i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 3 2 0 e1 2 3 2 9

Available online at www.sciencedirect.com

ScienceDirect

journal homepage: www.elsevier.com/locate/he

Simple and functional direct methanol fuel cell

stack designs for application in portable and
auxiliary power units

O. Barbera a,*, A. Stassi a, D. Sebastian a, J.L. Bonde b, G. Giacoppo a,

a
C. D'Urso a, V. Baglio a, A.S. Arico
a
CNR e ITAE, National Research Council of Italy, Institute for Advanced Energy Technologies, Via Salita S. Lucia
sopra contesse, 598126 Messina, Italy
b
IRD Fuel Cells A/S, Kullinggade 31, DK-5700 Svendborg, Denmark

article info abstract

Article history: Simple and functional stack designs were developed for direct methanol fuel cells (DMFCs)
Received 5 November 2015 with the aim of reducing capital costs. A simplified planar and monopolar ministack, with a
Received in revised form reduced number of components, was designed for portable applications. This planar
3 May 2016 design facilitated the manufacturing of a compact passive mode operation ministack. The
Accepted 14 May 2016 ministack provided an output power of 1.30 W under air breathing mode operation, at
Available online 4 June 2016 room pressure and temperature without any auxiliary. For application in auxiliary power
units (APU), a self-heating and modular bipolar stack, operating up to 90
C without in-
Keywords: ternal cooling, was designed. Bipolar stacks consisted of 5e10 cells (100 cm2 active area),
Direct methanol fuel cell providing an output power of 90e180 W. The recirculated methanol solution was used as
DMFC thermostating fluid. The stacks achieved a normalised power density per cell of
Stack design 180 mW cm2. Whereas, for the passive mode operation, ministack approached
Portable application 30 mW cm2 with 5 M methanol feed. These results favourably compare with the best
APU systems performance reported in the literature for DMFCs operating under similar conditions.
Air breathing © 2016 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.

oxidation kinetics and high cross-over rate of methanol are

Introduction
Direct Methanol Fuel Cells have received worldwide attention
because of several intrinsic advantages, such as simplicity of
operation, high theoretical energy density, easy refuelling.
DMFCs can be simpler in construction than proton exchange
membrane fuel cells (PEMFCs), and do not need pressurised
hydrogen gas storage, delivery or processing. Additional
simplification in DMFCs includes air breathing operation for
low power application [1e9] However, the sluggish electro-
the main factors that have hindered the commercialization of
DMFCs.
In recent years, technology developments have promoted
interest around DMFC applications in small portable power
devices and APU systems [2]. Both small power, fast recharg-
ing, portable devices and hundred-watt-class DMFC systems,
designed to supply power for medium-size electrical applica-
tions, have been developed [2]. High durability for DMFC sys-
tems has been recently demonstrated using dynamic load

* Correspondiong author.
E-mail address: orazio.barbera@itae.cnr.it (O. Barbera).
http://dx.doi.org/10.1016/j.ijhydene.2016.05.135
0360-3199/© 2016 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 3 2 0 e1 2 3 2 9
profiles [10]. Portable power devices with high energy density
and easy recharging are becoming important for many elec-
tronic devices, such as notebook computers, power tools and
cellular phones.
Advances in the development of portable fuel cells should
have a large impact on the development of modern electronic
devices [1e9]. If for portable application there are some
limits in terms of operating temperature range, for auxiliary
power units (APU) systems, the high temperature operation
approach appears more appropriate being methanol oxidation
a sluggish process [11e15].
The stack designs for portable and APU applications are
different both in terms of architecture and operating condi-
tions. According to these premises, the power output and the
DMFC system around the stack can be quite different for
portable operating conditions and APU applications. The aim
of this work was to develop simple and functional designs for
two different configurations of DMFC stacks i.e. monopolar
and bipolar types [12]. The monopolar configuration, in com-
bination with a passive mode ambient operation, can be
addressed to portable applications whereas the bipolar
configuration operating at relatively high temperatures is
selected for APU applications.
The developed stacks were characterized by simplicity of
design, as needed to reduce capital costs. The devices pro-
vided excellent performance for the specific application. The
monopolar stack was assembled by placing electrodes of the
same polarity on the same side of the solid electrolyte
(membrane) and then electrically connecting the anodes and
cathodes in series [3e6]. The bipolar design, consisted of a
number of repeating units of membrane-electrode assemblies
(MEA) and bipolar plates [14e27]. In general, large size fuel cell
stacks are characterised by thermal management issues [11].
These are generally addressed using liquid fluid-fed cooling
cells, which are allocated between the electrochemical cells
inside the stack. Air cooled cells can represent another option.
In the present work, our approach regarded the use of a
methanol solution recirculated through the anode compart-
ment as cooling fluid. The bipolar plates contained also in-
ternal flow fields to facilitate air exchange through natural
convection. In general, this approach allows the use of an easy
cell sealing concept and favours a balance-of-plant simplifi-
cation thus avoiding additional thermal managing systems.
Experimental
The direct methanol fuel cell mini-stack consisted of two
main housings (strings) composed of 5 cells connected in se-
ries; the housings were connected in parallel. The single cell
area was 4.85 cm2. The operating conditions were: ambient
pressure and temperature, air breathing, and methanol
diffusion by natural convection. Two methanol concentra-
tions were selected (1 M and 5 M) corresponding to operating
conditions providing high fuel utilization or high energy
density. In this passive, monopolar, DMFC, 5 electrodes were
allocated on both sides of the same membrane, forming a
5-cell string. Two printed circuit boards (PCB) were chosen to
clamp and support an MEA and to electrically connect the five
electrodes by thin gold strips.
12321
The MEAs were prepared by hot-pressing at 130
C for
3 min and 20 kg cm2. The electrodes consisted of carbon cloth
backings, diffusion and catalytic layers. An unsupported PtRu
catalyst was employed at the anode; it was mixed with 15 wt%
Nafion ionomer (Ion Power, 5 wt% solution) and deposited by a
doctor blade technique onto a gas diffusion layer (0.36 mm in
thickness). Whereas, an unsupported Pt catalyst was utilized
for cathode fabrication; the catalyst was mixed with 15 wt%
Nafion ionomer (Ion Power, Nafion solution) and deposited by
a doctor blade onto gas diffusion layers (0.28 mm in thick-
ness). A Nafion 115 membrane was used as the electrolyte
[5,28e30].
Polarization curves for the ministack were obtained using a
Potentiostat-Galvanostat AUTOLAB PGSTAT302. The monop-
olar stack was connected to an AUTOLAB electrochemical
apparatus, and then the methanol reservoir was filled with
30 cc of 1 M or 5 M methanol solution with the cathode side
exposed to the ambient atmosphere. By imposing a dis-
charging current density, ranging from 0 to 140 mA cm2 with
a step of 10 mA cm2, the corresponding stack potential was
recorded. Chrono-potentiometric discharge curves for the
DMFC ministack were registered at the constant current of
0.6 A.
A direct methanol fuel cell stack based on a bipolar
configuration, was instead developed for APU applications. It
was operated with a gradient pressure between the cathode
and anode compartments. The anode was fed with 1 M or 5 M
methanol solution. The active electrode area was 100 cm2.
Two bipolar stacks were tested consisting of 5 or 10 cells. The
MEA used in the bipolar stack consisted of 5-layers: two
anodic and cathodic gas diffusion layers (GDLs) and a catalyst-
coated membrane (CCM, three layers). These MEAs had the
following characteristics: (i) Nafion 115 membrane, (ii) (1:1)
PtRu/C 1.8 mg cm2 (anode catalyst loading) and Pt/C
1.2 mg cm2 (cathode catalyst loading), (iii) SGL35DC GDL. A
thermocouple was allocated inside a cooling channel in the
center of the stack, facing the cathodic side, in close contact
with the composite graphite plate. This was used as reference
for the internal temperature.
Tests were carried out by connecting the fuel cell stack
to a Hydrogenics fuel cell test station. Polarisation curves
were acquired by increasing the current in steps from 0 to
50 A and recording the corresponding potential. Chrono-
potentiometric curves for the DMFC bipolar stack were regis-
tered at a constant current of 30 A. The methanol cross-over
rate was determined using a conventional DMFC cell [2],
operating under conditions similar to those of the stack here
investigated. Methanol permeated through the membrane
and oxidised to CO2 at the cathode was determined using gas-
chromatographic analysis (Varian CP-2003-QUAD micro GC).
Results and discussion
Methodology: ministack design and assembling for portable
applications
The DMFC ministack was designed for operation at ambient
pressure and temperature, under air breathing and using
methanol diffusion by natural convection. However, since
12322 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 3 2 0 e1 2 3 2 9
there was some methanol cross-over through the membrane
(see below), this caused direct chemical combustion with air
and the stack was self-heating [19]. Special stack structure is
required for passive diffusion of the liquid fuel and to provide
access to the ambient air [31,32]. A survey of different ap-
proaches used in this field is provided in the following. Baglio
et al. [2e5] developed a three cell passive DMFC stack using
PBC plates for clamping the MEA and collecting current; the
single cells were connected in series externally. To improve
the performance and decrease the internal ohmic resistance,
a Ti grid was interposed between the PCB and MEA. An alter-
native design for planar stacks, based on polycarbonate plates
and 304 Stainless Steel current collectors, was developed in
Ref. [19]. Also in this case, a series connection of three cells
was obtained by external wires; the device achieved a
maximum power density of 8.60 mW cm2. A mDMFC planar
stack [20] was developed using PMMA for fixer plates, silicon
rubber for gaskets and a metallic mesh for collecting current.
In this case, methanol was fed using a flow-through configu-
ration with different flow patterns. The device achieved a
maximum power density of 24.75 mW cm2. Passive DMFC
prototypes, in the Watt range, were assessed in Refs. [21e23].
D'Urso et al. developed a planar mDMFC [33] consisting of
polycarbonate layer. Gas-liquid feed holes were machined
with anodic and cathodic micro-channels, arranged in the
same plane. Pt-based catalysts were directly deposited on the
walls of the micro-channels. A modular DMFC stack, con-
sisting of four cells sharing the methanol reservoir and
externally connected in series, was proposed in Ref. [34]. The
maximum power density of the single module was
27 mW cm2. A DMFC fed with pure methanol was designed
and tested in Ref. [35]. Yingli et al. developed a passive direct
methanol fuel cell [36] where SS316L, photochemically etched,
was used as current collector and a TiN-plated stainless steel
mesh was inserted between the MEA and current collectors.
An internal series connection of the cell was adopted in this
case and a power density of 16.9 mW cm2 was achieved.
Ligang and al. evaluated the performance of a planar passive
DMFC stack [37]. An unsaturated polyester plate, with high
rigidity and insulation, was used to fix the current collectors.
These were made from 0.5 mm of 316L stainless steel mesh.
The surfaces were electro-deposited with Au to decrease
contact resistance and prevent corrosion. The peak power was
26.00 mW cm2. A tubular shape for the fuel cell was
numerically evaluated in Ref. [38].
From this survey, it is derived that in most of the previous
studies, the passive DMFC design was mainly based on plastic
materials for clamping plates and metallic meshes to collect
current. The general approach for the cell connection mainly
regarded the use of external wires. Only a few DMFC stacks
were conceived with internal connections in combination
with the flip-flop technique. However, despite the promising
approach, the performance achieved was modest. According
to these premises, the design of the present fuel cell ministack
was driven by the idea of simplifying its structure by reducing
and integrating the components and streamlining the
connections.
The architecture of the ministack was conceived as planar
and modular, with each pair of current collectors and
clamping plates combined on a single component. The aim
was to maintain or improve performance for this simplified
design with respect to the literature reports and to manufac-
ture prototypes in the Watt range. According to a numerical
simulation, aiming to minimise the stack volume, the active
area of the present stack was divided in 10 single cells of
4.85 cm2. The successive step was the definition of the base
unit, to be repeated n-times (modular approach). In Fig. 1, this
base unit, (b), is depicted with the end units, (a), and the final
arrangement of the plate, (c). The ministack was composed by
two strings of 5 cells each, with two MEAs arranged in a planar
configuration, with 5 active electrodes allocated on both sides
of a single membrane. Accordingly, the PCB approach was
considered the best solution because this material was stiff
enough to clamp and support the MEA. Moreover, an electrical
path can be obtained on the surface of the board by the photo-
etching technique (a standard technique used for electronic
devices production). The conductive pathways were gold
plated. The choice of the PCB avoided the use of metallic
meshes and combined in a single component both current
collection and clamping functions.
Thus, the anode and cathode of each cell were internally
connected. There were only two external connections for each
string (positive and negative). In Fig. 2a, a general scheme of
the series connection of single cells, in a planar stack, is given.
In this type of connection, the MEA is crossed by the current
collector. In Fig. 2b, the string design for the planar fuel cell
stack is shown. The overall MEA, with five electrodes placed
over the same membrane, is assembled between the two PCB
plates. To make a series electrical connection, the anodic pole
of the first MEA (1- in Fig. 2b) is connected to the cathodic pole
of the second MEA (2 þ in Fig. 2b) using an internal conductive
washer (blue circles in Fig. 2b) and so on.
The MEA was laser cut as shown in Fig. 2c. The assembling
approach is shown in Fig. 3a where gaskets and MEA are
represented in an exploded view.
In correspondence to each active area, for air and methanol
feeding, five windows were machined on the boards (Fig. 3a).
Two 250 mm thick PTFE gaskets were used to prevent re-
actants leakage (Fig. 3a). The two strings made of five cells
were in contact with the internal methanol reservoir at the
anode. The PTFE gaskets avoided fuel leakage from the
reservoir. An exploded view of the stack is shown in Fig. 3b.
Thus, both anode sides of the two stack strings were exposed
to the methanol solution, which was located in the central
reservoir. While the cathode sides were directly exposed to
the ambient air, which spontaneously diffused to the elec-
trodes by natural convection. Ministack tightening and MEA
compression were assured by using washers and tie-rods.
Fig. 4a shows the 3D concept of the stack.
At an operating electrochemical current density of
100 mA cm2, the fuel utilization was reduced by about 20%
due to methanol cross-over. This contributed to the stack self-
heating through the direct catalytic combustion of methanol
and air but it also reduced the fuel efficiency. The methanol
cross-over through the membrane was determined from ex-
situ measurements (see below). The internal temperature of
the stack was estimated to reach 40e50
C. In this design, the
specific configuration of the windows influences fuel cell
performance and heat exchange [3,4,9]. The diffusion win-
dows were designed with or without internal ribs or holes, and
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 3 2 0 e1 2 3 2 9 12323

Fig. 1 e Modular approach for designing the fuel cell ministack plates, (a), end units, (b), base units, (c), final configuration (5
base units and 2 end units).

Fig. 2 e (a), General scheme of the flip e flop connection, (b), internal pathways for series connection of cells, “A” and “C”
refer to anode and cathode, (c) MEAs and PTFE gasket, laser cut for screws allocation.

served to sustain and contact the MEA active area. This im-
plies that the fluid dynamics of the reactants and the distri-
bution of the clamping force are influenced by the design of
the ribs. The down-selected configuration is shown in Fig. 4.
The window is characterized by a specific ratio of 63% be-
tween the open area and MEA contacting surface. The larger
the open area the higher the diffusion of the reactants and
heat exchange. Whereas the larger the MEA contacting sur-
face the lower the ohmic resistance [26,27]. The final
arrangement of the planar stack is shown in Fig. 4b.
Methodology: bipolar stack design for auxiliary power units
Being the power requirements for the APU purposes signifi-
cantly larger than for portable applications, a bipolar
“stacked” configuration, providing low series resistance, was
chosen. The stack was equipped with 5 or 10 single cells
(100 cm2 each) to evaluate the validity of the modular design.
The cells were fed with reactants via internal manifolds
and a “Z” shape of internal fluids circuits was selected. Thus,
the inlet and outlet of the reactants were positioned on
opposite sides of the bipolar plates in a counter flow config-
uration. Bipolar plates were made of graphite, fuel and
oxidant were uniformly distributed over the electrode surface
by a flow-field machined on the front side of a bipolar plate
(Fig. 5a, b). Anodized aluminium plates with steel tie rods
(14  M6) were used to clamp the stack. High purity, copper
was used to manufacture the current collectors plates. These
plates were gold-plated. Internal and external leakage was
prevented using PTFE gaskets. To optimize the pressure losses
at the anodic and cathodic flow fields, these were designed
using different channel dimensions. The channel size was
(width  height) 1.10  1.20 mm at the anode and
1.50  0.50 mm at the cathode [16,26,27]. Fig. 5 shows the
cathodic (a) and anodic (b) plates concept. As a consequence of
methanol cross-over and irreversibility of the electrochemical
12324 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 3 2 0 e1 2 3 2 9
Fig. 3 e Assembling concept: (a), of a single string, (b) and
two strings with the reservoir in between.
process, the stack was self-heating, and the temperature was
regulated by cooling paths machined on the rear side of the
graphite plates. Such paths were not fed with a cooling fluid
but left open to the environment to facilitate heat exchange
through natural convection. Instead, the recirculating meth-
anol solution through the anode was used as thermostating
fluid. Thus, the device and balance of plant were considerably
simplified without significant decrease in performance.
Due to the compact design and the presence of several
stacked cells, the direct combustion process, produced by the
methanol permeation through the membrane and the
exothermicity of the electrochemical reaction, caused an
Fig. 4 e (a), 3D view of the assembled modular ministack, (b), the ministack under testing.
increase of temperature up to 80e90
C. The stack was virtu-
ally assembled (Fig. 5c), to check the components alignment
and to review the whole stack design. The external di-
mensions and mass of the stack were about 4.15 kg, with
the external dimension of 179  165  110 mm
(width  height  length). The stack was initially assembled
with 5 cells and subsequently with 10 cells; the device ready
for testing is shown Fig. 6a, where current collectors, graphite
plates and clamping plates can be observed. Fig. 6b shows the
MEA configuration as observed after stack operation. Fig. 6c
shows the stack under operation.
Monopolar ministack results
A polarisation curve for the monopolar ministack, operating
under passive mode with 5 M methanol solution corre-
sponding to a high energy density is shown in Fig. 7.
The total series resistance at ambient temperature for
this stack was about 263 mU corresponding to 0.51 U cm2.
The ohmic resistance appeared reasonable for a Nafion 115
membrane-based fuel cell operating at room temperature [31].
This was not significantly different from that observed in the
literature for a DMFC single cell operating under passive mode
[2]. Thus, the contribution of the monopolar set-up to the
ohmic resistance appeared not relevant.
The fuel cell ministack achieved a maximum power of
1.3 W at 1.3 A, with a corresponding mean power density of
27 mW cm2 at 137 mA cm2. Due to the connection in par-
allel, the corresponding current for each string in the minis-
tack was 0.66 A.
The best single cell in the ministack showed a maximum
power density of about 30 mW cm2. The normalised perfor-
mance appears better than that reported in the literature
[7,29,34,38,39]. This could derive from a suitable combination
of simple stack design and stack components, facilitating
natural convection of reactants to the active catalyst sites and
easy removal of reaction products. A distribution of cell po-
tentials for the two stack strings is reported in Fig. 8.
Some variation in the distribution of cell voltages was
recorded as consequence of methanol convective flow. Gal-
vanostatic discharging curves, at 0.6 A with 5 M methanol
concentration, showed a voltage drop at about 3 h, instead of a
4 h discharge time as expected from the Faraday's law. The
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 3 2 0 e1 2 3 2 9
Fig. 5 e Cathodic (a), and anodic (b), graphite bipolar plates concept view, (c), bipolar stack virtual assembling.
Fig. 6 e (a), Manufactured 5-cells stack ready for testing, (b), photograph of a 5-layer MEA after operation during stack
dismantling, (c), 10-cells stack under operation.
fuel utilisation was about of 75%, as consequence of cross-
over. In the case of 1 M methanol concentration the
discharge time was 0.67 h vs. 0.8 h (theoretical) with a fuel
utilisation approaching 85% (Fig. 9a). However, the perfor-
mance with 1 M methanol was significantly lower due to mass
transport constraints and lower stack internal temperature,
as a consequence of the lower methanol cross-over (Fig. 9b).
This clearly indicates that a proper compromise for ministack
operation in terms of methanol concentration is necessary.
Thus, the higher internal temperature and the reduced
convective mass transport constraints, in the presence of
Fig. 7 e Polarisation curve for the monopolar ministack
under passive mode operation (room temperature, air
breathing, MeOH 5 M, natural convection).
12325
higher methanol concentration, resulted in a significantly
better performance operation with 5 M vs. 1 M, as shown in
Fig. 9. Depending on the specific requirements for portable
applications, different methanol concentration can be used to
get larger fuel utilisation or better power output.
Methanol cross-over is essentially related to the methanol
concentration, membrane properties, operating temperature
current density and MEA/electrode characteristics [40]. Since
it was not possible to precisely determine the cross-over of
methanol directly in the passive mode ministack, this was
determined in a conventional DMFC cell [40], operating under
similar conditions and using the same MEA type. Fig. 10 shows
the variation of cross-over with temperature for the two
methanol concentrations used in this work.
At 30
C with 5 M methanol concentration, the cross-over
rate reaches 1.56 105 mol cm2 min1 which is significantly
larger than the cross-over registered with 1 M methanol
concentration and it varies with temperature. The esti-
mated values of methanol permeation in the conventional
cell do not match exactly with that estimated for the passive
mode stack from chronopotentiometric curves. Effects such
as spontaneous diffusion vs. forced convection, operating
current density (methanol concentration gradient at the
interface) as well as evaporation effects may affect this
aspect. However, the main effects of temperature and con-
centration on methanol cross-over are well evident from
Fig. 10.
Thus, the internal temperature of the stack increases with
the larger concentration of methanol fed to the anode,
12326 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 3 2 0 e1 2 3 2 9

Fig. 8 e Distribution of cell potentials for the two strings of the stack at different current loads.

Fig. 9 e (a), Chrono-potentiometric discharge curves for the ministack at a current load of 0.6 A (passive mode operation), for
two different methanol concentrations (1 M and 5 M), (b), polarisation curves for the DMFC stack in the presence of two
methanol feed concentrations (1 M and 5 M), recorded before the discharging tests.

producing a better performance but resulting in a significant The 10-cell stack showed slightly better performance in the
decrease of fuel utilisation. activation region, probably due to the fact that the effective
temperature under operation was a few degrees higher for
this stack. Whereas the 5-cell stack was slightly better per-
Bipolar stack experimental results
forming at high current density. The normalised series resis-
tance slightly increased from 0.14 U cm2 to 0.16 U cm2 passing
Fig. 11a shows a comparison of the polarisation behaviour of
from 5 cells to 10 cells. These results show that the bipolar
the 5-cell and 10-cell stack. Minimal differences were recor-
stack was characterised by appropriate modular properties.
ded. In both cases, the internal temperature reached almost
However, a liquid cooling approach may be necessary for a
90
C. However, it was slightly larger (90
C vs. 88
C) for the
stack larger than 10 cells, especially if operating at high cur-
10-cell stack.
rent density. Alternatively, high temperature membranes
capable of operating in a wide temperature range and char-
-4
10 acterised by rapid start-up would be very appropriate. A short
1M CH3OH
-1

time test, with few start-up and shut down cycles, was carried
Methanol cross-over / mol cm min

5M CH3OH out by imposing a constant current of 30 A (Fig. 11b); the 5-cell

-2

bipolar stack delivered an average power of about 70.0 W at a

mean voltage of 2.3 V. During operation, some recoverable
voltage losses were registered. This was interpreted as due to
-5 both an increased coverage of CO-like species on the anode
10
surface and the cathode poisoning by cross-over. These losses
were almost recovered after each shut-down/start-up cycles.
Fig. 12 shows the distribution of cell voltages in the 10-cell
stack, measured at 10 and 20 A of current load. The cell voltage
distribution appeared quite homogeneous in the bipolar stack.
-6
However, the central cells were operating slightly better. This
10
30°C 60°C 90°C aspect may be related to the effect of compression. Moreover,
Temperature / °C this trend reflects the temperature profile along the stack. Due
to the exothermic reactions, the increase of temperature is
Fig. 10 e Methanol cross-over determined ex-situ as a larger for the central cells, being the terminal cells in contact
function of temperature and concentration for MEAs with the “cold” clamping plates. Specific experiments were
equivalent to those used in the present DMFC stacks. made to check the effect of methanol feed concentrations
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 3 2 0 e1 2 3 2 9 12327

Fig. 11 e (a), comparison between 5-cell (solid) and 10-cell stacks (dashed) polarisations, (b) time-test at 30 A with shut-
down/start-up cycles for the 5-cells stack, operating under self-heating mode and fed with 5 M Methanol. Stoichiometry:
l ¼ 4 for fuel (recirculated solution) l ¼ 4 for oxidant.

0.60 (Fig. 10). However, due to the gas flow at the cathode and the
0.50 recirculation of methanol solution at the anode, the difference
Cell Potential / V

in internal temperature was not much different in the two

0.40 cases. The best performance was achieved with 5 M methanol
0.30 especially at high current densities where the higher con-
centration resulted in lower mass transport constraints
0.20
(Fig. 13). According to cross-over data (Fig. 10), fuel utilisation
0.10 is certainly better with 1 M methanol solution. Thus, for
operation at low current density i.e. in the high voltage effi-
0.00
ciency region, the low methanol concentration is preferred. In
1 2 3 4 5 6 7 8 9 10
Cell number 10 A 20 A fact, no relevant difference is observed for the two concen-
trations in this region (Fig 13).
Fig. 12 e Distribution of cell voltages at different current Some applications, e.g. stationary type, do not strictly
load. require compact systems but a high fuel efficiency. A different
situation may occur for application of DMFC generators in
telecommunications or as APU, in campers and small boats. In
using similar recirculation rate (Fig. 13). If 1 M methanol this case, the DMFC systems could be more appealing if
concentration gives rise to a lower methanol consumption by characterised by high power and compact size.
effect of a lower cross-over, a 5 M methanol concentration is The simplified stack designs developed in this work are
instead more appropriate for fuel management in compact characterised by a reduced number of components and simple
DMFC systems. This is due to a proper increase in energy assembling properties; this results in compact devices with
density, with consequent reduction in weight and volume of excellent performance. In this regard, these designs offer a
methanol tanks. significant reduction in capital costs. The self-heating char-
At 90
C, the methanol cross-over with 5 M solution was acteristics, natural reactants convection for the passive mode
significantly larger than that recorded for the 1 M solution operation, simplified cooling strategy for the bipolar stack,
contribute to system simplification and provide cost-savings
in terms of operating expenditures.

Conclusions

Two types of DMFC short stacks characterised by functional

Fig. 13 e Performance comparison for the DMFC bipolar
stack in the presence of two methanol feed concentrations.
design and simplified operation characteristics were
developed. The approach was to use a minimum number of
components for stack assembling, to provide concrete per-
spectives to reduce capital costs and operating expenditures.
A compact design for monopolar configuration operating at
low temperature under passive mode was demonstrated. This
configuration was selected as proof-of-concept for portable
applications. It provided a normalised power density of about
30 mW cm2 under passive mode. A bipolar stack, with bipolar
plates made of graphite, was designed and manufactured for
12328 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 3 2 0 e1 2 3 2 9
high temperature operation, especially for application in
auxiliary power units. In this case, the operating temperature
was around 90
C and the normalised power density
approached 180 mWcm2. The electrochemical tests showed
that the simplified design was accompanied by an excellent
power density for both stacks in relation with the specific
applications. The specific design and the self-heating mode
may allow for a simplification of the DMFC system. Simplicity
of operation makes these designs useful for portable and APU
applications.
Acknowledgements
The research leading to these results has received funding
from the European Community's Seventh Framework Pro-
gramme (FP7/2011e2014) for the Fuel Cells and Hydrogen Joint
Technology Initiative under grant agreement DURAMET
n
278054.
references
[1] Liu H, Zhang J, editors. Electrocatalysis of direct methanol
fuel cells. Weinheim, Germany: Wiley-VCH Verlag GmbH &
Co. KGaA; 2009.
[2] Dillon R, Srinivasan S, Arico AS, Antonucci V. International
activities in DMFC R&D: status of technologies and potential
applications. J Power Sources 2004;127:112e26.
[3] Baglio V, Stassi A, Matera FV, Di Blasi A, Antonucci V,
Arico AS. Optimization of properties and operating
parameters of a passive DMFC mini-stack at ambient
temperature. J Power Sources 2008;180:797e802.
[4] Baglio V, Stassi A, Matera FV, Antonucci V, Arico  AS.
Investigation of passive DMFC mini-stacks at ambient
temperature. Electrochimica Acta 2009;54:2004e9.
[5] Baglio V, Stassi A, Matera FV, Kim H, Antonucci V, Arico  AS.
AC-impedance investigation of different MEA configurations
for passive-mode DMFC mini-stack applications. Fuel Cells
2010;1:124e31.
[6] Hsieh S, Chen I, Hwong C. Development of a four-cell DMFC
stack with air-breathing cathode and dendrite flow field. Int J
Energy Res 2014;38:1693e711.
[7] Ghayor R, Shakeri M, Sedighi K, Farhadi M. Experimental and
numerical investigation on passive and active mDMFC. Int J
Hydrogen Energy 2010;35:9329e37.
[8] Kamarudin SK, Achmad F, Daud WRW. Overview on the
application of direct methanol fuel cell (DMFC) for portable
electronic devices. Int J Hydrogen Energy 2009;34:6902e16.
[9] Sundarrajan S, Allakhverdiev SI, Ramakrishna S. Progress
and perspectives in micro direct methanol fuel cell. Int J
Hydrogen Energy 2012;37:8765e86.
[10] Kimiaie N, Wedlich K, Hehemann M, Lambertz R, Müller M,
Korte C, et al. Results of a 20 000 h lifetime test of a 7 kW
direct methanol fuel cell (DMFC) hybrid system e
degradation of the DMFC stack and the energy storage.
Energy Environ Sci 2014;7:3013e25.
[11] Cai W, Li S, Li C, Liang L, Xing W, Liu C. A model based
thermal management of DMFC stack considering the double-
phase flow in the anode. Chem Eng Sci 2013;93:110e23.
[12] Kim D, Lee J, Lim TH, Oh IH, Ha HY. Operational
characteristics of a 50W DMFC stack. J Power Sources
2006;155:203e12.
[13] Buttin D, Dupont M, Straumann M. Development and
operation of a 150 W air-feed direct methanol fuel cell stack.
J Appl Electrochem 2001;31:275e9.
[14] Argyropoulos P, Scott K, Taama W. An investigation of
scale-up on the response of the direct methanol fuel cell
under variable load conditions. J Appl Electrochem
2001;31:13e24.
[15] Squadrito G, Barbera O, Giacoppo G, Urbani F, Passalacqua E.
Computer aided fuel cell design and scale-up, comparison
between model and experimental results. J Appl Electrochem
2006;37:87e93.
[16] Giacoppo G, Barbera O, Carbone A, Gatto I, Sacca  A,
Pedicini R, et al. 1.5 kWe HT-PEFC stack with composite
MEA for CHP application. Int J Hydrogen Energy
2013;38:11619e27.
[17] Bates A, Mukherjee S, Hwang S, Lee SC, Kwon O, Choi GH,
et al. Simulation and experimental analysis of the clamping
pressure distribution in a PEM fuel cell stack. Int J Hydrogen
Energy 2013;38:6481e93.
[18] Yu B, Yang Q, Kianimanesh A, Freiheit T, Park SS, Zhao H,
et al. A CFD model with semi-empirical electrochemical
relationships to study the influence of geometric and
operating parameters on DMFC performance. Int J Hydrogen
Energy 2013;38:9873e85.
[19] Martin JJ, Qian W, Wang H, Neburchilov V, Zhang J,
Wilkinson DP, et al. Design and testing of a passive planar
three-cell DMFC. J Power Sources 2007;164:287e92.
[20] Liu X, Zhang B, Zhang Y, He H, Li J, Wang S, et al.
Development and characterization of a novel air-breathing
micro direct methanol fuel cell stack for portable
applications. J Micromechanics Microengineering
2010;20:1e11.
[21] Guo Z, Faghri A. Development of a 1 W passive DMFC. Int
Commun Heat Mass Transf 2008;35:225e39.
[22] Yuan Z, Zhang Y, Leng J, Gao Y, Liu X. Development of a
4-cell air-breathing micro direct methanol fuel cell stack. J
Power Sources 2012;202:134e42.
[23] Guo Z, Faghri A. Development of planar air breathing direct
methanol fuel cell stacks. J Power Sources 2006;160:1183e94.
[24] Zhao TS, Chen R, Yang WW, Xu C. Small direct methanol fuel
cells with passive supply of reactants. J Power Sources
2009;191:185e202.
[25] Vijayakumar R, Rajkumar M, Sridhar P, Pitchumani S. Effect
of anode and cathode flow field depths on the performance
of liquid feed direct methanol fuel cells (DMFCs). J Appl
Electrochem 2012;42:319e24.
[26] Lu Y, Reddy RG. Effect of flow fields on the performance of
micro-direct methanol fuel cells. Int J Hydrogen Energy
2011;36:822e9.
[27] Kim JH, Yang MJ, Park JY. Improvement on performance and
efficiency of direct methanol fuel cells using hydrocarbon-
based membrane electrode assembly. Appl Energy
2014;115:95e102.
[28] Zhang Y, Zhang P, He H, Zhang B, Yuan Z, Liu X, et al. A self-
breathing metallic micro-direct methanol fuel cell with the
improved cathode current collector. Int J Hydrogen Energy
2011;36:857e68.
[29] Zainoodin AM, Kamarudin SK, Masdar MS, Daud WRW,
Mohamad AB, Sahari J. Investigation of MEA degradation in a
passive direct methanol fuel cell under different modes of
operation. Appl Energy 2014;135:364e72.
[30] Zainoodin AM, Kamarudin SK, Masdar MS, Daud WRW,
Mohamad AB, Sahari J. High power direct methanol fuel cell
with a porous carbon nanofiber anode layer. Appl Energy
2014;113:946e54.
[31] Mallick RK, Thombre SB, Shrivastava NK. A critical review of
the current collector for passive direct methanol fuel cells. J
Power Sources 2015;285:510e29.
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 3 2 0 e1 2 3 2 9
[32] Yousefi S, Zohoor M. Conceptual design and statistical
overview on the design of a passive DMFC single cell. Int J
Hydrogen Energy 2014;39:5972e80.
[33] D'Urso C, Baglio V, Antonucci V, Arico AS, Specchia S,
Icardi UA, et al. Development of a planar mDMFC operating at
room temperature. Int J Hydrogen Energy 2011;36:8088e93.
[34] Wang L, He M, Hu Y, Zhang Y, Liu X, Wang G. A ‘4-cell’
modular passive DMFC (direct methanol fuel cell) stack for
portable applications. Energy 2015;82:229e35.
[35] Li X, Faghri A. Development of a direct methanol fuel cell
stack fed with pure methanol. Int J Hydrogen Energy
2012;37:14549e56.
[36] Zhu Y, Liang J, Liu C, Ma T, Wang L. Development of a passive
direct methanol fuel cell (DMFC) twin-stack for long-term
operation. J Power Sources 2009;193:649e55.
12329
[37] Feng L, Cai W, Li C, Zhang J, Liu C, Xing W. Fabrication and
performance evaluation for a novel small planar passive
direct methanol fuel cell stack. Fuel 2012;94:401e8.
[38] Ward T, Xu C, Faghri A. Performance and design analysis of
tubular-shaped passive direct methanol fuel cells. Int J
Hydrogen Energy 2011;36:9216e30.
[39] Obeisun OA, Meyer Q, Robinson J, Gibbs CW, Kucernak AR,
Shearing PR, et al. Development of open-cathode polymer
electrolyte fuel cells using printed circuit board flow-field
plates: flow geometry characterisation. Int J Hydrogen Energy
2014;39:18326e36.
 AS, Sebastian D, Schuster M, Bauer B, D'Urso C,
[40] Arico
Lufrano F, et al. Selectivity of direct methanol fuel cell
membranes. Membranes 2015;5:793e809.