Journal of Power Sources 195 (2010) 536–540
Contents lists available atScienceDirect
Journal of Power Sources
j o u r n a l h o m e p a g e :w w w . e l s e v i e r . c o m / l o c a t e / j p o w s o u r
Short communication
Water permeation analysis on gas diffusion layers of proton exchange
membrane fuel cells for Teflon-coating annotation
Yen-I Chou
a,∗, Zih-Yuan Siao
a,b, Yu-Feng Chen
a, Lung-Yu Sung
a, Wen-Mei Yang
b, Chi-Chuan Wang
aaEnergy and Environment Research Laboratories, Industrial Technology Research Institute, Chutung 310, Taiwan bDepartment of Mechanical Engineering, National Chiao Tung University, Hsinchu 300, Taiwan
a r t i c l e i n f o
Article history: Received 23 July 2009 Accepted 29 July 2009 Available online 7 August 2009 Keywords:
Proton exchange membrane fuel cells Gas diffusion layer
Hydrophobic Water permeation Water flooding Polytetrafluoroethylene
a b s t r a c t
This study presents an analysis of water permeation of a polytetrafluoroethylene (PTFE)-coated gas diffu-sion layer (GDL) to determine the influence of hydrophobic treatment on the GDL for diagnosis of water flooding. It is found that the behaviour of water drainage is controlled by the pore configuration instead of the hydrophobicity in GDL. Better water drainage is achieved by the action of the Teflon coating in modulating the GDL pore configuration to give both a larger average pore size and a wider distribution of pore size. The results show that water penetration through the GDL must overcome a threshold surface tension defined by the largest pore range. A 30 wt.% PTFE coating of a GDL is shown to generate a satis-factory pore configuration, explaining the improved cell polarization performance with a lower driven pressure (∼1.91 kPa) and a higher rate of water drainage.
© 2009 Elsevier B.V. All rights reserved.
1. Introduction
The proton exchange membrane fuel cell (PEMFC) is one of the most promising energy resources for the future due to its advan-tages of high power density at low temperature, very low or zero emission of greenhouse gases, long-term reliable operation and facility of system set-up[1–3]. However, it is well known that PEMFC suffers from water flooding, commonly from the reaction at the cathode, O2+ 4H++ 4e−→ 2H2O, especially during high power
density operation[4,5]. The excess water accumulates in the mem-brane electrode assembly (MEA), which blocks the oxygen feed and deteriorates the cell polarization performance considerably. For efficient water management, the gas diffusion layer (GDL) is commonly hydrophobically treated with polytetrafluoroethylene (PTFE) for flooding mitigation. A Teflon coating on a GDL is intended to engender vapour diffusion and water shear force for efficient drainage [3,6–8]. Velayutham et al. [9] examined the effects of PTFE contents in the range 7–30 wt.% and found a 23 wt.% Teflon-coated GDL to be optimal in terms of flooding diagnosis. Tüber et al. [10]demonstrated that a 20 wt.% PTFE coating produces a strongly hydrophobic GDL, which leads to superior cell performance at room temperature. The review by Li et al.[3]showed an optimal con-centration of PTFE within 20–40 wt.%, depending on the conditions used for the preparation of GDL samples. It is believed that the PTFE
∗ Corresponding author. Tel.: +886 3 5919277; fax: +886 3 5829782. E-mail address:[email protected](Y.-I. Chou).
coating on a GDL is a compromise between pore hydrophobicity and pore configuration[11,12]. Normally, too little PTFE coating results in insufficient hydrophobicity for water removal, while excessive PTFE loading causes an inferior cell performance, since it blocks the GDL pore void and reduces electrical conductivity.
Direct observation of flooding into GDL is quite complicated and may interfere with both the GDL specimen and the operat-ing conditions. Most studies have examined only the outcome of GDL hydrophobic treatment by cell polarization testing [12–15] or computational modelling[16,17]. However, these characteriza-tions were usually accompanied by many other operative effects, which inevitably obscure the actual flooding phenomenon inside PEMFC systems. This could explain the lack of the direct evidence needed to specify the degree of PTFE coating for flooding diagnosis. Recently, Benziger et al.[18]described a pressurized membrane filtration system for simulating water penetration through a GDL in order to evaluate the role of the PTFE coating. Firstly, they found that the water must achieve a threshold pressure to force water through the largest pores in the GDL. Secondly, they showed the largest pores in a GDL were effective for drainage of water flood-ing. And the water drainage was dominated by only a small portion of void in GDL. This result implies the water drainage was correlated largely with the pore configuration of the GDL, which differs from the common viewpoint of hydrophobic promotion. The objective of this study was to extend the experiments with water-pressurized GDL filtration to a broader range of operative pressures, combined with an examination of the influence of pore configuration on water drainage for various PTFE-coated GDLs. Moreover, the effect of PTFE 0378-7753/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
Fig. 1. A diagram of the pressurized water filtration system. The pressure is
mea-sured as the height of the hydrostatic head of the water above the GDL.
treatment on GDLs was examined with respect to modulation of adequate pore configuration instead of the surface hydrophobicity in GDL.
2. Experimental approach
2.1. Hydrophobic Teflon coating and characterization
The GDLs were all obtained from SIGRACET® GDL (SGL 10AA,
SGL Carbon Group, Germany) as carbon papers without PTFE or microporous layer (MPL) coating. The carbon papers were cut into 2.4 cm× 2.4 cm squares and immersed in a PTFE aqueous dispersion solution (DuPontTMTeflon®PTFE TE-3893) for 30 min. PTFE
emul-sions of 15–60 wt.% were prepared. The coated carbon papers were air-dried at 120◦C for 1 h, and subsequently sintered in a furnace at 350◦C under a nitrogen atmosphere with a temperature rising rate of 3◦C min−1for 30 min. The treated GDLs were characterized by optical microscopy (OM) (BX51, Olympus) and scanning electron microscopy (SEM) (JSM-7000F, JEOL) for surface morphological and structural observations, respectively. The hydrophobicity of the GDLs was characterized by the contact angle with a static water drop shape analysis system (DSA100, KRÜSS). The contact angle was determined as the average value of at least three measure-ments of 10L water droplets at randomly chosen regions on each GDL sample. The porous structure of the GDL was characterized with a mercury porosimeter (Autopore 9520, Micromeritics). The average pore size and size distribution were determined from the mercury intrusion by an applied pressure based on capillary law. 2.2. Water permeation characterization
The water permeation through GDL was characterized in a pres-surized membrane filtration cell. As shown inFig. 1, the GDL was placed on a laboratory-built polymethylmethacrylate (PMMA) cell attached to a 70 cm high, 1.4 cm diameter cylinder. Initially, a peri-staltic pump (MP-1000, EYELA) was used to introduce water into the cylinder slowly (0.2 g min−1) until the water started to pene-trate the GDL. When the water started to drain out of the GDL, the hydrostatic head, i.e. the height of the water in the cylinder, was recorded as the threshold pressure Pth. The pump was regulated
to maintain Pthin order to measure the drained water flow. The
Fig. 2. A model of the pressurized water penetration through the GDL with
trans-verse cylindrical pairs of different diameters under (a) P < Pth, (b) P = Pth, and (c)
P > Pth.
amount of water drainage was recorded with a precision balance (GF-300, A&D) with a time interval of 5 s. The water permeation behaviour at pressures >Pth, i.e. higher hydrostatic heads, was
char-acterized. Unlike the study reported by Benziger et al.[18], the water permeation characteristics and its influence subject to pore structure variation were examined above the threshold pressure.
3. Water permeation analysis
Fig. 2shows a diagram for water penetration through the GDL with various pore diameters under different driven pressures. In this model, it is assumed that the pores are cylindrical and dis-tributed non-uniformly. The penetration pressure P of water into a pore of size d in GDL can be described by the Young and Laplace relation[19]:
P =4 cos
d (1)
where is the surface tension of water, is the contact angle of water with the pore surface, and d is the pore diameter. As shown, the water is first expelled from the pores when the applied pressure is insufficient to override the surface tension (as shown in the case of P < PthinFig. 2(a)). When the pressure is increased to Pth,
the water can overcome the minimum surface tension and start to penetrate through the GDL via the largest pores (dmax) as shown in
Fig. 2(b). With this micro-sized pore configuration, a laminar flow should prevail through these pores, hence the water flow rate (Qth)
drained from dmaxcan be described by Darcy’s law[20]:
Qth= KthPth= kthAth
538 Y.-I. Chou et al. / Journal of Power Sources 195 (2010) 536–540
Fig. 3. SEM and OM images of (a) the uncoated and (b) the 100 wt.% Teflon-coated GDLs.
where kthis the intrinsic permeability coefficient of water at pore
dmax, Athis the total area of dmaxpores, is the viscosity of water,
and L is the penetration length of water at a GDL thickness of 0.38 mm. As shown inFig. 2(c), a further increase of the pressure to P1(when P > Pth) is applied for water penetration to
over-come the higher resistance from pores of small diameter (d1). In
this respect, the flow rate (Q1) at P1is related to the water fluxes
from pores dmaxand d1:
Q1= K1P1= KthP1+ K1P1 (3)
where K1is the apparent permeability of water at P1through dmax
and d1, and K1 is the permeability of water through only d1.
With a further rise of applied pressure to Pn, the pore sizes
larger than dnare all activated for water penetration. Thus, Knfor
all penetrable pore sizes (from dnto dmax) and the Knfor pore dn
are: Kn= Qn Pn (4) Kn = Kn− Kth− n−1
i=1 Ki=knAn L (5)where An is the total penetration area for pore size dn. Here,
the flow rate qn for the indicated pores dn can be obtained
from Pnd4n/128L and subsequently Ancan be evaluated from
nn(dn/2)2, where the pore number (nn) is determined by (Qn−
Kn−1Pn)/qn. Once Anis known, the permeability coefficient of kn
for pore size dncan be obtained from Eq.(5). According to this
analy-sis, the behaviour of water penetration and the pore configuration of GDL are correlated. It can be seen from the results presented above that drainage in the fuel cell is dominated by macropores; therefore, the water-pressurized GDL filtration in this work was characterized at hydrostatic heads of <70 cm, which focuses on pores with a diameter >30m. This study was designed to charac-terize water penetration behaviour with pressure-activated pore structure in order to explain the benefit of the Teflon coating for GDL.
4. Results and discussion
Fig. 3 shows typical SEM and OM images of GDLs with and without a PTFE coating. It is clear that the PTFE is coated on the cross space of the carbon fibres. The PTFE coating reduces the void volume and the pore number of the GDL.Fig. 4shows the pore distribution of GDLs with different PTFE loadings. The majority of pore sizes of the GDLs were in the range 20–400m (macropores) with a few in the range 0.2–10m (micropores). For the uncoated GDL, the porosity is 91.0% and the average pore diameter dav
is 61.4m. With the PTFE coating, the voids in both micro- and macro-regions are inevitably reduced. This is especially
conspicu-Fig. 4. Average pore size and porosity of GDLs with different PTFE coatings. Inset:
Fig. 5. Effect of hydrostatic pressure on the water flow rate for various Teflon-coated
GDLs. Inset: time-dependent water weight drained from uncoated GDL at various hydrostatic pressures.
ous for small pores, where the Teflon fills a considerable portion of voids, smoothing out the micropore region and leading to a larger average pore size. It is shown that the porosity is reduced to 78.3% but the average pore size is increased to 76.7m for a PTFE con-centration of 60 wt.%.
Fig. 5illustrates the flow rates of different PTFE-coated GDLs under various hydrostatic head heights and the inset shows the weight of water drained from the GDL. The linearity of water drained versus time indicates these flow rates are constant and sta-ble for the relevant pore structure regime.Fig. 5shows that, due to the hydrophobicity of GDL, no water drainage can be achieved when the height of the hydrostatic head is lower than that cor-responding to the penetration pressure Pth. These Pth values
are 2.72 kPa (27.7 cm), 2.58 kPa (26.3 cm), 1.91 kPa (19.5 cm), and 2.02 kPa (20.6 cm) for 0 wt.%, 15 wt.%, 30 wt.%, and 60 wt.% PTFE-coated GDLs, respectively. According to Eq. (1), the maximum pore diameter for each Pthvalue corresponds to 67.7m, 84.4 m,
117.6m, and 113.7 m for 0 wt.%, 15 wt.%, 30 wt.%, and 60 wt.% PTFE-coated GDLs, respectively. This result coincides with the pore analysis from mercury-intrusion porosimetry, which showed that the maximum pore size increased with increased concentration of PTFE. With more PTFE coating, the voids in the micropores are fur-ther filled, fur-thereby enhancing the wt.% of macropores and resulting in a decrease of Pthfor the initial draining. These characteristics
of GDLs with different amounts of PTFE coating are summarized in Table 1.
Table 1
Characteristics of GDLs coated with various PTFE emulsion concentrations. PTFE(%) Pmin(Pa) (◦) davg(m) dmaxa(m) (%)
0 2716.6 129.7 61.4 67.7 91.0
15 2577.4 139.1 64.8 84.4 90.9
30 1911.0 141.3 66.0 117.6 87.8
60 2018.8 142.9 76.7 113.7 78.3
ad
maxis calculated based on of 71.975 mN m−1at 298 K from Eq.(1) [21].
Fig. 6. Plots of Kversus d for various Teflon-coated GDLs.
Also shown inFig. 5, the 30 wt.% PTFE-coated GDL exhibits the superior water penetration with its lower Pthand higher flow rates.
This result is compatible with those of other studies, where the concentration of PTFE emulsion was 20–40 wt.%[22]for attainable water drainage and polarization performance. Besides, it is noted that the flow rate is not in linear proportion to the driving pres-sure, which implicates a discrepancy in Darcy’s law. Part of the explanation might be that the GDL has an uneven distribution of pore size and a tortuous pore structure, showing a departure from Darcy’s assumption of single cylindrical pore distribution. The K values of GDLs at various pore diameters inFig. 6indicate that 30 wt.% GDL has the desired Kcharacteristics applicable to a large pore range of 60–110m. The uncoated GDL also shows acceptable water drainage, but it is limited to pores <65m. It can be seen from Eq.(5)that the Kfor each range of pore size is controlled by two parameters; i.e. intrinsic permeability, k, and permeance area, A. The k property is independent of flowing fluid, but is often corre-lated with the configuration of porous media via the well known formula as k = cd2 where c is a dimensionless constant for pore
structural properties. A can be identified from the quantity of the indicated pore size and further related to the pore configuration of the GDL.
Fig. 7 shows the k values evaluated from various Teflon-coated GDLs with different pore diameters. Most k values are <2.0× 10−10m2 and are proportional to d2with a slope of 0.031
(c∼ 1/32), which is in good agreement with Poiseuille’s law: Q = 1
32 d2A
LP (6)
The inset inFig. 7shows the pore size distributions of different Teflon-coated GDLs can be obtained from this water penetration analysis. For an uncoated GDL, the small pore size range controls water drainage. With Teflon coating, the effective pore size of GDL shifts to a larger pore size range and the 30 wt.% Teflon-coated GDL provides the largest average pore size of 40–120m and the widest pore size distribution for superior water drainage with lower threshold pressures as well as efficient drainage flux. It is well known that a Teflon coating raises the GDL hydrophobicity and minimizes vapour condensation. However, the variation of pore configuration along with the Teflon coating can lead to significant improvement of water drainage. This manipulation of GDL pore
dis-540 Y.-I. Chou et al. / Journal of Power Sources 195 (2010) 536–540
Fig. 7. Linear plot of k versus d2for various Teflon-coated GDLs. Inset: macropore
size distribution of the Teflon-coated GDLs.
Fig. 8. The polarization curves of PEM fuel cells with uncoated and 30 wt.%
Teflon-coated GDLs.
tribution gives rise to an effective increase of macropores, assisting the water drainage at a smaller Pthand a higher flux performance.
For further verification of the results, polarization tests were done for evaluation of the Teflon coating. Voltage-controlled measure-ments were performed with a Gore PRIMEA®5621using anode and
cathode catalyst loadings of 0.45 (Pt-Ru) and 0.6 (Pt) mg cm−2at a cell temperature of 65◦C and at an ambient pressure of 1 atm. From two serpentine flow field channels, anode and cathode gases were fed with 1.5× and 3.0× H2/air stoichiometries at 70◦C and 80◦C
humidification, respectively.Fig. 8shows the polarization differ-ence between Teflon-coated and uncoated GDLs. From a current density of 400 mA cm−2, the cell voltage of the uncoated GDL was
lower than that of a 30 wt.% Teflon-coated GDL by approximately 5% and the difference was increased with a further increase of current density. When the current density exceeded 1000 mA cm−2, there was a further substantial drop of cell voltage for the untreated GDL due to considerable water flooding. In summary, with an appropri-ate Teflon coating to modulappropri-ate the pore configuration on the GDL, it is possible to achieve adequate water management for superior cell polarization performance.
5. Conclusions
The results of a water permeation analysis of various PTFE-coated GDLs are presented in this study. It was shown that the pore configuration of a GDL has a key role in the diagnosis of water flood-ing. In line with the proposed water penetration mechanism, water drainage began from the largest pores in a GDL, and the intrin-sic water permeability of a GDL was controlled by pore size only. The effect of a PTFE coating on the pore configuration is impor-tant, since it provides a dramatic improvement of water drainage from a GDL. The contribution of the pore configuration to drainage outweighs the influence of hydrophobicity. In this work, the exper-imental results showed that a 30 wt.% PTFE-coated GDL had larger macropores, which resulted in superior cell polarization perfor-mance featuring a favourable water drainage characteristic and a greater permeability efficiency at a much lower driven pressure.
Acknowledgements
Part of this work is financially supported by the Department of Industrial Technology, Ministry of Economic Affairs, Taiwan. Authors wish to express their thanks to Dr. Huan-Ruei Shiu for his help in GDL support and technique discussion.
References
[1] B.C.H. Steele, A. Heinzel, Nature 414 (2001) 345–352.
[2] F. Barbir, PEM Fuel Cells, Elsevier Academic Press, Burlington, 2005. [3] H. Li, Y. Tang, Z. Wang, Z. Shi, S. Wu, D. Song, J. Zhang, K. Fatih, J. Zhang, H. Wang,
Z. Liu, R. Abouatallah, A. Mazza, J. Power Sources 178 (2008) 103–117. [4] D. Natarajan, T.V. Nguyen, J. Power Sources 115 (2003) 66–80.
[5] C. Song, Y. Tang, J.L. Zhang, J. Zhang, H. Wang, J. Shen, S. McDermid, J. Li, P. Kozak, Electrochim. Acta 52 (2007) 2552–2561.
[6] C.Y. Wang, Chem. Rev. 104 (2004) 4727–4766.
[7] J.T. Gostick, M.W. Fowler, M.A. Ioannidis, M.D. Pritzker, Y.M. Volfkovich, A. Sakars, J. Power Sources 156 (2006) 375–387.
[8] S. Litster, C.R. Buie, T. Fabian, J.K. Eaton, J.G. Santiago, J. Electrochem. Soc. 154 (2007) B1049–B1058.
[9] G. Velayutham, J. Kaushik, N. Rajalakshmi, K.S. Dhathathreyan, Fuel Cells 4 (2007) 314–318.
[10] K. Tüber, D. Pocza, C. Hebling, J. Power Sources 124 (2003) 403–414. [11] J. Moreira, A.L. Ocampo, P.J. Sebastian, M.A. Smit, M.D. Salazar, P.D. Angel, J.A.
Montoya, R. Perez, L. Martinez, Int. J. Hydrogen Energy 28 (2003) 625–627. [12] G.G. Park, Y.J. Sohn, T.H. Yang, Y.G. Yoon, W.Y. Lee, C.S. Kim, J. Power Sources
131 (2004) 182–187.
[13] D. Bevers, R. Rogers, M.V. Bradke, J. Power Sources 63 (1996) 193–201. [14] E. Passalacqua, G. Squadrito, F. Lufrano, A. Patti, L. Giorgi, J. Appl. Electrochem.
31 (2001) 449–454.
[15] M. Prasanna, H.Y. Ha, E.A. Cho, S.-A. Hong, I.-H. Oh, J. Power Sources 131 (2004) 147–154.
[16] H.S. Chu, C. Yeh, F. Chen, J. Power Sources 123 (2003) 1–9. [17] K. Jiao, B. Zhou, J. Power Sources 169 (2007) 296–314.
[18] J. Benziger, J. Nehlsen, D. Blackwell, T. Brennan, J. Itescu, J. Membr. Sci. 261 (2005) 98–106.
[19] H.-J. Butt, K. Graf, M. Kappl, Physics and Chemistry of Interfaces, Wiley-VCH, Weinheim, 2006.
[20] W.G. Gray, C.T. Miller, Environ. Sci. Technol. 38 (2004) 5895–5901.
[21] D.R. Lide, CRC Handbook of Chemistry and Physics, 89th ed., CRC Press, Boca Raton, 2008.
[22] Y.W. Chen-Yang, T.F. Hung, J. Huang, F.L. Yang, J. Power Sources 173 (2007) 183–188.
