Effects of hydrophobicity of diffusion layer on the electroreduction of biomass derivatives in polymer electrolyte membrane reactors

Article abstract of DOI:10.1002/cssc.201402302

For the first time, the hydrophobicity design of a diffusion layer based on the volatility of hydrogenation reactants in aqueous solutions is reported. The hydrophobicity of the diffusion layer greatly influences the hydrogenation performance of two model biomass derivatives, namely, butanone and maleic acid, in polymer electrolyte membrane reactors operated at atmospheric pressure. Hydrophobic carbon paper repels aqueous solutions, but highly volatile butanone can permeate in vapor form and achieve a high hydrogenation rate, whereas, for nonvolatile maleic acid, great mass transfer resistance prevents hydrogenation. With a hydrophilic stainless-steel welded mesh diffusion layer, aqueous solutions of both butanone and maleic acid permeate in liquid form. Hydrogenation of maleic acid reaches a similar level as that of butanone. The maximum reaction rate is 340 nmolcm^-2 s^-1 for both hydrogenation systems and the current efficiency reaches 70%. These results are better than those reported in the literature.

Full text of DOI:10.1002/cssc.201402302

DOI: 10.1002/cssc.201402302
Full Papers
Effects of Hydrophobicity of Diffusion Layer on the
Electroreduction of Biomass Derivatives in Polymer
Electrolyte Membrane Reactors
[a]
[a]
[a]
[a]
[b]
[a]
Wei Chen, Gaohong He, Feilong Ge, Wu Xiao, Jay Benziger, and Xuemei Wu*
For the first time, the hydrophobicity design of a diffusion
layer based on the volatility of hydrogenation reactants in
aqueous solutions is reported. The hydrophobicity of the diffu-
sion layer greatly influences the hydrogenation performance of
two model biomass derivatives, namely, butanone and maleic
acid, in polymer electrolyte membrane reactors operated at at-
mospheric pressure. Hydrophobic carbon paper repels aque-
ous solutions, but highly volatile butanone can permeate in
vapor form and achieve a high hydrogenation rate, whereas,
for nonvolatile maleic acid, great mass transfer resistance pre-
vents hydrogenation. With a hydrophilic stainless-steel welded
mesh diffusion layer, aqueous solutions of both butanone and
maleic acid permeate in liquid form. Hydrogenation of maleic
acid reaches a similar level as that of butanone. The maximum
ꢀ
2
ꢀ1
reaction rate is 340 nmolcm
s
for both hydrogenation sys-
tems and the current efficiency reaches 70%. These results are
better than those reported in the literature.
Introduction
Nowadays, the decline of fossil reserves has forced
[
1]
people to seek renewable energy sources. Biofuels
are generally viewed as promising renewable energy
sources, since they are environmentally friendly and
compatible with the existing infrastructure. The US
Department of Energy has set goals that 30% of the
liquid petroleum transportation fuel and 25% of in-
dustrial organic chemicals should be produced from
[
2]
biomass by the year 2025.
There are several critical techniques to produce bi-
ofuels from biomass, such as catalytic hydrogenation
of C=C, C=O, and even CꢀOꢀC bonds; hydrodeoxy-
genation of carbohydrates for the purposes of up-
grading energy density and improving thermal and
chemical stability; and the storage of intermittent re-
[
2b,c,3]
Figure 1. Schematics of the device for the electrocatalytic hydrogenation of bio-oils in
a PEM reactor. Temperature control is employed at the humidifier, anode inlet, PEM reac-
tor, and cathode circulating tank. a) Current collecting plate; b) serpentine flow channel
newable electricity in the form of liquid fuel.
Polymer electrolyte membrane (PEM) reactors have
attracted much attention as promising electrochemi- in graphite blocks; c) MPL of the gas diffusion layer (GDL); d) macroporous layer of the
GDL; e) CL; f) proton exchange membrane.
cal hydrogen pumps and electrocatalytic hydrogena-
[
4,5]
tion devices.
As shown in Figure 1, with a power
supply, any catalytic hydrogenation or electroreduc-
tion should be possible in PEM reactors, regardless of the feasi-
[5a,b,6]
bility of thermodynamics.
The most significant advantage
[
a] W. Chen, Prof. G. He, F. Ge, Prof. W. Xiao, Prof. X. Wu
State Key Laboratory of Fine Chemicals, Research and
Development Center of Membrane Science and Technology
Dalian University of Technology, 2 Linggong Road
Dalian, 116024 (P.R. China)
of PEM hydrogenation reactors over conventional three-phase
reactors is the mild reaction conditions, especially atmospheric
pressure, due to the ability to produced adsorbed hydrogen
atoms readily in situ on the catalyst surface through current/
Fax: (+86)411-8498-6291
E-mail: xuemeiw@dlut.edu.cn
[5c,6b,c]
potential control. In studies by Benziger and Green et al.,
Homepage: http://gs1.dlut.edu.cn/Supervisor/Front/dsxx/new/Default.aspx?-
WebPageName=wuxuemei
they reported a negligible hydrogen transport resistance for
the hydrogenation of acetone in PEM reactors, which was
barely achieved in conventional heterogeneous reactors, even
upon imposing a large portion of working capital to reach
[
b] Prof. J. Benziger
Department of Chemical Engineering
Princeton University, Engineering Quadrangle A407
Princeton, NJ 08544-5263 (USA)
[7]
pressures of tens of MPa. Moreover, reactions in PEM reactors
ChemSusChem 2015, 8, 288 – 300
288
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
feature high selectivity and the reactors can be operated in
a sustainable and environmentally friendly way, for instance,
the hydrogen feedstock can be made from the electrolysis of
water, with the electricity needed being supplied by renewable
power sources, such as solar, wind, geothermal, and tidal pow-
stainless-steel welded meshes (SSWMs) are used as hydropho-
[15]
bic and -philic diffusion media, respectively. Aqueous solu-
tions of butanone (C=O) and maleic acid (C=C) are selected as
model biomass-derived compounds because of their large vol-
atility differences. We find that the hydrophobic carbon paper
repels an aqueous solution of butanone at atmospheric pres-
sure, but butanone can permeate through the carbon paper in
the form of vapor due to its high volatility to give a high hy-
drogenation reaction rate. However, for the nonvolatile aque-
ous solution of maleic acid, mass transport through the hydro-
phobic carbon paper is the rate-limiting step, which results in
an almost zero reaction rate for hydrogenation. With the hy-
drophilic SSWM DL, the hydrogenation of maleic acid can
reach a similar level as that of butanone.
[
5c,6b,c,8]
ers.
Therefore, hydrogenation in PEM reactors appears
to be attractive from environmental and/or economic points of
view.
Several studies have been successfully carried out on the
electrocatalytic reduction of organic compounds in PEM reac-
tors. Ma et al. achieved the cogeneration of electricity and val-
[
9]
uable chemicals. Pintauro et al. studied the hydrogenation of
soybean oil and found an improved fatty acid hydrogenation
[
5a,
selectivity with a lower reaction temperature and pressure.
b,6a,10]
Acetone, which is a model compound of biomass deriva-
tives, was investigated by Benziger and the group of Green, in
[
5c,6c]
terms of the operating parameters and different catalysts.
Results and Discussion
The hydrogenation of o-xylene and the selective electrocatalyt-
Effects of the DL on hydrogenation reactions
ic reduction of levulinic acid with a PEM flow cell reactor were
[
8,11]
also investigated.
In previous studies, PEM reactors with hydrophobic GDLs were
directly used for electrocatalytic hydrogenation of biomass de-
rivatives. It is assumed that the biomass-derived aqueous solu-
All previous studies took advantage of the configuration of
the PEM fuel cell (PEMFC), which provided adsorbed hydrogen
atoms easily through an electrode reaction and usually had
a hydrophobic GDL to repel liquid water and improve H /O
[5c,6b,c]
tions transfer through the DL in the form of a liquid.
However, liquid water has proven difficult to transfer through
the hydrophobic carbon cloth or carbon paper at atmospheric
2
2
[
12]
transport. Based on previous studies, at atmospheric pres-
sure, liquid water could barely flow through the hydrophobic
[12b,13,16]
pressure.
It is necessary to determine the transfer form
[
13]
GDL because of the high surface energy. Therefore, we can
imagine that the hydrophobic GDL would not be a good fit for
the hydrogenation of bio-oils because bio-oils contain about
of the biomass derivative reactants in aqueous solutions
through the DL, which will greatly influence the mass transfer
and thus electrocatalytic hydrogenation rate. Because the
chemical composition of bio-oils is complicated (boiling points
from a few dozen degrees to more than 4008C with 35% non-
[
3a,6c,14]
25 wt% of liquid water and are hydrophilic in nature.
Great mass transport resistance should result as the aqueous
solution of bio-oil flows through the hydrophobic GDL, and
hence, decrease the reaction rate of hydrogenation. However,
Benziger and Green et al. reported negligible acetone transfer
resistance through a hydrophobic GDL with an aqueous solu-
tion of acetone as the feedstock, and they still assumed that
a liquid form of butanone transferred through the hydropho-
[3a,14]
volatile compounds),
two compounds with a large differ-
ence in volatility, namely, butanone and maleic acid, have been
selected as the bio-oil model compounds herein. Electrocata-
lytic hydrogenations of their aqueous solutions were investi-
gated systematically with cathode DLs of different hydropho-
bicity, for example, hydrophobic carbon paper and hydrophilic
SSWMs.
[
5c,6c]
bic GDL.
We show herein that acetone transfers through
hydrophobic GDL in vapor form, due to its high volatility, in-
stead of in the liquid form. Although the chemical composition
of bio-oils is complicated, including 35% of nonvolatile com-
Electrochemical catalytic hydrogenation of butanone and
maleic acid in PEM reactors with different DLs are shown in
Figures 2 and 3, respectively. It is shown clearly that hydropho-
bicity/-philicity of the DL has a great effect on the hydrogena-
tion of different model compounds. In Figure 2, butanone ach-
ieves similar levels of conversion, reaction rate, and current ef-
ficiency, regardless of the hydrophobic/-philic nature of the DL.
This indicates that, for volatile substances such as butanone,
mass transfer resistance in DLs is not the governing factor for
the hydrogenation reaction, although the hydrogenation per-
formances increase with increasing air permeability or mesh
size of the SSWMs. The result coincides with that reported by
Green et al. and Benziger for the hydrogenation of acetone in
PEM reactors. In their work, they also suggested considering
the ignored mass transfer of acetone in the overall resistance
[
3a,6c,14]
pounds,
it is essential to clarify the mass transport of an
aqueous solution of bio-oil through diffusion layers (DLs) of
different hydrophobicity and its relationship with the volatility
of different bio-oil components. These works are instructive for
the design of DLs and will pave the way for the hydrogenation
of aqueous solutions of bio-oil with PEM reactors and greatly
improve the efficiency of hydrogenation for biofuel produc-
tion. To the best of our knowledge, there have been few inves-
tigations that focus on the structure of PEM reactors, especially
the hydrophobicity of GDLs and their influence on the hydro-
genation efficiency with an aqueous solution of bio-oil as the
feedstock.
[5c,6b,c]
Herein, the effects of the hydrophobicity of the DL on the
electrocatalytic hydrogenation of biomass-derived compounds
with different volatility in PEM reactors have been investigated.
Carbon papers treated with polytetrafluoroethylene (PTFE) and
of the PEM reactor.
However, the hydrophobicity/-philicity
of the DL greatly influences the hydrogenation of maleic acid,
as shown in Figure 3. With the hydrophilic SSWM DL, maleic
acid achieves a similar level of hydrogenation as that of buta-
ChemSusChem 2015, 8, 288 – 300
www.chemsuschem.org
289
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Figure 3. Conversion, current efficiency, and reaction rate for the hydrogena-
tion of maleic acid in PEM reactors as functions of a) hydrophobic carbon
papers with different air permeabilities; and b) hydrophilic SSWMs with dif-
Figure 2. Conversion, current efficiency, and reaction rate for the hydrogena-
tion of butanone in PEM reactors as functions of a) hydrophobic carbon
papers with different air permeabilities; and b) hydrophilic SSWMs with dif-
ꢀ1
ferent meshes. Reaction conditions: 15 mLmin maleic acid, 20 sccm hydro-
ꢀ1
ꢀ2
ferent meshes. Reaction conditions: 15 mLmin butanone, 20 sccm
gen, 0.5m maleic acid, 408C, 38 mAcm , 4 h.
ꢀ
1
(
0
sccm=standard centimeter per minute; 1 sccmꢁ1 mmols ) hydrogen,
ꢀ2
.5m butanone, 408C, 38 mAcm , 4 h.
reaction rate with increasing SSWM mesh (Figure 2b). For the
hydrogenation of maleic acid, mass transfer of maleic acid
through the DL plays a more important role on the reaction
rate.
none, whereas with the hydrophobic carbon paper the reac-
tion rate is almost zero and no product (succinic acid) is de-
tected. This indicates that the hydrophobic GDL repels maleic
acid and the mass transfer of maleic acid through the hydro-
phobic DL is the rate-limiting resistance. From interesting ob-
servations made in our experiments, we propose the following
question: why do different components in aqueous solutions
behave differently as they penetrate through DLs of different
hydrophobicity? This is investigated in the following sections.
It is noted from Figures 2b and 3b that the mesh of SSWM
has different influences on the hydrogenation performances of
butanone and maleic acid. Compared with carbon papers, the
hydrophilic SSWMs possess larger pore sizes (as shown in
Table 1); therefore, there is a tradeoff between the void frac-
tion and the contact area between the SSWM and catalyst
layer (CL). As the SSWM mesh increases, the void fraction de-
creases and then less reactant can reach the catalyst surface.
Better contact between the SSWM and CL allows uniform de-
livery of electrons and increases the reaction rate. From the ex-
perimental data obtained, electron transfer greatly influences
the hydrogenation of butanone, which results in an improved
Form of bio-oil components in the aqueous solution
through the DL
The ability of a liquid to penetrate through a solid surface has
great correlation with the surface energy of the liquid on the
surface. Therefore, the contact angle has been measured for
several liquid reagents on carbon papers and SSWMs. As
shown in Table 1, the wettability of hydrophobic carbon
papers (treated with Teflon) in aqueous solution is poor. For in-
stance, the contact angles of the solution of maleic acid
remain around 1408, regardless of the concentration. Although
butanone exhibits good wettability on carbon paper (contact
angle is zero; for example, a drop of butanone penetrates
through the DL immediately), the contact angles of aqueous
solutions of butanone are larger than 1208, even for the high-
est concentration in water at room temperature (2m). It is
known that a larger contact angle indicates a higher surface
[15c,17]
energy and lower surface wettability.
Benziger et al. stud-
ChemSusChem 2015, 8, 288 – 300
www.chemsuschem.org
290
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Table 1. Properties of the DLs and contact angles of aqueous solutions of butanone and maleic acid.
[
a]
[b]
DL
Thickness
mm]
Pore diameter
[mm]
Air permeability of
Void fraction of
SSWMs [%]
Contact angle [8]
3
ꢀ2 ꢀ1
[
GDL [cm cm
s
]
water
butanone
maleic acid
[c]
0.5m
1m
2m
100%
0.5m
1m
2m
[
d]
GDL34 BC
GDL24 BC
GDL25 BC
GDL35 BC
SSWM 100 mesh
SSWM 200 mesh
SSWM 300 mesh
SSWM 400 mesh
315
235
235
325
212
121
109
86
8–15
8–15
8–15
8–15
154
72
0.35
0.6
1.0
1.5
–
–
–
–
–
–
–
140
144
144
143
126
123
127
124
138
142
140
138
127
116
120
112
134
136
135
139
118
109
113
97
123
129
123
124
110
92
0
0
0
0
0
0
0
0
140
137
143
138
125
122
119
110
141
142
141
144
124
109
104
100
140
137
143
137
114
105
96
–
36.8
32.1
29.5
27.0
46
33
89
89
94
[a] Data for the GDLs are provided by SGL Co., about 38–48 mm for carbon paper, 0.09–0.12 mm for microporous layer (MPL), and 8–15 mm for the crack in
the MPL. [b] Measuring errors are less than 58 for the carbon paper GDL and less than 88 for the SSWMs. Aqueous solutions of butanone and maleic acid
were used. [c] 2m is the maximum concentration of an aqueous solution of butanone at room temperature. [d] The liquid drop penetrates through the
DLs immediately.
ied the penetration of water through hydrophobic GDLs, for
ode flow channel experiments were conducted, in which an
aqueous solution of butanone or maleic acid was circulated
through the cathode flow channel sealed by a piece of the DL.
The observations are shown in Figure 4 and Table 2. Aqueous
solutions of both butanone and maleic acid with concentra-
tions of 0.5, 1.0, and 2.0m could flow through hydrophilic
SSWMs in the form of a liquid, even at the lowest flow rate
[
13]
example, carbon paper and carbon cloth. They found that
no water flowed through the hydrophobic GDLs until pres-
sures of 5–10 kPa were applied to overcome the surface
energy of the water/Teflon interface in the largest pores. In
this study, the electrocatalytic hydrogenation reaction in the
PEM reactor operates at atmospheric pressure, which is also
one of the most important advantages of PEM reactors. From
contact angle measurements, it is reasonable to predict that
aqueous solutions of both butanone and maleic acid cannot
flow through the hydrophobic carbon papers in the form of
a liquid; this is different from the concept in the literature that
an aqueous solution of acetone transfers through the GDL in
ꢀ
1
(15 mLmin ; Figure 4b), but no drop of liquid comes out of
the hydrophobic carbon paper during the experimental period
ꢀ
1
of 2 h, even at the highest flow rate (80 mLmin ; Figure 4c).
However, when a glass slide was put above the hydrophobic
carbon paper, a layer of small condensed droplets emerged
quickly on the glass surface close to the carbon paper and the
droplets got bigger with time (Figure 4c). It is interesting that
as droplets butanone is detected by GC, whereas no maleic
[
5c,6c]
liquid form in a PEM hydrogenation reactor.
From the re-
sults given in Table 1, it is surprising that the contact angles of
solutions of butanone and
maleic acid with SSWMs are
mostly larger than 908, although
the stainless-steel material is
quite hydrophilic (contact angle
is 528ꢂ58). The measured con-
tact angles of SSWMs exhibit the
combined effect of the micro-
pores in the meshes (full of air
and hydrophobic) and the stain-
[
18]
less-steel metal itself.
Further
measurement of the hydrostatic
head shows only about 100 Pa is
necessary for the penetration of
water through the SSWMs; this
pressure can safely be provided
by the peristaltic pump for feed-
stock circulation. (The measured
pressure at the cathode inlet is
in the range of 1.5–10.5 kPa for
ꢀ
1
flow rates of 15–80 mLmin .)
To further clarify how the
cathode reactant penetrates
through the DLs at atmospheric
Figure 4. Photograph of cathode flow channel experiments a) with hydrophobic carbon paper as the DL, b) with
hydrophilic SSWMs as the DL, c) a dewdrop on the glass slide above the cathode flow channel with carbon paper
as the DL, d) with hydrophobic carbon paper as the DL and cyclohexane or ethanol as the feed. Experimental con-
ꢀ1
pressure in PEM reactors, cath- ditions: 0.5, 1.0, and 2.0m butanone, or maleic acid; 408C; 2 h; 15 and 80 mLmin circulating fluid.
ChemSusChem 2015, 8, 288 – 300
www.chemsuschem.org
291
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
ity were quantitatively measured. As shown in
Figure 5, after 2 h of circulation in the cathode flow
channel, 40–50% of the butanone feedstock perme-
ates through the carbon paper in the form of vapor,
and the percentage of butanone evaporation in-
creases with the air permeability of the carbon
paper. Therefore, the increasing amount of butanone
on the cathode catalyst surface accounts for the in-
crease in the reaction rate observed in Figure 2a.
However, the percentage evaporation of maleic acid
through carbon paper is less than 5%, due to the
low vapor pressure of maleic acid. Thus, no product
of the hydrogenation of 0.5m maleic acid with the
hydrophobic GDL is observed in Figure 3a.
Table 2. Permeation of different liquids through DLs from the cathode flow channel
experiments.
[
a]
DL
Water
Maleic acid
Butanone
Ethanol Cyclohexane
0.5m 1.0m 2.0m 0.5m 1.0m 2.0m 100%
GDL34 BC
GDL24 BC
GDL25 BC
GDL35 BC
SSWM
ꢀ
ꢀ
ꢀ
ꢀ
+
ꢀ
ꢀ
ꢀ
ꢀ
+
ꢀ
ꢀ
ꢀ
ꢀ
+
ꢀ
ꢀ
ꢀ
ꢀ
+
ꢀ
ꢀ
ꢀ
ꢀ
+
ꢀ
ꢀ
ꢀ
ꢀ
+
ꢀ
ꢀ
ꢀ
ꢀ
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
1
00 meshes
SSWM
00 meshes
SSWM
00 meshes
SSWM
00 meshes
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
2
3
4
The results show that, at atmospheric pressure,
the hydrophilicity/-phobicity of the DL has a great
effect on the permeation form of the aqueous solu-
tions of bio-oil. A schematic model is proposed for
the mass transfer through DL, as shown in Figure 6.
For hydrophobic DLs, the reactants in aqueous solutions
cannot penetrate through them in the form of a liquid because
of the high surface energy. In this case, the volatility of the
components plays an important role on the mass transfer of
[
a]+: Permeable; ꢀ: nonpermeable. Aqueous solutions of maleic acid and butanone
were used.
1
acid is detected by H NMR spectroscopy. From cathode flow
channel experiments, it can be seen that butanone penetrates
through the hydrophobic carbon papers in the form of a vapor
due to its volatility (the boiling point of butanone is 79.68C, sa-
turated vapor pressure of 24.0 kPa at 408C, data from
Aspen 7.2), and then reacts with adsorbed hydrogen atoms on
the catalyst surface. However, maleic acid is a nonvolatile sub-
stance with a boiling point of 355.58C and a saturated vapor
pressure of 0.046 Pa at 408C (data from Aspen 7.2). It barely
penetrates through the hydrophobic carbon papers in the
form of vapor-like butanone and the droplets are only con-
densed water vapor (Figure 4c). As a result, the poor mass
transfer of maleic acid through hydrophobic carbon paper pre-
vents its participation in the hydrogenation reaction on the
cathode CL. In case of hydrophilic SSWMs, it is easy for the
polar aqueous solutions to permeate through SSWMs, as
driven by a peristaltic pump for feedstock circulation, and then
participate in hydrogenation in the form of liquid. Mass trans-
fer resistance is not the governing factor. Therefore, the hydro-
genation of maleic acid and butanone reaches similarly high
levels, as shown in Figures 2 and 3. In the literature, hydropho-
bic or amphiphilic solvents, such as cyclohexane and ethanol,
were also reported as solvents
Figure 5. Percentage of evaporation as a function of air permeability of
carbon paper. Experimental conditions: 0.5m butanone or maleic acid, 408C,
ꢀ1
2
h, 15 mLmin circulating fluid.
for PEM hydrogenation reactor-
[
5a,b,9–11]
s.
As shown in Figure 4d,
this kind of solvent can perme-
ate through both the hydropho-
bic carbon papers and the hy-
drophilic SSWMs immediately
due to their low surface energy.
Therefore, the mass transfer re-
sistance through DLs can be ig-
nored for these nonaqueous sol-
vent (low polarity) systems.
Furthermore, the percentages
of butanone and maleic acid
that permeate through carbon
Figure 6. Schematic model for the mass transfer of aqueous solutions of bio-oil in PEM reactors through a) a hy-
drophobic DL for volatile compounds, b) a hydrophobic DL for nonvolatile compounds, and c) a hydrophilic DL
papers of different air permeabil- for both volatile and nonvolatile compounds.
ChemSusChem 2015, 8, 288 – 300
www.chemsuschem.org
292
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
cathode reactants through DLs. The volatile components, such
as butanone in the bio-oils, can transfer through the DLs in
the form of vapor and take part in the hydrogenation reaction
on the cathode catalyst surface (Figure 6a). However, vapor
transfer can barely occur for nonvolatile compounds such as
maleic acid. Only water vapor transfers through the DLs (Fig-
ure 6b). With hydrophilic DLs, an aqueous solution of both vol-
atile and nonvolatile bio-oil can penetrate through them in the
form of a liquid and the concentration of reactant in the cath-
ode CL is mainly related to the feed concentration rather than
the volatility (Figure 6c). Based on this analysis, different hy-
drogenation rates for butanone and maleic acid with DLs of
different hydrophobicity (Figures 2 and 3) can be interpreted.
Whether in vapor or liquid form, volatile butanone in the aque-
ous solution can easily transfer through both hydrophobic and
energy on both hydrophobic and -philic surfaces. In this case,
the hydrophobicity of the DL will not greatly influence the
mass transfer of the reactants, and the resistance of DLs can
[5a,b,9–11]
be ignored, as reported in the literature.
Electroreduction of aqueous solution of butanone
In this section, results from investigations into the operating
parameters to improve the performance of the PEM reactors as
much as possible are discussed. Based on the above discus-
sion, both hydrophobic and -philic DLs can be used for the hy-
drogenation of an aqueous solution of volatile butanone. Here-
inafter, commercial hydrophobic carbon paper GDLs were used
as the DL because of their advantages for PEMFC applications,
such as stability, good conductivity, and protection of the
[
12b]
-philic DLs. Although the reaction rate increases with air per-
CL.
Specifically, GDL35 BC (with the highest air permeability
3
ꢀ2 ꢀ1
meability of the GDL or mesh of the SSWMs, mass transfer re-
sistance through the DLs is not the rate-limiting factor for the
hydrogenation of volatile butanone, and results in a similar
high level of reaction rate. However, there is great difference
between the mass transfer of nonvolatile maleic acid through
the hydrophobic and -philic DL. Maleic acid could barely pene-
trate through the hydrophobic carbon paper in the form of
a vapor. In this case, mass transfer of maleic acid through the
DL becomes the rate-limiting resistance, which results in a very
low reaction rate. Therefore, under atmospheric pressure, a hy-
drophilic DL should be used to promote flow of an aqueous
solution of maleic acid through the DL in the form of a liquid,
and thus, improve the reaction rate. It is worth mentioning
that the chemical composition of bio-oils is complicated and
the pyrolytic oil includes about 35% nonvolatile com-
of 1.5 cm cm s ) was selected due to its high performance
(Figure 2).
Figure 7a depicts the effect of current density on the elec-
trocatalytic hydrogenation of an aqueous solution of buta-
none. Current density is calculated by using the geometric
2
area of the membrane electrode assembly (MEA; 5.29 cm ). As
seen in Figure 7a, conversion and reaction rate go through
a maximum with current density. As the current density in-
creases, more protons are transported across the PEM, and
subsequently, reduced to adsorbed hydrogen atoms on the
cathode catalyst surface, which is readily available for hydroge-
nation with the adsorbed organic reactant species. There is
a competition of protons between hydrogenation and hydro-
0
0
gen evolution. Based on Gibbs free energy (DG ; DG of hydro-
0
gen evolution is zero, whereas DG values for the hydrogena-
[
14]
pounds; thus, it is better to use a hydrophilic DL for the hy-
drogenation of the water-soluble bio-oils to ensure more com-
ponents are involved in the reaction. Hydrophobic DLs can
also be used when extra pressure is applied, but the operation
capital will increase.
tion of butanone and maleic acid are ꢀ25.79 and
ꢀ
1
ꢀ122.94 kJmol , respectively), hydrogenation is more favora-
ble than hydrogen evolution. Therefore, the latter will increase
only when the adsorbed hydrogen atoms are present in excess
relative to those required for hydrogenation. At low current
ꢀ
2
The results in this study provide highlights for the design of
DLs in PEM reactors operating at atmospheric pressure. To ach-
ieve high performances, the DLs should be designed according
to the solvent polarity as well as to the volatility of the reac-
tants. When using highly polar compounds, such as water, as
the solvent, the selection of the DL depends on the volatility
of the solute. For volatile reactants, such as acetone, butanone,
density (less than 95 mAcm ), hydrogenation is the dominant
reaction, for example, the reaction is limited by the lack of ad-
sorbed hydrogen atoms. Thus, the current efficiency is high
ꢀ
2
(reaching nearly 100% at 38 mAcm ) and the hydrogenation
reaction rate increases with increasing current density. At high
current density, excess adsorbed hydrogen atoms block the ad-
sorption of butanone, which results in a decrease in the hydro-
genation rate and conversion. Byproduct hydrogen gas will
then be generated and decrease the current efficiency. The
whole process matches the modified Langmuir–Hinshelwood
mechanism well, for which the rate of hydrogen adsorption is
[
5c,6b,c]
and furfural,
both hydrophilic and -phobic DLs are avail-
able for hydrogenation. The reactants can transfer through hy-
drophilic DLs in the form of an aqueous solution, but transfer
through hydrophobic DLs in the form of a vapor because of
their high surface energy. We suggest that this is similar to the
case of the direct methanol fuel cell (DMFC) with conventional
hydrophobic GDLs. For nonvolatile reactants in aqueous solu-
tion, such as maleic acid, they cannot transfer in the vapor
form, so the only mass transfer form is as an aqueous solution.
Thus, hydrophilic DLs must be used to ensure mass transfer of
maleic acid and the occurrence of hydrogenation. When hydro-
phobic or amphiphilic solvents, such as cyclohexane and etha-
nol, are used, they can flow through both hydrophobic and
[5c]
set to equal the current density.
As seen in Figure 7a, the maximum value of the reaction
ꢀ
2
ꢀ1
ꢀ1
ꢀ1
rate is about 340 nmolcm
s
(204 mmolgPt min ) at
ꢀ
2
95 mAcm , and its current efficiency is 68%. The performance
is much better than that reported by Green et al.
ꢀ
1
ꢀ1
(70.7 mmolgPt min for reaction rate and 44% for current ef-
[
6c]
ficiency at 0.1 V and 408C, acetone/hydrogen system). Addi-
tionally, the hydrogenation rate in PEM reactors also has ad-
vantages over that of heterogeneous catalytic hydrogenation.
The traditional heterogeneous catalytic hydrogenation of buta-
-philic DLs in the form of a liquid because of the low surface
ChemSusChem 2015, 8, 288 – 300
www.chemsuschem.org
293
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Figure 7. Conversion, current efficiency, and reaction rate for the hydrogenation of butanone in the electrocatalytic PEM reactor as functions of a) current,
ꢀ
1
b) butanone concentration, c) temperature, and d) reaction time. Reaction conditions: a) 15 mLmin butanone, 20 sccm hydrogen, 408C, 1m butanone, 3 h,
ꢀ
1
ꢀ2
ꢀ1
ꢀ2
b) 15 mLmin butanone, 20 sccm hydrogen, 57 mAcm , 408C, 3 h; c) 15 mLmin butanone, 20 sccm hydrogen, 57 mAcm , 1m butanone, 3 h;
ꢀ
1
ꢀ2
d) 15 mLmin butanone, 20 sccm hydrogen, 38 mAcm , 408C, 0.5m butanone.
none is usually achieved in a stirred stainless-steel autoclave at
a maximum with butanone concentration. This tendency is
similar to that of the acetone/hydrogen system reported by
Green et al. and indicates a Langmuir–Hinshelwood–Hougen–
[
19]
relatively high temperature and pressure. The hydrogenation
rate depends significantly on the catalyst. The highest reaction
ꢀ
1
ꢀ1
[6c]
rate of 850 mmolgPt min was reported by Breen et al. with
Watson (LHHW) rate equation. The decrease in the reaction
RuꢀPt/activated carbon (AC) as the catalyst at 3 bar and 308C;
rate at a high butanone concentration corresponds to the sit-
uation in which adsorbed butanone blocks the adsorption of
hydrogen atoms. Additionally, after 3 h of reaction with 2m bu-
tanone, part of the CL peeled off from the catalyst-coated
membrane (CCM) and stuck to the carbon paper; this did not
happen at lower concentration. Damage to the CL suggests
another reason for the decrease in hydrogenation rate at high
concentration. This is discussed in a later section.
[
19a]
this was attributed to the bimetallic effects.
However, when
monometallic platinum was used as the catalyst, the reaction
ꢀ
1
ꢀ1
rate decreased to 34.8 mmolg min with a graphite sup-
Pt
ꢀ1
ꢀ
1
port and to 1.2 mmolgPt min with an AC support, which
was much smaller than that reported herein. The much higher
activity of monometallic platinum in the electrohydrogenation
reaction in the PEM reactor might be due to a different reac-
tion mechanism, sufficient in situ hydrogen atoms supplied by
Butanone conversion, reaction rate, and current efficiency
also go through a maximum with increasing temperature, as
shown in Figure 7c. In particular, the reaction rate decreases
greatly when temperature increases to 608C. It is known that
both the activity of the catalyst and mass transfer of reactants
increase with temperature, and thus, lead to the increase in
the reaction rate. However, with increasing temperature, the
loss of butanone increases, too. That is, butanone volatilization
loss from the condenser (as shown in Figure 1) and permeation
loss through the Nafion membrane to the anode aggravate
the decrease in adsorbed butanone as well as the reaction
[
5c]
an electrochemical hydrogen pump, and the high platinum
catalyst utility in the unique design of the three-dimensional
electrode of the fuel cell. Therefore, electrohydrogenation with
a PEM reactor is more attractive than that of heterogeneous
catalytic hydrogenation in terms of both the reaction rate and
operating conditions.
Figure 7b shows the effects of butanone concentration on
conversion, current efficiency, and reaction rate. Butanone con-
version decreases with increasing butanone concentration,
whereas the reaction rate and current efficiency go through
ChemSusChem 2015, 8, 288 – 300
www.chemsuschem.org
294
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
rate. High temperature also aggravates swelling of the Nafion
cases can be ignored. In this case, the optimal operating cur-
rent density and maximum reaction rate are closely related to
the active area of the catalyst.
[
20]
membrane and Nafion binder, which contributes to the dra-
matic reduction of the reaction rate at 608C.
The effects of reaction time on conversion, current efficiency,
and reaction rate are shown in Figure 7d. Reaction rate and
current efficiency decrease with time, whereas conversion in-
creases as the reaction proceeds. After 10 h of reaction, the
average reaction rate reduced to 70% of that at 2 h and the
trend to decrease became faster and faster. There are several
reasons for this result. As the reaction proceeds, the concentra-
tion of butanone decreases and the concentration of the prod-
uct, 2-butanol, increases, leading to product inhibition on the
reaction and then a decrease in the reaction rate. In addition,
butanone in the solution may accelerate swelling of Nafion
membrane and Nafion binder as the reaction proceeds, result-
ing in damage to the three-phase interface and even to the CL
structure (as shown later in Figure 9c). Additionally, the long-
term performance of the PEM reactors can also degrade due
to agglomeration of platinum nanoparticles or peeling off of
platinum nanoparticles from the carbon support, as well as oxi-
dation of the carbon support under electrochemical condi-
Figure 8b shows the effect of increasing the maleic acid
concentration on conversion, current efficiency, and reaction
rate. There are maximum values of current efficiency and reac-
tion rate with increasing maleic acid concentration. The con-
version decreases with increasing concentration of maleic acid.
The trends are similar to those of butanone and also indicate
a LHHW rate equation. Excess adsorbed maleic acid blocks the
adsorption of hydrogen atoms at a high maleic acid concentra-
tion, leading to a decrease in the reaction rate. The maximum
reaction rate appears at 2m maleic acid, which is a higher con-
centration than that of the butanone system (decreased dra-
matically at 2m). This is related to less swelling of the Nafion
membrane and Nafion binder in the aqueous solution of
maleic acid. At 3m maleic acid, the performance significantly
reduces due to damage to the cathode CL structure, as dis-
cussed in a later section. It is worth mentioning that, for a high
maleic acid concentration of 3m, maleic acid crystals were
found to adhere to the anode CL after 3 h of reaction due to
the penetration of maleic acid through the Nafion membrane
with liquid water, which may also affect the performance of
the PEM reactor by decreasing the concentration of maleic
acid and utilization of the anode active area. Therefore, we
suggest that a high concentration of nonvolatile organic com-
pounds should not be used for hydrogenation in PEM reactors
for a thin Nafion-based CCM.
[
21]
tions; contact with water and organic compounds can accel-
[6c]
erate degradation. It is worth mentioning that Green et al.
found that no catalyst deactivation was observed for 7 h of re-
action when a conventional hot-pressed gas diffusion elec-
trode (GDE) with Nafion 117 was used. The stability might be
due to the much thicker membrane (175 mm for Nafion 117
versus 18 mm for Nafion/PTFE composite membrane used
herein), the antiswelling PTFE binder for CL, and protection of
the CL from the hot-pressed carbon paper. However, the
proton conductivity of hot-pressed CL is relatively low, and the
long-term stability of the GDE is still doubtful because the
membrane can be swelled by the organic reactant.
Figure 8c shows that the maleic acid conversion, reaction
rate, and current efficiency all go through a maximum value
when the reactor temperature increases. The maximum values
are around 408C and the reaction rate suffers from a serious
decrease at 608C. This is similar to the effect of temperature
on the hydrogenation of butanone. The only difference is the
nonvolatility of maleic acid. As shown in Figure 8d, prolonging
the reaction time increases the conversion, but decreases the
current efficiency and reaction rate of maleic acid. The average
reaction rate and conversion at 11 h are 50% of those at 2 h
and the conversion of maleic acid is almost unchanged after
10 h (from 22.8% at 10 h to 22.9% at 11 h), which indicates
a very low reaction rate after 10 h.
Electroreduction of aqueous solution of maleic acid
As suggested previously that nonvolatile organic components
in bio-oil, such as maleic acid, are impermeable to hydropho-
bic DLs but permeable to hydrophilic ones in the form of
a liquid. According to Figure 3b, the SSWM with 100 meshes
has an improved hydrogenation performance. Therefore, it was
selected as the DL in the following hydrogenation optimization
for an aqueous solution of maleic acid.
Degradation of Nafion-based CCM in PEM reactors
Figure 8 shows the conversion, current efficiency, and reac-
tion rate for the hydrogenation of maleic acid in the electroca-
talytic PEM reactor as functions of current density, maleic acid
concentration, reactor temperature, and reaction time. Current
efficiency decreases with increasing current density, whereas
conversion and reaction rate go through a maximum with cur-
rent density (Figure 8a); this indicates a typical Langmuir–Hin-
shelwood mechanism similar to that of the hydrogenation of
Previous results showed that the performance of the PEM reac-
tors might be closely related to the stability of the CCM, for ex-
ample, the Nafion/PTFE membrane and CL. SEM images of the
surface of the cathode CL after hydrogenation of the two
model compounds were recorded (Figures 9 and 10). In both
cases, the CL structure was gradually damaged (such as crack-
ing or even peeling off) at higher temperature and reactant
concentration. The effect of reactant concentration is the most
significant (see Figures 9a and 10a); the CL structures are
peeled off in 2m butanone and 3m maleic acid. It is worth
mentioning that when SSWM was used as the GDL for the hy-
drogenation of 2m butanone, the solution became cloudy,
which indicated the dissolution of the CL. If a carbon paper
ꢀ
2
ꢀ1
butanone. The maximum reaction rate is 345 nmolcm
s
at
ꢀ
2
a current density of 95 mAcm . This is similar to that of buta-
ꢀ
2
ꢀ1
ꢀ2
none (334 nmolcm
s
at 95 mAcm ). The comparable per-
formance of SSWM with 100 meshes for maleic acid to that of
the commercial GDL for butanone indicates that the mass
transfer resistance of the reactants through the DL in both
ChemSusChem 2015, 8, 288 – 300
www.chemsuschem.org
295
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Figure 8. Conversion, current efficiency, and reaction rate for the hydrogenation of maleic acid in an electrocatalytic PEM reactor as functions of a) current
ꢀ
1
density, b) maleic acid concentration, c) reaction temperature, and d) reaction time. Reaction conditions: a) 20 sccm hydrogen, 15 mLmin maleic acid, 408C,
ꢀ
1
ꢀ2
ꢀ1
ꢀ2
1
m maleic acid, 3 h; b) 20 sccm hydrogen, 15 mLmin maleic acid, 57 mAcm , 408C, 3 h; c) 20 sccm hydrogen, 15 mLmin maleic acid, 57 mAcm , 1m
ꢀ
1
ꢀ2
maleic acid, 3 h; d) 20 sccm hydrogen, 15 mLmin maleic acid, 57 mAcm , 408C, 1m maleic acid.
GDL was used, the solution remained clear during the 3 h of
reaction, although part of CL peeled off from the membrane
and stuck to the GDL after the reaction. This phenomenon
confirmed the protective effect of the MPL for the CL. We sug-
gest that the damage to the CL is related to the Nafion/PTFE
membrane and Nafion binder in the CL. It is well known that
Nafion swells in water and organic compounds, such as ke-
tones and alcohols, and the swelling ratio increases with con-
was attributed to the swelling and expansion of both the
Nafion binder in the CL and Nafion membrane with changes
to the water environment. Degradation usually appeared after
[22]
being operated for a long time. However, in our experiment,
degradation was observed after only a short time (<11 h), es-
pecially when operating at high concentration. This is under-
standable because only oxygen gas and water are present in
the cathode of the PEMFC and DMFC, but there is organic so-
lution in the cathode of the PEM reactors. This comparison
proved, from another aspect, that organic compounds, espe-
cially a high concentration of ketone and alcohol, could accel-
erate the degradation of the CCM due to swelling of the
Nafion binder in the CL and Nafion membrane. Additionally,
from data in Figure 8d, the CL almost loses activity after 11 h
of reaction. However, the SEM images do not show much
damage to the CL structure at 408C with 1m maleic acid at dif-
ferent reaction times. Thus, in addition to the inhibition effect
of the product, another possible reason for the dramatic de-
crease in catalyst activity observed in Figure 8d might be the
agglomeration of platinum nanoparticles or the peeling off of
platinum nanoparticles from their carbon supports, as proven
[
6c,21]
centration and temperature.
Nafion is also used as the
binder in the catalyst ink, which is then coated on the surface
of Nafion membrane by spraying. High concentrations of buta-
none and maleic acid, as well as high temperature, can exacer-
bate swelling of the Nafion membrane and Nafion binder. The
resulting swelling stress in the CCM will damage the three-
phase interface and even peel off the CL (as shown in Figure-
s 9a,b and 10a,b). This is one of the major reasons for the de-
creased reaction rate observed in Figures 7 and 8. Because the
Nafion membrane and Nafion binder swell more in butanone
than in the aqueous solution of maleic acid (Figures 9a and
10a), the PEM reactor can tolerate a higher concentration of
maleic acid. Some researchers already observed the degrada-
tion of the cathode of the CCM in PEMFCs and DMFCs, which
[21]
in PEMFCs. Direct contact with the organic solution may also
ChemSusChem 2015, 8, 288 – 300
www.chemsuschem.org
296
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Figure 10. SEM images of the surface of the cathode CL after hydrogenation
of maleic acid with SSWMs at varying a) concentrations of butanone (408C,
3
h), b) reaction temperatures (1m maleic acid, 3 h), and c) reaction times
(
1
408C, 1m maleic acid). General reaction conditions: 20 sccm hydrogen,
5 mLmin circulation fluid, 57 mAcm .
ꢀ
1
ꢀ2
swelling of the Nafion membrane and Nafion binder in organic
compounds has a great effect on the stability of the CL, the ra-
tional design of new PEM materials for electrochemical reac-
tors is necessary. Good swelling resistance and low permeabili-
ty of organic compounds across the membrane are required to
[20,24]
maintain a high efficiency of the PEM reactors.
Compared
with Nafion, non-fluorinated polymer electrolyte materials
often possess good swelling resistance to ketones and alco-
hols. For example, acetone and ethanol are good dispersants
of Nafion, but good precipitating reagents of some non-fluori-
nated polymer electrolytes, such as sulfonated poly(phthalazi-
none ether sulfone ketone) and quaterized poly(ether ether
ketone). The methanol permeability of non-fluorinated PEMs is
usually one order of magnitude lower than that of Nafion
Figure 9. SEM images of the surface of the cathode CL after hydrogenation
of butanone with carbon paper at varying a) concentrations of butanone
(408C, 3 h), b) reaction temperatures (1m butanone, 3 h), and c) reaction
times (408C, 1m butanone). General reaction conditions: 20 sccm hydrogen,
ꢀ
1
ꢀ2
1
5 mLmin circulation fluid, 57 mAcm .
[25]
exacerbate degradation of Pt/C compared with the PEMFC.
Thus, it is necessary to change the CCM after every reaction to
guarantee duplication of the data.
membrane. Based on these aspects, some non-fluorinated
polymer electrolytes can be investigated as alternatives. The
work is ongoing in our group.
Based on this analysis, we suggest that there is much room
for improvement to the MEAs used in PEM reactors for the cat-
alytic hydrogenation of bio-oil, especially when operated in
high concentration. Design of the hydrophilicity of DLs is es-
sential to achieve high efficiency of the PEM reactors, as dis-
cussed herein. Rational hydrophilicity modifications include
acidification or hybrid inorganic nanomaterials of traditional
carbon material DLs and application of hydrophilic materials,
Conclusions
The effects of hydrophilicity/-phobicity of the DL on the elec-
trocatalytic reduction of aqueous solutions of butanone and
maleic acid in PEM reactors were investigated. The commercial
hydrophobic carbon paper GDL did not allow permeation of
the aqueous solution at atmospheric pressure due to the high
surface energy of the aqueous solutions on the hydrophobic
[
23]
such as stainless-steel welded mesh and fiber felt. Because
ChemSusChem 2015, 8, 288 – 300
www.chemsuschem.org
297
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
surface of the GDLs. When the hydrogenation reactant was
volatile, such as butanone, the reactant could permeate
through the hydrophobic GDL in the form of vapor instead of
as a liquid, and the reaction rate increased with the air perme-
ability of the GDLs. When the reactant was nonvolatile, such as
maleic acid, the hydrogenation reaction barely occurred be-
cause of the very low reactant vapor pressure transferring
through the hydrophobic GDLs to the cathode CL. In this case,
a hydrophilic DL should be used to allow the aqueous solution
of nonvolatile compounds to flow through the DL in the form
of a liquid. This conclusion can be used for the catalytic hydro-
genation of aqueous solutions of bio-oil in PEM reactors. Be-
cause the chemical composition of bio-oils is complicated and
the pyrolytic oil includes about 35% nonvolatile compounds, it
is better to use a hydrophilic DL to ensure the hydrogenation
of more compounds; the protective effect of hydrophilicity to
the CL should also be taken into account.
Electrocatalytic hydrogenation of butanone and maleic acid were
carried out in a PEMFC reactor, which was a standard single cell
(
http://www.fuelcell-china.com.cn) with serpentine channels set in
2
graphite blocks (5.29 cm for geometric area). The DL for the
anode was always GDL35 BC, whereas different cathodes were
used, according to the experimental requirements. Viton or PET
gaskets with different thicknesses were used to seal the mem-
brane–graphite interface. Silicone rubber was used to seal the DL–
gasket interface to guarantee that no reactant leaked from it. The
hydrogenation setup is shown in Figure 1. Dry hydrogen at a flow
ꢀ
1
of 20 mLmin was maintained by a mass flow controller (Seven-
star D07-19B Anlog Mass Flowmeter), and was introduced to the
anode by passing through a humidifier. With an external potential
applied (GWInSTEK GPD-3303S DC Power Supply), hydrogen on
the anode CL was oxidized to give protons. The protons were
transported across the PEM and subsequently reduced to hydro-
gen atoms adsorbed on the cathode catalyst surface, which were
readily available for hydrogenation with the adsorbed liquid reac-
tant species. The side reaction was the recombination of excess ad-
sorbed hydrogen atoms into hydrogen gas again at the cathode.
The hydrogenation reaction was operated in a batch recycle mode.
The total volume of the cathode reactant was 100 mL, with a circu-
Commercially hydrophobic carbon paper and hydrophilic
SSWMs were used as DLs for the hydrogenation of aqueous
solutions of butanone and maleic acid, respectively. The PEM
reactors exhibited comparable performances with either of the
DLs at optimal current density, reactant concentration, and
temperature. The maximum reaction rate was reached at
ꢀ1
lation flow rate of 15 mLmin , unless otherwise stated. Because
the stability of the CL was indeterminate, a new CCM was used for
every reaction to guarantee comparability. In addition, the flow
rate of the cathode hydrogen was measured every hour with
a soap bubble flowmeter to check the total charge–product bal-
ance. The balances were closed with 100–115% for all experiments
reported herein.
ꢀ
2
ꢀ1
around 340 nmolcm
s
for both hydrogenation systems (at
ꢀ
2
a current density of 95 mAcm ), and the current efficiency
reached around 70%. The results were much better than those
reported in the literature. However, the CCM degraded during
the reaction because organic compounds could accelerate
swelling of the Nafion binder in the CL and the Nafion/PTFE
composite membrane, especially under high temperature and
concentration conditions. Non-fluorinated polymer electrolytes
might also be good alternatives for the CL binder and mem-
brane. By solving the problem of membrane swelling, PEM re-
actors will play a more important role in the catalytic hydroge-
nation of biomass derivatives.
Although the hydrogenation reactions of butanone and maleic
acid are spontaneous and will generate electricity when conducted
in a fuel cell model (standard open-circuit voltage of the hydroge-
nation reaction is 0.133 V for butanone (C=O) and 0.637 V for
[27]
maleic acid (C=C) in standard state), the measured voltages and
[6c]
currents are very small because of polarization of the cell. There-
fore, an external applied potential was used to accelerate the hy-
drogenation reaction. The focus of this paper was on the hydroge-
nation rate, which was related to the number of adsorbed hydro-
gen atoms (e.g., current density), so a tensiostatic operation was
used in all experiments to provide a comparable current for differ-
ent operating conditions.
Experimental Section
General
Cathode flow channel experiments
Butanone (99.5%, Fisher), 2-butanol (99.5%, Fisher), n-butanol
(
99.5%, Fisher), ethanol (99.5%, Fisher), cyclohexane (99.5%,
Fisher), maleic acid (99.5%, Guangfu, P.R. China), and succinic acid
99.5%, Guangfu, P.R. China). The MEA for PEM reactors consisted
To observe the form of cathode reactants (liquid or vapor) intui-
tively, cathode flow channel experiments were conducted. The
cathode flow channel was covered by a piece of DL (carbon paper
or SSWM), which was sealed on the graphite plate with silicone
rubber. Cathode liquid (100 mL), for example, aqueous solutions of
butanone or maleic acid, was pumped at circulating flow rates of
(
of a CCM sandwiched between two pieces of DLs. Commercial
CCM and DLs were used to guarantee duplication of the experi-
ments. The CCM (Sunrise Power, P.R. China) was fabricated by
spray-coating the catalyst ink onto both sides of the Nafion/PTFE
composite membrane (thickness of (18.0ꢂ2.0) mm). The catalyst
ꢀ1
15 and 80 mLmin (the maximum flow rate of the peristaltic
pump in our device) for 2 h. The solution and the cathode graphite
plate were both heated to 408C, and the MPL side of the GDL was
placed upward to simulate the reaction conditions. During the ex-
periment, a piece of glass slide was put above the DL to condense
the vapor released from the DLs, and thus, the mass transfer form
(liquid or vapor) of the reactant across the DL could be observed
intuitively. The concentration of the cathode circulating liquid after
2 h of circulation and the liquid drop on the glass slide were ana-
(
70 wt% Pt/C, Johnson Matthey) loadings of the anode and cath-
ꢀ2
[26]
ode were 0.2 and 0.4 mg cm , respectively. The electrochemi-
Pt
2
ꢀ1
cally active surface area of the cathodes was 55.8 m g , which
was provided by Sunrise Power Co. Hydrophobic GDLs were kindly
provided by SGL Tech., Germany, and consisted of carbon paper
with a MPL on one side. The products and their air permeability
are listed in Table 1. The hydrophilic DLs were commercial 316L
SSWMs (Anping, P.R. China), including 100, 200, 300, 400 meshes,
as listed in Table 1. SSWMs were ultrasonically cleaned in ethanol
for about 10 min before use.
1
lyzed by GC for butanone and H NMR spectroscopy for maleic
acid. The amount of reactant that evaporated through the GDL
could be calculated based on the concentration difference before
ChemSusChem 2015, 8, 288 – 300
www.chemsuschem.org
298
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
and after the experiment, and denoted as the percentage of evap-
oration.
ner, J. A. Dumesic, Science 2008, 322, 417–421; c) G. W. Huber, S. Iborra,
A. Corma, Chem. Rev. 2006, 106, 4044–4098; d) A. B. Lovins, E. K. Datta,
O.-E. Bustnes, J. G. Koomey, N. J. Glasgow, Winning the Oil Endgame: In-
novation for Profits, Jobs, and Security, Rocky Mountain Institute, Snow-
mass, 2005.
3] a) A. V. Bridgwater, G. V. C. Peacocke, Renewable Sustainable Energy Rev.
2000, 4, 1–73; b) S. Czernik, A. V. Bridgwater, Energy Fuels 2004, 18,
590–598; c) D. P. Mohan, U. Charles, Jr., P. H. Steele, Energy Fuels 2006,
Analytic methods
[
[
The liquid product of butanone hydrogenation was analyzed by
GC (FuLi GC-9790II), equipped with a 10% PEG-20000 packed
column (3 mꢁ3 mm; PEG=polyethylene glycol) and a flame ioni-
zation detector (FID). sec-Butanol was the only detected hydroge-
nation product in the liquid sample. Quantitative n-butanol was
added to the aqueous solution as an internal standard. GC was un-
suitable for the maleic acid hydrogenation system because the
boiling point of maleic acid and its hydrogenation product were
2
0, 848–889; d) D. C. Elliott, T. R. Hart, Energy Fuels 2009, 23, 631–637;
e) T. P. Vispute, H. Y. Zhang, A. Sanna, R. Xiao, G. W. Huber, Science 2010,
30, 1222–1227.
3
4] a) X. Wu, J. Benziger, G. He, J. Power Sources 2012, 218, 424–434; b) X.
Wu, G. He, L. Yu, X. Li, ACS Sustainable Chem. Eng. 2014, 2, 75–79; c) A.
Abdulla, K. Laney, M. Padilla, S. Sundaresan, J. Benziger, AIChE J. 2011,
5
7, 1767–1779; d) J. M. Sedlak, J. F. Austin, A. B. Laconti, Int. J. Hydrogen
1
very high. Instead, H NMR spectroscopy (Bruker Avance II 400M)
Energy 1981, 6, 45–51; e) C. Casati, P. Longhi, L. Zanderighi, F. Bianchi, J.
Power Sources 2008, 180, 103–113.
[28]
was used to detect the liquid products of maleic acid. Succinic
1
acid was the only product detected through H NMR spectroscopy.
[5] a) W. D. An, J. K. Hong, P. N. Pintauro, K. Warner, W. Neff, J. Am. Oil
Chem. Soc. 1999, 76, 215–222; b) W. D. An, J. K. Hong, P. N. Pintauro, K.
Warner, W. Neff, J. Am. Oil Chem. Soc. 1998, 75, 917–925; c) J. Benziger,
Ind. Eng. Chem. Res. 2010, 49, 11052–11060.
To detect the concentration of maleic acid in the cathode flow
channel experiment, quantitative succinic acid was added to the
circulating liquid as an internal standard substance. Gas products
of butanone and maleic acid hydrogenation were also analyzed by
using a GC-7890 TECHCOMP GC equipped with a thermal conduc-
tivity detector (TCD-01 packed column). Hydrogen was the only
product in the gas sample for both hydrogenation systems. Con-
tact angles of the DLs with different concentrations of butanone
and maleic acid were measured by using a Dataphysics OCA20 in-
strument (Germany). The size of the drop was 4 mL and the test
temperature was 25.98C. The measuring errors were less than 58
for carbon paper GDL and 88 for SSWM. Reaction rate, conversion,
and current efficiency were used as the evaluation criteria for the
hydrogenation reactions. Conversion could be easily obtained from
[
6] a) P. N. Pintauro, M. P. Gil, K. Warner, G. List, W. Neff, Ind. Eng. Chem. Res.
005, 44, 6188–6195; b) S. K. Green, J. Lee, H. J. Kim, G. A. Tompsett,
2
W. B. Kim, G. W. Huber, Green Chem. 2013, 15, 1869; c) S. K. Green, G. A.
Tompsett, H. J. Kim, W. B. Kim, G. W. Huber, ChemSusChem 2012, 5,
2
410–2420.
7] a) M. P. Dudukovi c´ , F. Larachi, P. L. Mills, Chem. Eng. Sci. 1999, 54, 1975–
995; b) M. P. Dudukovi c´ , F. Larachi, P. L. Mills, Catal. Rev. 2002, 44, 123–
246; c) U. K. Singh, M. A. Vannice, Appl. Catal. A 2001, 213, 1–24.
[
1
[8] L. Xin, Z. Zhang, J. Qi, D. J. Chadderdon, Y. Qiu, K. M. Warsko, W. Li,
ChemSusChem 2013, 6, 674–686.
[
9] a) X. Z. Yuan, Z. F. Ma, Q. Z. Jiang, W. S. Wu, Electrochem. Commun. 2001,
, 599–602; b) X.-Z. Yuan, Z.-F. Ma, Q.-G. He, J. Hagen, J. Drillet, V. M.
Schmidt, Electrochem. Commun. 2003, 5, 189–193.
10] W. An, J. K. Hong, P. N. Pintauro, J. Appl. Electrochem. 1998, 28, 947–
54.
11] R. Fonocho, C. L. Gardner, M. Ternan, Electrochim. Acta 2012, 75, 171–
78.
3
1
the GC spectrogram GC and H NMR spectrum. The reaction rate
[
[
[
was the average value during the reaction time and obtained on
9
ꢀ2
ꢀ1
the basis of either MEA geometric area (nmolcm s ) or weight of
ꢀ1
ꢀ1
Pt catalyst (mmolgPt min ). Current efficiency is defined by Equa-
tion (1), in which n is the number of electrons transferred, F is the
Faraday constant, I is the current observed in the experiment, and
t is the reaction time. Because the hydrogen permeation current of
1
12] a) EG&G Technical Services, Fuel Cell Handbook, 7th ed., U.S. Department
of Commerce, Sprinfield, VA, 2002; b) S. Park, J.-W. Lee, B. N. Popov, Int.
J. Hydrogen Energy 2012, 37, 5850–5865; c) S. Park, J.-W. Lee, B. N.
Popov, J. Power Sources 2006, 163, 357–363; d) L. Cindrella, A. M.
Kannan, J. F. Lin, K. Saminathan, Y. Ho, C. W. Lin, J. Wertz, J. Power Sour-
ces 2009, 194, 146–160; e) A. El-kharouf, T. J. Mason, D. J. L. Brett, B. G.
Pollet, J. Power Sources 2012, 218, 393–404.
ꢀ
2
the CCM is relatively small (2.6 mAcm at 808C, obtained from
Sunrise Power, P.R. China) compared with our current density, its
effect can be ignored.
[
[
[
13] J. Benziger, J. Nehlsen, D. Blackwell, T. Brennan, J. Itescu, J. Membr. Sci.
molproduct n F
Current efficiency ¼
ꢃ 100 %
ð1Þ
2
005, 261, 98–106.
I t
14] M. Garcia-Perez, A. Chaala, H. Pakdel, D. Kretschmer, C. Roy, Biomass Bio-
energy 2007, 31, 222–242.
15] a) T. Hottinen, M. Mikkola, T. Mennola, P. Lund, J. Power Sources 2003,
118, 183–188; b) J. G. Liu, G. Q. Sun, F. L. Zhao, G. X. Wang, G. Zhao,
L. K. Chen, B. L. Yi, Q. Xin, J. Power Sources 2004, 133, 175–180; c) P. Yi,
L. Peng, X. Lai, M. Li, J. Ni, Int. J. Hydrogen Energy 2012, 37, 11334–
Acknowledgements
We thank financial support from the National Science Founda-
tion for Distinguished Young Scholars of China (grant no.
11344.
[
16] U. Pasaogullari, C. Y. Wang, Electrochim. Acta 2004, 49, 4359–4369.
2
2
1125628), National Science Foundation of China (grant nos.
1176044 and 20976027), the Fundamental Research Funds for
[17] A. Arvay, E. Yli-Rantala, C. H. Liu, X. H. Peng, P. Koski, L. Cindrella, P. Kaur-
anen, P. M. Wilde, A. M. Kannan, J. Power Sources 2012, 213, 317–337.
[
18] Technical Information by Firsttenangstroms, “Contact Angle Measure-
ments on a Fine Metal Grid” to be found under http://www.firsttenang-
stroms.com/pdfdocs/CAOnMetalGrid.pdf, 2000.
the Central Universities of China (DUT13ZD204), and the Program
for Liaoning Excellent Talents in University.
[
19] a) J. Breen, R. Burch, K. Griffin, C. Hardacre, M. Hayes, X. Huang, S.
Obrien, J. Catal. 2005, 236, 270–281; b) F. Gao, R. Li, M. Garland, J. Mol.
Catal. A 2007, 272, 241–248; c) B. S. Akpa, C. D’Agostino, L. F. Gladden,
K. Hindle, H. Manyar, J. McGregor, R. Li, M. Neurock, N. Sinha, E. H. Stitt,
D. Weber, J. A. Zeitler, D. W. Rooney, J. Catal. 2012, 289, 30–41.
Keywords: biomass · diffusion layers · electrochemistry · fuel
cells · hydrogenation
[1] S. Chu, A. Majumdar, Nature 2012, 488, 294–303.
[20] K. A. Mauritz, R. B. Moore, Chem. Rev. 2004, 104, 4535–4585.
[
2] a) A. J. Ragauskas, C. K. Williams, B. H. Davison, G. Britovsek, J. Cairney,
C. A. Eckert, W. J. Frederick, J. P. Hallett, D. J. Leak, C. L. Liotta, J. R. Mie-
lenz, R. Murphy, R. Templer, T. Tschaplinski, Science 2006, 311, 484–489;
b) E. L. Kunkes, D. A. Simonetti, R. M. West, J. C. Serrano-Ruiz, C. A. Gart-
[21] a) Y. Shao, J. Wang, R. Kou, M. Engelhard, J. Liu, Y. Wang, Y. Lin, Electro-
chim. Acta 2009, 54, 3109–3114; b) Y. Shao, G. Yin, Y. Gao, J. Power Sour-
ces 2007, 171, 558–566; c) S. Zhang, X.-Z. Yuan, J. N. C. Hin, H. Wang,
K. A. Friedrich, M. Schulze, J. Power Sources 2009, 194, 588–600; d) S. D.
ChemSusChem 2015, 8, 288 – 300
www.chemsuschem.org
299
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Knights, K. M. Colbow, J. St-Pierre, D. P. Wilkinson, J. Power Sources
004, 127, 127–134.
22] a) P. Liu, G.-P. Yin, K.-D. Cai, Electrochim. Acta 2009, 54, 6178–6183; b) S.
F. Boschet, B. Ameduri, Prog. Polym. Sci. 2011, 36, 1521–1557; e) G.
Merle, M. Wessling, K. Nijmeijer, J. Membr. Sci. 2011, 377, 1–35.
[26] a) D. Xing, G. He, Z. Hou, P. Ming, S. Song, Int. J. Hydrogen Energy 2011,
36, 2177–2183; b) D. Xing, G. He, Z. Hou, P. Ming, S. Song, Int. J. Hydro-
gen Energy 2013, 38, 8400–8408.
2
[
[
[
[
Mu, C. Xu, Y. Gao, H. Tang, M. Pan, Int. J. Hydrogen Energy 2010, 35,
2872–2876.
23] a) L. Cindrella, A. M. Kannan, R. Ahmad, M. Thommes, Int. J. Hydrogen
Energy 2009, 34, 6377–6383; b) Y. Li, X. Zhang, L. Nie, Y. Zhang, X. Liu, J.
Power Sources 2014, 245, 520–528.
24] a) A. S. Aricꢂ, S. Srinivasan, V. Antonucci, Fuel Cells 2001, 1, 133–161;
b) C. Y. Du, T. S. Zhao, W. W. Yang, Electrochim. Acta 2007, 52, 5266–
[27] N. A. Lange, J. A. Dean in Lange’s Handbook of Chemistry, 11th ed., Vol. 9
(Ed.: J. A. Dean), McGraw-Hill, San Francisco, 1973.
[28] S. S. Vaghela, G. Ramachandraiah, P. K. Ghosh, D. Vasudevan, J. Appl.
Electrochem. 2002, 32, 1189–1192.
5271; c) X. Li, A. Faghri, J. Power Sources 2013, 226, 223–240.
25] a) X. Wu, G. He, S. Gu, W. Chen, P. Yao, J. Appl. Polym. Sci. 2007, 104,
1002–1009; b) X. Wu, G. He, S. Gu, Z. Hu, P. Yao, J. Membr. Sci. 2007,
Received: April 14, 2014
295, 80–87; c) X. Wu, W. Chen, X. Yan, G. He, J. Wang, Y. Zhang, X. Zhu,
Revised: July 16, 2014
J. Mater. Chem. A 2014, 2, 12222–12231; d) G. Couture, A. Alaaeddine,
Published online on October 15, 2014
ChemSusChem 2015, 8, 288 – 300
www.chemsuschem.org
300
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim