcoholic
group in the molecule.
Bibliography
[1] R.L. Orbach, Energy Production: A Global Perspective, 2012.
[2] D.J. Murphy, C.A.S. Hall, Ann. N. Y. Acad. Sci. 1219 (2011) 52–72.
[3] T.K. Ghosh, M.A. Prelas, Energy Resources and Systems. Volume 1 - Fundamentals
and Non-Renewable Resources, 2009.
The electrochemical reforming of alcohols in both acid and alkaline
media has been extensively studied for a range of catalyst types [49]. In
terms of reduced energy cost, Pd-CeO2/C outperforms most catalyst
types aside from Rh/C [21], while the 97% conversion efficiency of
ethanol by Pd-CeO2/C is the highest observed so far in electroreforming
of alcohols.
[4] S. van Renssen, Nat. Clim. Chang. 10 (2020) 799–801.
[5] A. Lavacchi, H. Miller, F. Vizza, Nanotechnol. Electrocatal. Energy (2013).
[6] P. Nikolaidis, A. Poullikkas, Renew. Sustain. Energy Rev. 67 (2017) 597–611.
[7] L. Barreto, A. Makihira, K. Riahi, Int. J. Hydrogen Energy 28 (2003) 267–284.
[8] S. Sharma, S.K. Ghoshal, Renew. Sustain. Energy Rev. 43 (2015) 1151–1158.
[9] J.A. Turner, Science 305 (2004) 972–974.
[10] S. Marini, P. Salvi, P. Nelli, R. Pesenti, M. Villa, M. Berrettoni, G. Zangari, Y. Kiros,
Electrochim. Acta 82 (2012) 384–391.
4. Conclusions
[11] M. Carmo, D.L. Fritz, J. Mergel, D. Stolten, Int. J. Hydrogen Energy 38 (2013)
4901–4934.
This paper describes the performance of an electrochemical reformer
using Pd-CeO2/C as anode catalyst operating at 60 ◦C and fed with an
alkaline solution of various alcohols. The fuels studied were: formate,
MeOH, EtOH, 1,2-propandiol, ethylene glycol, glycerol and 1,4-butane-
diol. The results highlight the excellent and stable activity of Pd-CeO2/C.
The amount of electrical energy required for H2 production for all al-
[12] Y.X. Chen, A. Lavacchi, H.A. Miller, M. Bevilacqua, J. Filippi, M. Innocenti,
A. Marchionni, W. Oberhauser, L. Wang, F. Vizza, Nat. Commun. 5 (2014) 4036.
[13] Z. Hu, M. Wu, Z. Wei, S. Song, P.K. Shen, J. Power Sources 166 (2007) 458–461.
[14] T. Take, K. Tsurutani, M. Umeda, J. Power Sources 164 (2007) 9–16.
[15] V. Bambagioni, M. Bevilacqua, C. Bianchini, J. Filippi, A. Lavacchi, A. Marchionni,
F. Vizza, P.K. Shen, ChemSusChem 3 (2010) 851–855.
[16] F. Vitse, M. Cooper, G.G. Botte, J. Power Sources 142 (2005) 18–26.
[17] M. Cooper, G.G. Botte, J. Electrochem. Soc. 153 (2006) A1894.
cohols is in the range of 6.0 – 17.7 kWh kgH ꢀ 1. In particular, 13.7 kWh
2
¨
[18] H.A. Miller, M. Bellini, F. Vizza, C. Hasenohrl, R.D. Tilley, Catal, Sci. Technol. 44
ꢀ 1
kgH
from EtOH results in an electrical energy saving of 33.3 kWh
2
(2016) 7540–7590.
kgH ꢀ 1 as compared with that reported by the DOE for PEM electrolyzer.
[19] M. Bellini, J. Filippi, H.A. Miller, W. Oberhauser, F. Vizza, Q. He, H. Grützmacher,
ChemCatChem 9 (2017) 746–750.
2
Formate is also a very promising fuel since its electroreforming process
[20] J. Mahmoudian, M. Bellini, M.V. Pagliaro, W. Oberhauser, M. Innocenti, F. Vizza,
H.A. Miller, ACS Sustain. Chem. Eng. (2017).
ꢀ 1
to hydrogen and carbonate requires only 6 kWh kgH . The other aspect
studied was the selectivity of alcohol oxidation in2 batch experiments.
The electroreformer here presented was selective for MeOH oxidation to
carbonate (95%), EtOH to acetate (100%), 1,2-propandiol oxidation to
lactate (84%), ethylene glycol to glycolate (81%). The other polyols
(ethylene glycol and 1,4-butanediol) were oxidized to more complex
mixtures of partial oxidation products. For the first time we have
demonstrated how an electrochemical reformer can operate with seven
alcohol fuels. It can be envisaged that this technology can accommodate
high variability in bioalcohol availability and market fluctuation as well
as geographical and seasonal variability. The Pd-CeO2/C catalyst ex-
hibits superior conversion efficiencies over the whole range of alcohols
studied. Combined with high stability and selectivity for valuable
chemicals distinguishes this catalyst as ideal candidate for the indus-
trialization of electrochemical reforming in the future.
[21] M.V. Pagliaro, M. Bellini, M. Bevilacqua, J. Filippi, M.G. Folliero, A. Marchionni,
H.A. Miller, W. Oberhauser, S. Caporali, M. Innocenti, F. Vizza, RSC Adv. 7 (2017)
13971–13978.
[22] C. Bianchini, P.K. Shen, Chem. Rev. 109 (2009) 4183–4206.
[23] S. Zinoviev, F. Müller-Langer, P. Das, N. Bertero, P. Fornasiero, M. Kaltschmitt,
G. Centi, S. Miertus, ChemSusChem 3 (2010) 1106–1133.
[24] N.S. Bentsen, C. Felby, Biotechnol. Biofuels 5 (2012) 1–10.
[25] E. Kirtay, Energy Convers. Manag. 52 (2011) 1778–1789.
[26] C.B. Field, J.E. Campbell, D.B. Lobell, Trends Ecol. Evol. 23 (2008) 65–72.
[27] H. Miller, J. Ruggeri, A. Marchionni, M. Bellini, M. Pagliaro, C. Bartoli, A. Pucci,
E. Passaglia, F. Vizza, Energies 11 (2018) 369.
[28] T. Vo, K. Purohit, C. Nguyen, B. Biggs, S. Mayoral, J.L. Haan, ChemSusChem 8
(2015) 3853–3858.
[29] M. Bellini, M. Bevilacqua, M. Innocenti, A. Lavacchi, H.A. Miller, J. Filippi,
A. Marchionni, W. Oberhauser, L. Wang, F. Vizza, J. Electrochem. Soc. 161 (2014)
D3032–D3043.
[30] L.Q. Wang, M. Bellini, J. Filippi, M. Folliero, A. Lavacchi, M. Innocenti,
A. Marchionni, H.A. Miller, F. Vizza, Appl. Energy 175 (2016) 479–487.
[31] L. Wang, A. Lavacchi, M. Bevilacqua, M. Bellini, P. Fornasiero, J. Filippi,
M. Innocenti, A. Marchionni, H.A. Miller, F. Vizza, ChemCatChem 7 (2015)
2214–2221.
CRediT authorship contribution statement
[32] A. Lenarda, M. Bellini, A. Marchionni, H.A. Miller, T. Montini, M. Melchionna,
F. Vizza, M. Prato, P. Fornasiero, Inorganica Chim. Acta 470 (2018) 213–220.
[33] M. Melchionna, M. Bevilacqua, P. Fornasiero, Mater. Today Adv. 6 (2020)
100050–100061.
Marco Bellini: Investigation, Formal analysis. Maria V. Pagliaro:
Investigation, Visualization. Andrea Marchionni: Investigation, Visu-
alization. Jonathan Filippi: Writing - review & editing, Visualization.
Hamish A. Miller: Investigation, Formal analysis. Manuela Bev-
ilacqua: Investigation. Alessandro Lavacchi: Investigation, Validation,
Data curation. Werner Oberhauser: Investigation. Jafar Mahmou-
dian: Investigation. Massimo Innocenti: Conceptualization. Paolo
Fornasiero: Conceptualization, Writing - original draft, Supervision.
Francesco Vizza: Conceptualization, Supervision, Project administra-
tion, Funding acquisition.
[34] A. Beltram, M. Melchionna, T. Montini, L. Nasi, R.J. Gorte, M. Prato, P. Fornasiero,
Catal. Today 253 (2015) 142–148.
[35] G. Valenti, M. Melchionna, T. Montini, A. Boni, L. Nasi, E. Fonda, A. Criado,
A. Zitolo, S. Voci, G. Bertoni, M. Bonchio, P. Fornasiero, F. Paolucci, M. Prato, A.C.
S. Appl, Energy Mater. 3 (2020) 8509–8518.
[36] V. Bambagioni, C. Bianchini, Y. Chen, J. Filippi, P. Fornasiero, M. Innocenti,
A. Lavacchi, A. Marchionni, W. Oberhauser, F. Vizza, ChemSusChem 5 (2012)
1266–1273.
[37] S.D. Minteer, Biochemical Production of Other Bioalcohols: Biomethanol,
Biopropanol, Bioglycerol, and Bioethylene Glycol, Elsevier, 2011.
[38] H.A. Miller, A. Lavacchi, F. Vizza, M. Marelli, F. Di Benedetto, F. D’Acapito,
Y. Paska, M. Page, D.R. Dekel, Angew. Chemie Int. Ed. 55 (2016) 6004–6007.
[39] L. Thannimalay, S. Yusoff, Z.Z. Norliyana, Aust. J. Basic Appl. Sci. 7 (2013)
421–431.
Declaration of Competing Interest
[40] J.E.A. Seabra, I.C. Macedo, M.R.L.V. Leal, Greenhouse Gases Emissions Related to
Sugarcane Ethanol, Editora Edgard Blücher (2014).
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
[41] D. Pimentel, T.W. Patzek, Nat. Resour. Res. 14 (2005) 65–76.
[42] P.W. Gallagher, W.C. Yee, H.S. Baumes, 2015 Energy Balance for the Corn-Ethanol
Industry, 2016.
˜
[43] C.F. Ruviaro, M. Gianezini, F.S. Brandao, C.A. Winck, H. Dewes, J. Clean. Prod. 28
Acknowledgements
(2012) 9–24.
[44] J.E.A. Seabra, I.C. Macedo, H.L. Chum, C.E. Faroni, C.A. Sarto, Biofuels Bioprod.
Biorefining 5 (2011) 519–532.
The authors would like to thank the Italian ministry MIUR through
the PRIN 2017 “Novel Multilayered and Micro-Machined Electrode
Nano-Architectures for Electrocatalytic Applications (Fuel Cells and
Electrolyzers).
[45] C.A.S. Hall, B.E. Dale, D. Pimentel, Sustainability 3 (2011) 2413–2432.
˜
[46] M. Simoes, S. Baranton, C. Coutanceau, Appl. Catal. B Environ. 93 (2010) 354–362.
[47] Y. Kwon, K.J.P. Schouten, M.T.M. Koper, ChemCatChem 3 (2011) 1176–1185.
[48] A. Marchionni, M. Bevilacqua, C. Bianchini, Y.X. Chen, J. Filippi, P. Fornasiero,
A. Lavacchi, H. Miller, L. Wang, F. Vizza, ChemSusChem 6 (2013) 518–528.
[49] H.A. Miller, A. Lavacchi, F. Vizza, Curr. Opin. Electrochem. 21 (2020) 140–145.
9