Electrochemical and Solid-State Letters, 14 (6) B51-B54 (2011)
B51
C The Electrochemical Society
1099-0062/2011/14(6)/B51/4/$28.00 V
Through-Plane Water Transport Visualization in a PEMFC by
Visible and Infrared Imaging
M. M. Daino,a,* Z. Lu,a,c,** J. M. LaManna,a,d J. P. Owejan,b T. A. Trabold,b,e
and S. G. Kandlikara,**,z
a
b
Mechanical Engineering Department, Rochester Institute of Technology, Rochester, New York 14623, USA
GM Electrochemical Energy Research Laboratory, Honeoye Falls, New York 14472, USA
In this study, water transport and thermal profile in the through-plane direction of a proton exchange membrane fuel cell (PEMFC) gas
diffusion layer (GDL) are reported. Direct optical and infrared access to both cathode and anode GDLs are provided in a typical 50
cm2 test section. Dynamic visualization (pixel resolution of 0.6 lm at 28 Hz) of liquid water transport and emergence in the gas distribution channels and diffusion layers are reported and the underlying transport processes are discussed. The temperature distributions
across the anode and cathode GDL are also measured with a high resolution infrared camera with a pixel resolution of 5 lm at 30 Hz.
C 2011 The Electrochemical Society.
V
[DOI: 10.1149/1.3560163] All rights reserved.
Manuscript submitted January 19, 2011; revised manuscript received February 8, 2011. Published March 10, 2011. This was Paper
713 presented at the Las Vegas, Nevada, Meeting of the Society, October 10–15, 2010.
Water management in proton exchange membrane fuel cells
(PEMFCs) is widely studied due to its impact on performance.1–4 A
variety of imaging techniques have been employed to monitor liquid
water, generally in the plane of the active area (in-plane). The most
prominent techniques include (a) visible imaging,5 which has high
spatial3 (<1 lm) and temporal resolutions3 (>1 kHz) but requires
modifications to test section materials for optical access, and (b) neutron
radiography,6 which can quantify the amount of water in actual PEMFCs
but current spatial and temporal resolutions are significantly lower. Infrared imaging has been applied to PEMFCs (at somewhat lower resolution
than visible) but is typically used to obtain in-plane temperature distributions on the gas diffusion layer (GDL) surface7,8 and end plates.9 An
in-depth review of these techniques is given elsewhere.10
With recent improvements in neutron detector technology enabling resolution of finer length scales (20 lm), neutron methods
are being applied to the through-plane dimension.11 Several other
techniques for measurement of water distributions are summarized
by St.-Pierre.12 Among them, x-ray and nuclear magnetic resonance
imaging and various spectroscopy methods are hindered by spatial
resolution, temporal resolution, or incompatibility with typical test
section materials. As a consequence, researchers must increase the
through-plane thickness of each layer, run experiments to provide
time-averaged measurements, and=or change the inherent properties
of the fuel cell assembly.
The limitations of these diagnostic methods combined with the
imperative need to develop robust GDL mass transport have motivated development of more advanced imaging techniques. Through
cross-sectional imaging at visible and infrared wavelengths, Ishikawa et al. investigated freezing phenomena under shutdown conditions.13,14 In the current work we expand upon this method to image
liquid water formation within the GDL and discuss the mechanisms
by which droplets are transferred to the gas channels. Additionally,
high resolution infrared images elucidate the through-plane temperature distribution, a parameter that must be well characterized to
predict condensation within the GDL.
Experimental
PEMFC design.— A PEMFC was designed to allow visible and
infrared transmission through a window revealing the cross-sections
of the cathode and anode GDLs and channels. The graphite flow
* Electrochemical Society Student Member.
** Electrochemical Society Active Member.
c
Present address: Ford Motor Company, Dearborn, Michigan 48126, USA.
d
Present address: Department of Mechanical, Aerospace and Biomedical
Engineering, University of Tennessee, Knoxville, Tennessee 37996, USA.
e
Present address: Center for Sustainable Mobility, Rochester Institute of
Technology, Rochester, New York 14623, USA.
z
E-mail: [email protected]
fields of a commercial 50 cm2 fuel cell were modified by recessing
the lands 180 lm to represent the compressed thickness of the
GDLs. Sapphire was chosen as the window material for its transmittance of visible and mid-wave infrared light and a more realistic
thermal boundary condition (k 25 W=m K) than the typically
used glass (k 1 W=m K). A pocket was machined into the side of
the flow fields to hold the sapphire flush with the unmodified perimeters. A schematic of the assembled cell is shown in Fig. 1a with a
close-up shown in Fig. 1b. Additional details of the fuel cell design
and experiment are discussed in a previous publication.15
PEMFC materials and testing conditions.— The catalyst coated
membrane (CCM) was based on an 18 lm thick Gore membrane
with 0.3 mg=cm2 Pt loading on both sides. The GDL material was
MRC (Mitsubishi Rayon Co.) 105 treated with polytetrafluoroethylene (PTFE) and coated with an in-house microporous layer (MPL)
by General Motors. In all experiments, dry, clean hydrogen and air
were supplied in a counter flow configuration; no external humidification was provided for either flow. The fuel cell was operated at
ambient temperature and pressure under a constant stoichiometric
ratio of An=Ca ¼ 2.0=3.0. The preliminary test revealed a comparable performance (IV curve) to the standard 50 cm2 cell, indicating
that the sapphire windows had little influence on performance.
Results and Discussion
Simultaneous anode and cathode channel visualization.—
Although cathode and anode flow channels have been visualized previously, this is typically done in separate experiments.16,17 In this
work, the simultaneous visualization of anode and cathode channels
was obtained at various magnifications. Figures 2a–2d show the
cross-sectional view under loads of 0.05, 0.30, 0.70, and 1.10 A=cm2,
respectively. Images at higher current densities were excluded due to
increased cell temperature and the change of the phase transition into
the channels. Several important observations can be made from this
figure. First, under lower current densities, only a small amount of
water accumulated in the cathode channels, while there was a significantly greater amount of water in the form of film along the anode
channels. This is in agreement with the neutron radiography studies
of Owejan et al.18 Secondly, at a given current density, the greatest
amount of condensation in the anode channels occurred towards the
inlet although both inlet streams had no external humidification. The
anode water appeared as condensation on the sapphire window and
under no operating conditions did droplets emerge from the anode
GDL as was routinely observed in the cathode. Additionally, the
location of the anode condensation migrated toward the outlet as a
function of current density. The condensed water in the anode was
periodically removed by water slugs in the channels (see Fig. 2b).
Lastly, liquid water appears in anode channels mostly via condensation mechanism instead of liquid transport through the GDL.
Downloaded 10 Mar 2011 to 198.208.159.20. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp
B52
Electrochemical and Solid-State Letters, 14 (6) B51-B54 (2011)
A=cm2, water droplets emerged from the GDL and wicked onto the
window and channel walls where they were removed by gas shear.
At one droplet formation site, small droplets appeared to be condensing on the GDL cross-section. To isolate the small droplets, the
cell was returned to OCV and purged on the cathode until all water
was removed. The load was then returned to 0.4 A=cm2 and the
cross-section was continuously monitored. The condensing small
droplets returned to the site within a minute. In order to isolate the
condensation, the video was processed using image subtraction and
subsequent thresholding. An initial color image of the dry GDL
cross-section (Fig. 3a, obtained after purge) was subtracted from
each subsequent frame (e.g., see Fig. 3b), resulting in a difference
sequence. A minimum change in intensity (threshold) was applied
to the difference sequence to identify potential water detection locations. These pixels were defined as water detection locations (Fig.
3c) if potential detection occurred independently on at least two
color planes. Finally, to give spatial reference (w.r.t. the GDL crosssection) to the binary water detection frames, the initial dry grayscale image was added to the water detection frames (Fig. 3d).
Example images of dry, wet, water detection, and water detection
with the GDL cross-section are shown in Figs. 3a–3d, respectively.
These results demonstrate that the water droplets and transport on
the cross-section of the GDL can be captured with a digital microscope and quantified with video processing. More interesting is that
water condensation inside the GDL is implied. However, in the current design, the MPL and catalyst layer could not be observed, and
thus no implicit conclusion on the location of the condensation site
or the consequent condensation front were obtained. The design of a
new setup allowing the observation of the MPL at higher magnification is under way.
Figure 1. (Color online) (a) Modified cathode flow field to accomodate the
compressed thickness of the GDL and placement of the sapphire window
and (b) Schematic of assmbled PEMFC with close-up of cathode and anode
cross-sections.
A similar phenomenon has been reported19 at a higher temperature
of 80 C. This indicates that water flooding and the mass transport
resistance in the anode are not as serious as in the cathode because
liquid water does not accumulate inside the porous GDL and block
hydrogen pathways. In contrast, liquid water in cathode channel
appears through both condensation mechanism and liquid water
transport through the GDL.
GDL cross-section observation.— In an additional experiment, in
order to increase the cross-section of the observed GDL, an additional layer of Toray TGP-H-060 10 wt % PTFE was used on the
channel side of the cathode MRC GDL. At a current density of 0.4
Cross-sectional observation with high resolution infrared
imaging.— The use of visible light microscopes to observe water is
further complicated by its transparency. There is thus a clear need
for other techniques to improve visualization of water. In this study,
an infrared imaging technique was applied.
Figure 4a shows an example IR image of the PEMFC at a current
density of 0.8 A=cm2 with MRC GDLs. A uniform emissivity of 0.9
(determined from separate calibration experiments for GDL) was
used (the color bar is in C). Note that water film and condensed
droplets are clearly observed in the anode and cathode channels,
respectively. This indicates that infrared imaging could be used to
identify water within an operating fuel cell at the microscale.
The temperature gradient across the GDLs shown in Fig. 4a can
be clearly illustrated by using a digital smoothing filter, averaging
the temperatures in the in-plane direction, and mapping them to a
color spectrum. The result is shown in Fig. 4b. Note the unexpected
result of a greater temperature gradient across the anode compared
to the cathode GDL. Additionally, neither GDL showed a significant
temperature gradient. Furthermore, the GDL gradients can be
Figure 2. (Color online) Cross-section of
transparent PEMFC as a function of current density: (a) j ¼ 0.05 A=cm2, (b)
j ¼ 0.30 A=cm2, (c) j ¼ 0.70 A=cm2, (d)
j ¼ 1.10 A=cm2.
Downloaded 10 Mar 2011 to 198.208.159.20. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp
Electrochemical and Solid-State Letters, 14 (6) B51-B54 (2011)
B53
Figure 3. (Color online) Visualization of
water on the cathode GDL cross-section at
j ¼ 0.4 A=cm2: (a) the dry initial image,
(b) image with water present, (c) water
detection locations, and (d) water detection with GDL.
Figure 4. (Color online) (a) Infrared
image of a PEMFC cross-section at 0.8
A=cm2 and (b) Color infrared image of
the temperature distribution across anode
and cathode GDLs in a PEMFC at 0.8
A=cm2.
Downloaded 10 Mar 2011 to 198.208.159.20. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp
B54
Electrochemical and Solid-State Letters, 14 (6) B51-B54 (2011)
affected by the amount of liquid water present. The water that was
routinely observed on the cathode GDL (Fig. 2) could have masked
the true temperature gradient. Since water was never observed on
the anode GDL, it is not expected to affect the temperature profile.
Conclusion
The main contribution of this work was the visualization of liquid water and measurement of the temperature profile across both
cathode and anode GDLs in the through-plane direction. Visible
observations with a digital microscope revealed anode GDL water
transport is dominantly in the vapor phase as opposed to the cathode
GDL where liquid water was routinely observed. High magnification observations of the cathode GDL revealed the condensation of
small droplets and implied the existence of condensation within the
GDL but could not be confirmed due to current limits on test section
design. Development of a new test section to reveal the complete
MPL and catalyst layers is being pursued. Infrared imaging technique of the GDL cross-section with 5 lm pixel resolution at 30 Hz
was tested to reveal the temperature gradients across GDLs in an
operating PEMFC.
Acknowledgments
This work was supported by the New York State Energy
Research and Development Authority under contract no. 9575.
Rochester Institute of Technology assisted in meeting the publication
costs of this article.
References
1. W. Dai, H. Wang, X. Z. Yuan, J. J. Martin, D. Yang, J. Qiao, and J. Ma, Int.
J. Hydrogen Energy, 34, 23 (2009).
2. S. G. Kandlikar and Z. Lu, J. Fuel Cell Sci. Technol., 6, 4 (2009).
3. Z. Lu, M. M. Daino, C. Rath, and S. G. Kandlikar, Int. J. Hydrogen Energy, 35, 9
(2010).
4. C. Y. Wang, Chem. Rev., 104, 10 (2004).
5. F. Y. Zhang, X. G. Yang, and C. Y. Wang, J. Electrochem. Soc., 153, A225 (2006).
6. T. A. Trabold, J. P. Owejan, J. J. Gagliardo, D. L. Jacobson, D. S. Hussey, and
M. Arif, in Handbook of Fuel Cells: Advance Electrocatalysis, Matererial,
Diagnostics and Durability, Vols. 5 and 6, W. Vielstich, H. A. Gasteiger, and
H. Yokokawa, Editors, pp. 658–672, John Wiley & Sons, Chichester, UK
(2009).
7. A. Hakenjos, H. Muenter, U. Wittstadt, and C. Hebling, J. Power Sources, 131, 1
(2004).
8. R. Shimoi, M. Masuda, K. Fushinobu, Y. Kozawa, and K. Okazaki, J. Energy
Resour. Technol., 126, 4 (2004).
9. L. S. Martins, J. E. F. C. Gardolinski, J. V. C. Vargas, J. C. Ordonez, S. C. Amico,
and M. M. C. Forte, Appl. Therm. Eng., 29, 14 (2009).
10. M. M. Daino and S. G. Kandlikar, in Proceeding ASME International Conference
Nanochannels, Microchannels, and Minichannels, ASME, p. 467 (2009).
11. D. S. Hussey, D. L. Jacobson, M. Arif, J. P. Owejan, J. J. Gagliardo, and T. A.
Trabold, J. Power Sources, 172, 1 (2007).
12. J. St-Pierre, J. Electrochem. Soc., 154, B724 (2007).
13. Y. Ishikawa, T. Morita, K. Nakata, K. Yoshida, and M. Shiozawa, J. Power
Sources, 163, 2 (2007).
14. Y. Ishikawa, T. Mortia, K. Nakata, K. Yoshida, and M. Shiozawa, J. Power
Sources, 179, 2 (2008).
15. M. M. Daino, A. Lu, J. M. LaManna, J. P. Owejan, T. A. Trabold, and S. K. Kandlikar, ECS Trans., 33(1), 1423 (2010).
16. D. Spernjak, A. K. Prasad, and S. G. Advani, J. Power Sources, 170, 2 (2007).
17. X. Liu, H. Guo, F. Ye, and C. F. Ma, Int. J. Hydrogen Energy, 33, 3 (2008).
18. J. P. Owejan, J. J. Gagliardo, J. M. Sergi, S. G. Kandlikar, and T. A. Trabold,
Int. J. Hydrogen Energy, 34, 8 (2009).
19. S. Ge and C. Y. Wang, J. Electrochem. Soc., 154, B998 (2007).
Downloaded 10 Mar 2011 to 198.208.159.20. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp