- Email: [email protected]
Characteristics of proton exchange membrane fuel cells cold start with silica in cathode catalyst layers
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 3 5 ( 2 0 1 0 ) 5 5 5 2 e5 5 5 7
Available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/he
Characteristics of proton exchange membrane fuel cells cold start with silica in cathode catalyst layers Zhili Miao a,b, Hongmei Yu a,*, Wei Song a,b, Lixing Hao a,b, Zhigang Shao a,*, Qiang Shen a,b, Junbo Hou a,b,1, Baolian Yi a a
Laboratory of Fuel Cells, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, 457 Zhongshan Road, Dalian, Liaoning 116023, PR China b Graduate School of the Chinese Academy of Sciences, Beijing 100039, PR China
article info
abstract
Article history:
In this study, a novel strategy is reported to improve the cold start performance of proton
Received 23 December 2009
exchange membrane (PEM) fuel cells at subzero temperatures. Hydrophilic nano-oxide
Received in revised form
such as SiO2 is added into the catalyst layer (CL) of the cathode to increase its water
9 March 2010
storing capacity. To investigate the effect of nanosized SiO2 addition, the catalyst coated
Accepted 11 March 2010
membranes (CCMs) with 5 wt.% and without nanosized SiO2 are fabricated. Although at
Available online 18 April 2010
normal operation conditions the cell performance with nanosized SiO2 was not so good as
Keywords:
run even at 100 mA cm2 for about 25 min and latter failed very shortly. Even at 10 C, the
that without SiO2, cold start experiments at 8 C showed that the former could start and Proton exchange membrane fuel cell
addition of SiO2 dramatically increased the running time before the cell voltage dropped to
Cold start
zero. These results further experimentally proved the cold start process was strongly
Catalyst layer
related with the cathode water storage capacity. Also, the performance degradation during
Nanosized SiO2
8 cold start cycles was evaluated through polarization curves, cyclic voltammetry (CV) and electrochemical impedance spetra (EIS). Compared with the cell without SiO2 addition, the cell with 5 wt.% SiO2 indicated no obvious degradation on cell performance, electrochemical active surface area and charge transfer resistance after experiencing cold start cycles at 8 C. ª 2010 Professor T. Nejat Veziroglu. Published by Elsevier Ltd. All rights reserved.
1.
Introduction
Cold start capacity and survivability are very important to proton exchange membrane fuel cells (PEMFC) for their vehicular applications. It is well-known that the product water changes to ice or frost when the cell encounters subzero environment. When the membrane, local pores and ionomer in the cathode CL are insufficient to retain all the produced water, the open pores will be filled completely with ice, resulting in lack of reactant gases, reducing electrochemical active surface area (ECA) and even shutting down the cell [1,2].
If the operating temperature of the cell is raised above the freezing point before the cathode CL is filled completely with ice, the cell will be able to start under subzero temperatures successfully without any external aid. Thus, the waterholding capacity of the cathode CL is important to self-cold start of the cell. In the past several years, the studies of cold start have been focused on the related freeze degradation, the gas purge process and cold start behavior [3e10]. The images of the morphology of fuel cell after cold start process, illustrated the magnitude of damages. Water freezing delaminates the
* Corresponding authors. Tel.: þ86 411 84379051; fax: þ86 411 84379185. E-mail address: [email protected] (H. Yu). 1 Present address: Department of Physical Chemistry, University of Leoben, A-8700, Austria. 0360-3199/$ e see front matter ª 2010 Professor T. Nejat Veziroglu. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ijhydene.2010.03.045
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 3 5 ( 2 0 1 0 ) 5 5 5 2 e5 5 5 7
catalyst layer from both the membrane and GDL [4]. The pore volume changed due to the material destruction or other degradations caused by water phase transition in the CL and the membrane. It was reported that the hydrogen desorption peak of the CL, as observed by the CV tests, would decrease after 10 cold start cycles. There was 5.4% degradation of power density per cold start experiment [7]. It was also reported that the cells could realize startup successfully at subzero temperature with external aids [11]. Also, the single cell was able to startup at 5 C if it was pre-purged [12,13]. However, there are few reports referring to the successful self-startup from temperature lower than 5 C. To improve the water storing capacity of the CL and increase the cell performance at low humidity, hydrophilic materials was reported to be added into the CL. Jung et al. [14] added nanosized SiO2 particles into the CL. Performance of the cells with SiO2 decreased at 100% relative humidity (RH), but increased at 40% and 60% RH. The silica sol (prepared from hydrolysis of TEOS solution) and catalyst inks were mixed and then sprayed on the membrane [15]. When the cell was operated at 120 C with 24% RH (inlet gases), the cell with 5 wt. % TEOS (relative to Nafion content in the CL) in the cathode CL showed the best performance. When nanosized SiO2 was added into the cathode CL, the hydrophilicity of the cathode CL would increase according to the contact angle measurement [16]. The cell performance was relatively higher at both full and low humidity with the addition of 5 wt.% nanosized SiO2. In this work, cold start capacity of cells with and without nanosized SiO2 (5 wt.% relative to the whole cathode CL) in the cathode CL were comparatively investigated. The different operating conditions such as the temperatures and the start current densities were performed on the two cells. The effect of cold start on the cell performance was also studied by analyzing the polarization curves, CV and EIS during 8 cold start cycles at the start current density of 50 mA cm2.
2.
Experimental
2.1.
Preparation of MEAs and assembly of single cell
Two MEAs were prepared, which were a standard MEA and a SiO2-doped MEA, respectively. To prepare the standard MEA, the catalyst ink was firstly formed by mixing 40 wt.% Pt/C (Johnson Matthey Inc.), 5 wt.% Nafion solution (DuPont) and isopropyl alcohol. The catalyst ink was then sprayed onto a PTFE substrate and finally transferred onto a Nafion 212 membrane to form a CCM. To prepare the SiO2-doped MEA, additional 5 wt.% nanosized SiO2 (based on the whole cathode CL) was added into the cathode catalyst ink. Both of the anode and the cathode were the same platinum loading of 0.4 mg cm2. The fabricated CCMs were pressed with wetproofed Toray carbon paper to form the MEAs. The effective area of the MEA was 4 cm2.
2.2.
Cold start process
First, the cells which experienced pre-conditioning, performance test and electrochemical measurements, were purged
5553
by RH 20% gas until the RH of outflow gas reached 30% (both the anode and the cathode). Second, reactant gases were fed at a rate of 30 ml min1 for about 1 min to ensure an open circuit voltage above 1.0 V. Then, the cell was started up at 50 mA cm2 with 30 ml min1 flow rate for both H2 and O2 [12], after the cell was kept at the temperature between 8 C and 10 C for 1.5 h in a climate chamber. After the startup procedure, the cell was taken out of the climate chamber and characterized with cell performance and electrochemical measurement tests. Finally, the cell was purged and put into the climate chamber again for the next cold start cycle. Each cell repeated the cold start process eight times.
2.3. Cell performance test and electrochemical measurement After the cold start process, the cell was operated at 50 mA cm2 for 30 min. The operating temperature was 60 C. Then, the characterization of the cell performance was conducted with gas flow rates of 100/30 mL min1 of O2/H2 at atmospheric pressure, respectively, which was recorded by a KFM 2030 impedance meter (Kikusui, Japan). After the performance measurement, in-situ EIS was also measured by this impedance meter. The perturbation amplitude for the sinusoidal signal was 150 mA (peak to peak) over a frequency range of 10 kHze0.5 Hz. A cyclic voltammetry test was applied to measure the ECA of the cathode electrode using a BI-STAT 2 channel Potentiostat (USA, Princeton). The humidified hydrogen and ultra pure nitrogen gas were fed to the anode and the cathode with flow rates of 30 and 50 mL min1, respectively. The CV measurement was carried out between 0 and 1.0 V with a sweep rate of 50 mV s1.
3.
Results and discussion
3.1.
Cell performance
Initial cell performance of the two cells is shown in Fig. 1. The performance of the cell with 5 wt.% nanosized SiO2 in the cathode CL was lower than that of the cell without SiO2 in the whole current density region. It might be attributed to the isolative characteristic of the nanosized SiO2 or due to the high humidity of the reactive gas. This result is very consistent with the previous work [14]. The two cells started up from 8 C at 50 mA cm2 and the variations of voltages during the startup are shown in Fig. 2. The voltage of the cell with 5 wt.% SiO2 was lower than the one without SiO2 during the initial period, which accords well with the cell performance shown in Fig. 1. After 15 min operation, the voltage of the cell without SiO2 fell off. On the contrary, the cell with 5 wt.% nanosized SiO2 was operating continually for more than 2 h. The cold start performance of the PEMFC was measured by the time a cell could operate before it was shutdown by ice filling in the cathode CL [17]. Due to hydrophilicity of the nanosized SiO2, the cathode CL of the cell with 5 wt.% nanosized SiO2 could retain more product water than the one without SiO2. If the time during which the cell could operate is used to judge the cold start performance of PEMFC,
5554
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 3 5 ( 2 0 1 0 ) 5 5 5 2 e5 5 5 7
1.0
0 wt.% 5 wt.%
Voltage / V
0.8
0.6
0.4
0.2
0.0 0.0
0.2
0.4
0.6
0.8
1.0
-2
Current density / A cm
Fig. 1 e Initial performance of the two cells: with 5wt. % nanosized SiO2 in the cathode catalyst layer; with no SiO2 in the cathode catalyst layer.
compared with the cell without SiO2, the cold start capability of the cell with 5 wt.% nano SiO2 was better. For the evaluation of the cold start experiments, the total product water is calculated according to the following equation [6]. mH2 O ¼
MH2 O 2F
Z I dt
(1)
Where mH2 O is the total product water, MH2 O the molecular weight of water and I the cell current. For the cell without SiO2, the total product water was 4.2 mg cm2 during the startup process. To further illustrate the effect of current density on the cold start process, the cold start experiment results are compared in Fig. 3. In Fig. 3a, the cell without SiO2 shut down within 1 min at the current densities of 100 and 200 mA cm2, and only lasted for 15 min at 50 mA cm2. However, the cell with 5 wt.% nanosized SiO2 was able to startup from 8 C at
Fig. 2 e Cell voltages during startup at 50 mA cmL2 from L8 C.
Fig. 3 e Cold start processes of the cells after different cold start cycles: (a) with no SiO2 (b) with 5 wt.% nanosized SiO2.
50 mA cm2 and ran throughout the investigated time (Fig. 3b). Furthermore, the cell with 5 wt.% nanosized SiO2 was able to operate longer than the one without SiO2 at 100 mA cm2. In the case of low current density, the reaction current distributed uniformly throughout the pores in CL, and the water retaining capacity of CL could be fully utilized. However, in the case of high current density, the reaction current was not distributed uniformly. Water was produced more quickly and less time was available for the product water to transfer to the membrane. Consequently, the ice formation was more severe in the cathode CL at higher current densities. Higher amount of product water could form an ice sheet at the electrode interface before water retaining capacity of the cathode CL was fully utilized [9]. Fig. 4 gives the performance for the two cells during 8 cold start cycles. The cold start tests were operated at 8 C and at 50 mA cm2 with 30 ml min1 flow rate for both H2 and O2. Fig. 4a shows that there is a small performance loss of the cell without SiO2. The voltage difference between the 2nd cold start cycle and the 1st cold start cycle was more than 0.04 V at 1 A cm2. However, no performance loss was observed in the whole investigated current density range from the 2nd cold start cycle to the fourth cold start cycle. There was a visible
5555
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 3 5 ( 2 0 1 0 ) 5 5 5 2 e5 5 5 7
10
1.2
Voltage / V
0.8 0.6 0.4
-2
0.2
Current density 0 wt.% 5 wt.%
0.8
6 0.6 4
0.4
2
0.2
0
0.0 00:02:28
0.0
0.2
0.4
0.6
0.8
1.0
-2
Current density / A cm
00:07:26
Fig. 5 e Time evolution of cell voltages during the startup from L10 C at 50 mA cmL2.
1.2
1st 2nd 3rd 4th 5th 6th 7th 8th
1.0
0.8
Voltage / V
00:04:57 Time / min
0.0
b
8
Voltage / V
1st 2nd 3rd 4th 5th 6th 7th
1.0
1.0
Current density / mA cm
a
0.6
0.4
0.2
operation of the cell with 5 wt.% nanosized SiO2 lasted longer. This was mainly attributed to the hydrophilic SiO2 addition into the cathode CL. The CL was able to retain more product water, and thus the cell shutdown was delayed. Since ice formation of the cell with SiO2 is reduced, SiO2 could help the ionomer water absorption. This opens more time for the water in the ionomer of the cathode CL to diffuse into the membrane [19,20]. According to Eq. (1), the total product water was 2.51 and 1.78 mg cm2, for the cell with 5 wt.% and without nanosized SiO2 in the cathode CL, respectively.
3.2.
0.0 0.0
0.2
0.4
0.6
0.8
Electrochemical active surface area
1.0
Current Density / A cm-2
Fig. 4 e Cell performances of the cells after different cold start cycles (a) with no SiO2 (b) with 5 wt.% nanosized SiO2.
voltage loss at high current density for the 5th cold start cycle. The cell could not operate at more than 600 mA cm2 after the 5th cold start cycle. This situation had been observed previously [4], and it was most likely related with the water blocking effect. Since this cell could not startup successfully during the cold start, the product water formed ice to hinder the access of the reactant gas to the reaction site. As a result, both the reversible and irreversible performance lost contributes to such phenomenon [18]. Fig. 4(b) indicates that there is no visible performance loss for the cell with 5 wt.% nanosized SiO2 within the 8 cold start cycles. The cell could startup from 8 C at 50 mA cm2 successfully. Due to the addition of nanosized SiO2, the water capacity of the cathode CL increased. The ice did not fill all the pores of the cathode CL, and then the oxygen reduction reaction could proceed during the cold start process. Therefore, the cell performance did not have an obvious performance loss with the cold start. When the cells were operated under 10 C at 50 mA cm2, two characteristics in the voltage curves can be seen in Fig. 5. First, they were almost the same at the initial performance measurements at low current density. Second, the cold start
The ECAs of the cells with no SiO2 and 5 wt.% nanosized SiO2 in the cathode CL were also investigated. The CV curves are shown in Fig. 6 and the calculated ECA results are shown in Fig. 7. The ECA reflected the active surface area of the cathode CL and could be calculated on the base of the hydrogen desorption peak area (HDA) [21]. A HDA decrease of the cell with no SiO2 is observed in Fig. 6(a). This result is consistent with the previously reported results [22]. The decreased cell performance was also observed. However, there was no significant performance decrease of the cell with 5 wt.% nanosized SiO2 in the cathode CL with the cold start cycles, indicated in Fig. 6(b). Furthermore, the ECA’s of the cell with 5 wt.% nanosized SiO2 did not change, as shown in Fig. 7, while for the one with no SiO2, the ECA was quite different e it decreased from 39.8 g1 m2 to 25.2 g1 m2 after the first cold start. This indicates that the number of catalytically active sites decreased. After the 8th cold start, the ECA of the cell with no SiO2 was 18.0 g1 m2.
3.3.
Electrochemical impedance spectra
Fig. 8 shows the effect of the cold start cycles with the start current density of 100 mA cm2 on the EIS responses for the cells with 5 wt.% and without nanosized SiO2. Generally, the anodic polarization is very small and could be neglected. EIS spectra of H2/O2 PEMFC are practically determined by the cathode [23]. The impedance spectrum includes a highest frequency intercept with the real axis denoting the sum of
5556
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 3 5 ( 2 0 1 0 ) 5 5 5 2 e5 5 5 7
a
1 3 5 7
0.4
Z' / ohm cm
2
0.3
0.2
0.1
0.77ohm cm2 0.0 0.2
0.4
0.6
0.8
1.0
2
Z / ohm cm
b
0.4
1 3 5 7
Z' / ohm cm
2
0.3
0.2
0.1
0.70 ohm cm
2
0.0 0.2
0.4
0.6
Z / ohm cm
Fig. 6 e CV of the cells at different cold start cycles (a) with no SiO2 (b) with 5 wt.% nanosized SiO2.
interfacial contact and material bulk resistance (R1), a higher frequency arc reflecting the combination of a charge transfer resistance (Rct), a double layer capacitance (Cdl) and a lower frequency arc correlated with the mass transport process. The double layer capacitance is considered to be a constant phase element. 50
0 wt.% 5 wt.%
ECA / g-1m2
40
30
20
10
0 0
1
2
3
4
5
6
7
Times of cold start Fig. 7 e Cold start effect on the ECA of the cells.
8
0.8
1.0
2
Fig. 8 e Impedance response of the cells at 100 mA cmL2 with different cycles of cold start (a) with no SiO2 (b) with 5 wt.% nanosized SiO2 in the cathode CLs.
The values of R1 did not increase with the cold start cycles in Fig. 8. This indicates that the sum of membrane resistance and the contact resistance between the membrane, electrodes and bipolar plates remained constant. At 100 mA cm2, the Rct of the cell without SiO2 changed from 0.47 Ucm2 to 0.77 Ucm2 after seven cycles in Fig. 8(a). The increase rate was 43.66 mU cm2 per cycle, which was consistent with the cell performance results. With the cold start cycle, the cell with hydrophilic CCM showed an obvious increase trend in the polarization resistance [7]. The cell without SiO2 was not able to startup from 8 C at 50 mA cm2 because of the ice in the pores of the cathode CL. The pore volume expansion was caused by ice formation. Once the ice had melted, a thin water film would form and block the pores in the cathode CL, and so oxygen diffusion would be blocked by the thin water film, which would induce a decay in the cell performance. Also, the increase of Rct ascribed to the catalyst specific surface area decrease should be considered. In addition, Rct of the cell with 5 wt.% nanosized SiO2 fluctuated less than the one without SiO2, in Fig. 8(b). The increase rate was 8.1 mU cm2 per cycle. The charge transfer process related to the oxygen reduction reaction was influenced by the proton transport and oxygen infiltration to the catalyst sites. Due to the hydrophilicity of nanosized SiO2, the cell was able to startup from 8 C at 50 mA cm2 for 8 cycles with no visible performance decrease. The performance was
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 3 5 ( 2 0 1 0 ) 5 5 5 2 e5 5 5 7
not influenced by the ice formation. Thus, the charge transfer resistance changes of the cell with 5 wt.% nanosized SiO2 were not obvious with the number of the cold start cycles.
4.
Conclusion
The hydrophilic nano SiO2 was added in the cathode catalyst layer and its effect on the cold start behavior was investigated as well as the related degradation by comparative study at different current densities and temperatures. The addition of SiO2 made the cell capable of starting and running even at 100 mA cm2 from 8 C for about 25 min while the cell without the SiO2 failed very shortly after the cold start. Even at 10 C the cold start performance was dramatically improved for the cell with SiO2. Furthermore, although the catalyst layer with SiO2 could hold more water than that without SiO2, there was no evident performance loss, cell resistance increase and ECA decay observed for the cell with nano SiO2 during the 8 cold start cycles.
Acknowledgments This work was financially supported by the National High Technology Research and Development Program of China (863 Program, No. 2007AA05Z123) and the National Natural Science Foundations of China (No. 20636060, No. 20876154).
references
[1] Mao L, Wang CY. Analysis of cold start in polymer electrolyte fuel cells. J Electrochem Soc 2007;154:B139e46. [2] Tajiri K, Tabuchi Y, Kagami F, Takahashi S, Yoshizawa K, Wang CY. Effects of operating and design parameters on PEFC cold start. J Power Sources 2007;165:279e86. [3] Hishinuma Y, Chikahisa T, Kagami F, Ogawa T. The design and performance of a PEFC at temperature below freezing. JSME Int J Ser B 2004;47:235e41. [4] Yan Q, Toghiani H, Lee YW, Liang K, Causey H. Effect of subfreezing temperatures on a PEM fuel cell performance. J Power Sources 2006;160:1242e50. [5] Ahluwalia RK, Wang X. Rapid self-start of polymer electrolyte fuel cell stacks from subfreezing temperatures. J Power Sources 2006;162:502e12. [6] Oszcipok M, Riemann D, Kronenwett U, Kreideweis M, Zedda M. Statistic analysis of operational influences on the cold start behaviour of PEM fuel cells. J Power Sources 2005; 145:407e15.
5557
[7] Song W, Hou JB, Yu HM, Hao LX, Shao ZG, Yi BL. Sub-freezing endurance of PEM fuel cells with different catalyst-coated membranes. J Appl Electrochem 2009;39:609e10. [8] Tajiri K, Tabuchi Y, Wang CY. Isothermal cold start of polymer electrolyte fuel cells. J Electrochem Soc 2007;154: B147e52. [9] Mao L, Wang CY, Tabuchi Y. A multiphase model for cold start of polymer electrolyte fuel cells. J Electrochem Soc 2007; 154:B341e51. [10] Ge SH, Wang CY. In situ imaging of liquid water and ice formation in an operating PEFC during cold start. Eletrochemical Solid Lett 2006;9:A499e503. [11] Sun SC, Yu HM, Hou JB, Shao ZG, Yi BL, Ming PW, et al. Catalytic hydrogen/oxygen reaction assisted the proton exchange membrane fuel cell (PEMFC) start up at subzero temperature. J Power Sources 2008;177:137e41. [12] Hou JB, Yi BL, Yu HM, Hao LX, Song W, Fu Y, et al. Investigation of resided water effects on PEM fuel cell after cold start. Int J Hydrogen Energy 2007;32:4503e9. [13] Oszcipok M, Zedda M, Riemann D, Geckeler D. Low temperature operation and influence parameters on the cold start ability of portable PEMFCs. J Power Sources 2006;154: 404e11. [14] Jung UH, Park KT, Park EH, Kim SH. Improvement of lowhumidity performance of PEMFC by addition of hydrophilic SiO2 particles to catalyst layer addition of hydrophilic SiO2 particles to catalyst layer. J Power Sources 2006;159:529e32. [15] Vengatesan S, Kim HJ, Lee SY, Cho EA, Ha HY, Oh IH, et al. High temperature operation of PEMFC: a novel approach using MEA with silica in catalyst layer. Int J Hydrogen Energy 2008;33:171e8. [16] Miao ZL, Yu HM, Song W, Zhao D, Hao LX, Yi BL, et al. Effect of hydrophilic SiO2 additive in cathode catalyst layers on proton exchange membrane fuel cells. Electrochemical Commun 2009;11:787e90. [17] Ge SH, Wang CY. Characteristics of subzero startup and water/ice formation on the catalyst layer in a polymer electrolyte fuel cell. Eletrochim Acta 2007;52:4825e35. [18] Oszcipok M, Zedda M, Hesselmann J, Huppmann M, Wodrich M, Junghardt M, et al. Portable proton exchange membrane fuel-cell systems for outdoor applications. J Power Sources 2006;157:666e73. [19] Jiao K, Li XG. Effects of various operating and initial conditions on cold start performance of polymer electrolyte membrane fuel cells. Int J Hydrogen Energy 2009;34: 8171e84. [20] Jiao K, Li XG. Cold start analysis of polymer electrolyte membrane fuel cells. Int J Hydrogen Energy 2010;35:5077e94. [21] Ralph TR, Hards GA, Keating JE, Campbell SA, Wilkinson DP, Davis M, et al. Low cost electrodes for proton exchange membrane fuel cells. J Electrochem Soc 1997;144:3845e57. [22] Ge SH, Wang CY. Cyclic voltammetry study of ice formation in the PEFC catalyst layer during cold start. J Electrochem Soc 2007;154:B1399e406. [23] Ciureanu M, Roberge R. Eletrochemical impedance study of PEM fuel cells. Experimental diagnostics and modeling of air cathodes. J Phys Chem B 2001;105:3531e9.
Available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/he
Characteristics of proton exchange membrane fuel cells cold start with silica in cathode catalyst layers Zhili Miao a,b, Hongmei Yu a,*, Wei Song a,b, Lixing Hao a,b, Zhigang Shao a,*, Qiang Shen a,b, Junbo Hou a,b,1, Baolian Yi a a
Laboratory of Fuel Cells, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, 457 Zhongshan Road, Dalian, Liaoning 116023, PR China b Graduate School of the Chinese Academy of Sciences, Beijing 100039, PR China
article info
abstract
Article history:
In this study, a novel strategy is reported to improve the cold start performance of proton
Received 23 December 2009
exchange membrane (PEM) fuel cells at subzero temperatures. Hydrophilic nano-oxide
Received in revised form
such as SiO2 is added into the catalyst layer (CL) of the cathode to increase its water
9 March 2010
storing capacity. To investigate the effect of nanosized SiO2 addition, the catalyst coated
Accepted 11 March 2010
membranes (CCMs) with 5 wt.% and without nanosized SiO2 are fabricated. Although at
Available online 18 April 2010
normal operation conditions the cell performance with nanosized SiO2 was not so good as
Keywords:
run even at 100 mA cm2 for about 25 min and latter failed very shortly. Even at 10 C, the
that without SiO2, cold start experiments at 8 C showed that the former could start and Proton exchange membrane fuel cell
addition of SiO2 dramatically increased the running time before the cell voltage dropped to
Cold start
zero. These results further experimentally proved the cold start process was strongly
Catalyst layer
related with the cathode water storage capacity. Also, the performance degradation during
Nanosized SiO2
8 cold start cycles was evaluated through polarization curves, cyclic voltammetry (CV) and electrochemical impedance spetra (EIS). Compared with the cell without SiO2 addition, the cell with 5 wt.% SiO2 indicated no obvious degradation on cell performance, electrochemical active surface area and charge transfer resistance after experiencing cold start cycles at 8 C. ª 2010 Professor T. Nejat Veziroglu. Published by Elsevier Ltd. All rights reserved.
1.
Introduction
Cold start capacity and survivability are very important to proton exchange membrane fuel cells (PEMFC) for their vehicular applications. It is well-known that the product water changes to ice or frost when the cell encounters subzero environment. When the membrane, local pores and ionomer in the cathode CL are insufficient to retain all the produced water, the open pores will be filled completely with ice, resulting in lack of reactant gases, reducing electrochemical active surface area (ECA) and even shutting down the cell [1,2].
If the operating temperature of the cell is raised above the freezing point before the cathode CL is filled completely with ice, the cell will be able to start under subzero temperatures successfully without any external aid. Thus, the waterholding capacity of the cathode CL is important to self-cold start of the cell. In the past several years, the studies of cold start have been focused on the related freeze degradation, the gas purge process and cold start behavior [3e10]. The images of the morphology of fuel cell after cold start process, illustrated the magnitude of damages. Water freezing delaminates the
* Corresponding authors. Tel.: þ86 411 84379051; fax: þ86 411 84379185. E-mail address: [email protected] (H. Yu). 1 Present address: Department of Physical Chemistry, University of Leoben, A-8700, Austria. 0360-3199/$ e see front matter ª 2010 Professor T. Nejat Veziroglu. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ijhydene.2010.03.045
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 3 5 ( 2 0 1 0 ) 5 5 5 2 e5 5 5 7
catalyst layer from both the membrane and GDL [4]. The pore volume changed due to the material destruction or other degradations caused by water phase transition in the CL and the membrane. It was reported that the hydrogen desorption peak of the CL, as observed by the CV tests, would decrease after 10 cold start cycles. There was 5.4% degradation of power density per cold start experiment [7]. It was also reported that the cells could realize startup successfully at subzero temperature with external aids [11]. Also, the single cell was able to startup at 5 C if it was pre-purged [12,13]. However, there are few reports referring to the successful self-startup from temperature lower than 5 C. To improve the water storing capacity of the CL and increase the cell performance at low humidity, hydrophilic materials was reported to be added into the CL. Jung et al. [14] added nanosized SiO2 particles into the CL. Performance of the cells with SiO2 decreased at 100% relative humidity (RH), but increased at 40% and 60% RH. The silica sol (prepared from hydrolysis of TEOS solution) and catalyst inks were mixed and then sprayed on the membrane [15]. When the cell was operated at 120 C with 24% RH (inlet gases), the cell with 5 wt. % TEOS (relative to Nafion content in the CL) in the cathode CL showed the best performance. When nanosized SiO2 was added into the cathode CL, the hydrophilicity of the cathode CL would increase according to the contact angle measurement [16]. The cell performance was relatively higher at both full and low humidity with the addition of 5 wt.% nanosized SiO2. In this work, cold start capacity of cells with and without nanosized SiO2 (5 wt.% relative to the whole cathode CL) in the cathode CL were comparatively investigated. The different operating conditions such as the temperatures and the start current densities were performed on the two cells. The effect of cold start on the cell performance was also studied by analyzing the polarization curves, CV and EIS during 8 cold start cycles at the start current density of 50 mA cm2.
2.
Experimental
2.1.
Preparation of MEAs and assembly of single cell
Two MEAs were prepared, which were a standard MEA and a SiO2-doped MEA, respectively. To prepare the standard MEA, the catalyst ink was firstly formed by mixing 40 wt.% Pt/C (Johnson Matthey Inc.), 5 wt.% Nafion solution (DuPont) and isopropyl alcohol. The catalyst ink was then sprayed onto a PTFE substrate and finally transferred onto a Nafion 212 membrane to form a CCM. To prepare the SiO2-doped MEA, additional 5 wt.% nanosized SiO2 (based on the whole cathode CL) was added into the cathode catalyst ink. Both of the anode and the cathode were the same platinum loading of 0.4 mg cm2. The fabricated CCMs were pressed with wetproofed Toray carbon paper to form the MEAs. The effective area of the MEA was 4 cm2.
2.2.
Cold start process
First, the cells which experienced pre-conditioning, performance test and electrochemical measurements, were purged
5553
by RH 20% gas until the RH of outflow gas reached 30% (both the anode and the cathode). Second, reactant gases were fed at a rate of 30 ml min1 for about 1 min to ensure an open circuit voltage above 1.0 V. Then, the cell was started up at 50 mA cm2 with 30 ml min1 flow rate for both H2 and O2 [12], after the cell was kept at the temperature between 8 C and 10 C for 1.5 h in a climate chamber. After the startup procedure, the cell was taken out of the climate chamber and characterized with cell performance and electrochemical measurement tests. Finally, the cell was purged and put into the climate chamber again for the next cold start cycle. Each cell repeated the cold start process eight times.
2.3. Cell performance test and electrochemical measurement After the cold start process, the cell was operated at 50 mA cm2 for 30 min. The operating temperature was 60 C. Then, the characterization of the cell performance was conducted with gas flow rates of 100/30 mL min1 of O2/H2 at atmospheric pressure, respectively, which was recorded by a KFM 2030 impedance meter (Kikusui, Japan). After the performance measurement, in-situ EIS was also measured by this impedance meter. The perturbation amplitude for the sinusoidal signal was 150 mA (peak to peak) over a frequency range of 10 kHze0.5 Hz. A cyclic voltammetry test was applied to measure the ECA of the cathode electrode using a BI-STAT 2 channel Potentiostat (USA, Princeton). The humidified hydrogen and ultra pure nitrogen gas were fed to the anode and the cathode with flow rates of 30 and 50 mL min1, respectively. The CV measurement was carried out between 0 and 1.0 V with a sweep rate of 50 mV s1.
3.
Results and discussion
3.1.
Cell performance
Initial cell performance of the two cells is shown in Fig. 1. The performance of the cell with 5 wt.% nanosized SiO2 in the cathode CL was lower than that of the cell without SiO2 in the whole current density region. It might be attributed to the isolative characteristic of the nanosized SiO2 or due to the high humidity of the reactive gas. This result is very consistent with the previous work [14]. The two cells started up from 8 C at 50 mA cm2 and the variations of voltages during the startup are shown in Fig. 2. The voltage of the cell with 5 wt.% SiO2 was lower than the one without SiO2 during the initial period, which accords well with the cell performance shown in Fig. 1. After 15 min operation, the voltage of the cell without SiO2 fell off. On the contrary, the cell with 5 wt.% nanosized SiO2 was operating continually for more than 2 h. The cold start performance of the PEMFC was measured by the time a cell could operate before it was shutdown by ice filling in the cathode CL [17]. Due to hydrophilicity of the nanosized SiO2, the cathode CL of the cell with 5 wt.% nanosized SiO2 could retain more product water than the one without SiO2. If the time during which the cell could operate is used to judge the cold start performance of PEMFC,
5554
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 3 5 ( 2 0 1 0 ) 5 5 5 2 e5 5 5 7
1.0
0 wt.% 5 wt.%
Voltage / V
0.8
0.6
0.4
0.2
0.0 0.0
0.2
0.4
0.6
0.8
1.0
-2
Current density / A cm
Fig. 1 e Initial performance of the two cells: with 5wt. % nanosized SiO2 in the cathode catalyst layer; with no SiO2 in the cathode catalyst layer.
compared with the cell without SiO2, the cold start capability of the cell with 5 wt.% nano SiO2 was better. For the evaluation of the cold start experiments, the total product water is calculated according to the following equation [6]. mH2 O ¼
MH2 O 2F
Z I dt
(1)
Where mH2 O is the total product water, MH2 O the molecular weight of water and I the cell current. For the cell without SiO2, the total product water was 4.2 mg cm2 during the startup process. To further illustrate the effect of current density on the cold start process, the cold start experiment results are compared in Fig. 3. In Fig. 3a, the cell without SiO2 shut down within 1 min at the current densities of 100 and 200 mA cm2, and only lasted for 15 min at 50 mA cm2. However, the cell with 5 wt.% nanosized SiO2 was able to startup from 8 C at
Fig. 2 e Cell voltages during startup at 50 mA cmL2 from L8 C.
Fig. 3 e Cold start processes of the cells after different cold start cycles: (a) with no SiO2 (b) with 5 wt.% nanosized SiO2.
50 mA cm2 and ran throughout the investigated time (Fig. 3b). Furthermore, the cell with 5 wt.% nanosized SiO2 was able to operate longer than the one without SiO2 at 100 mA cm2. In the case of low current density, the reaction current distributed uniformly throughout the pores in CL, and the water retaining capacity of CL could be fully utilized. However, in the case of high current density, the reaction current was not distributed uniformly. Water was produced more quickly and less time was available for the product water to transfer to the membrane. Consequently, the ice formation was more severe in the cathode CL at higher current densities. Higher amount of product water could form an ice sheet at the electrode interface before water retaining capacity of the cathode CL was fully utilized [9]. Fig. 4 gives the performance for the two cells during 8 cold start cycles. The cold start tests were operated at 8 C and at 50 mA cm2 with 30 ml min1 flow rate for both H2 and O2. Fig. 4a shows that there is a small performance loss of the cell without SiO2. The voltage difference between the 2nd cold start cycle and the 1st cold start cycle was more than 0.04 V at 1 A cm2. However, no performance loss was observed in the whole investigated current density range from the 2nd cold start cycle to the fourth cold start cycle. There was a visible
5555
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 3 5 ( 2 0 1 0 ) 5 5 5 2 e5 5 5 7
10
1.2
Voltage / V
0.8 0.6 0.4
-2
0.2
Current density 0 wt.% 5 wt.%
0.8
6 0.6 4
0.4
2
0.2
0
0.0 00:02:28
0.0
0.2
0.4
0.6
0.8
1.0
-2
Current density / A cm
00:07:26
Fig. 5 e Time evolution of cell voltages during the startup from L10 C at 50 mA cmL2.
1.2
1st 2nd 3rd 4th 5th 6th 7th 8th
1.0
0.8
Voltage / V
00:04:57 Time / min
0.0
b
8
Voltage / V
1st 2nd 3rd 4th 5th 6th 7th
1.0
1.0
Current density / mA cm
a
0.6
0.4
0.2
operation of the cell with 5 wt.% nanosized SiO2 lasted longer. This was mainly attributed to the hydrophilic SiO2 addition into the cathode CL. The CL was able to retain more product water, and thus the cell shutdown was delayed. Since ice formation of the cell with SiO2 is reduced, SiO2 could help the ionomer water absorption. This opens more time for the water in the ionomer of the cathode CL to diffuse into the membrane [19,20]. According to Eq. (1), the total product water was 2.51 and 1.78 mg cm2, for the cell with 5 wt.% and without nanosized SiO2 in the cathode CL, respectively.
3.2.
0.0 0.0
0.2
0.4
0.6
0.8
Electrochemical active surface area
1.0
Current Density / A cm-2
Fig. 4 e Cell performances of the cells after different cold start cycles (a) with no SiO2 (b) with 5 wt.% nanosized SiO2.
voltage loss at high current density for the 5th cold start cycle. The cell could not operate at more than 600 mA cm2 after the 5th cold start cycle. This situation had been observed previously [4], and it was most likely related with the water blocking effect. Since this cell could not startup successfully during the cold start, the product water formed ice to hinder the access of the reactant gas to the reaction site. As a result, both the reversible and irreversible performance lost contributes to such phenomenon [18]. Fig. 4(b) indicates that there is no visible performance loss for the cell with 5 wt.% nanosized SiO2 within the 8 cold start cycles. The cell could startup from 8 C at 50 mA cm2 successfully. Due to the addition of nanosized SiO2, the water capacity of the cathode CL increased. The ice did not fill all the pores of the cathode CL, and then the oxygen reduction reaction could proceed during the cold start process. Therefore, the cell performance did not have an obvious performance loss with the cold start. When the cells were operated under 10 C at 50 mA cm2, two characteristics in the voltage curves can be seen in Fig. 5. First, they were almost the same at the initial performance measurements at low current density. Second, the cold start
The ECAs of the cells with no SiO2 and 5 wt.% nanosized SiO2 in the cathode CL were also investigated. The CV curves are shown in Fig. 6 and the calculated ECA results are shown in Fig. 7. The ECA reflected the active surface area of the cathode CL and could be calculated on the base of the hydrogen desorption peak area (HDA) [21]. A HDA decrease of the cell with no SiO2 is observed in Fig. 6(a). This result is consistent with the previously reported results [22]. The decreased cell performance was also observed. However, there was no significant performance decrease of the cell with 5 wt.% nanosized SiO2 in the cathode CL with the cold start cycles, indicated in Fig. 6(b). Furthermore, the ECA’s of the cell with 5 wt.% nanosized SiO2 did not change, as shown in Fig. 7, while for the one with no SiO2, the ECA was quite different e it decreased from 39.8 g1 m2 to 25.2 g1 m2 after the first cold start. This indicates that the number of catalytically active sites decreased. After the 8th cold start, the ECA of the cell with no SiO2 was 18.0 g1 m2.
3.3.
Electrochemical impedance spectra
Fig. 8 shows the effect of the cold start cycles with the start current density of 100 mA cm2 on the EIS responses for the cells with 5 wt.% and without nanosized SiO2. Generally, the anodic polarization is very small and could be neglected. EIS spectra of H2/O2 PEMFC are practically determined by the cathode [23]. The impedance spectrum includes a highest frequency intercept with the real axis denoting the sum of
5556
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 3 5 ( 2 0 1 0 ) 5 5 5 2 e5 5 5 7
a
1 3 5 7
0.4
Z' / ohm cm
2
0.3
0.2
0.1
0.77ohm cm2 0.0 0.2
0.4
0.6
0.8
1.0
2
Z / ohm cm
b
0.4
1 3 5 7
Z' / ohm cm
2
0.3
0.2
0.1
0.70 ohm cm
2
0.0 0.2
0.4
0.6
Z / ohm cm
Fig. 6 e CV of the cells at different cold start cycles (a) with no SiO2 (b) with 5 wt.% nanosized SiO2.
interfacial contact and material bulk resistance (R1), a higher frequency arc reflecting the combination of a charge transfer resistance (Rct), a double layer capacitance (Cdl) and a lower frequency arc correlated with the mass transport process. The double layer capacitance is considered to be a constant phase element. 50
0 wt.% 5 wt.%
ECA / g-1m2
40
30
20
10
0 0
1
2
3
4
5
6
7
Times of cold start Fig. 7 e Cold start effect on the ECA of the cells.
8
0.8
1.0
2
Fig. 8 e Impedance response of the cells at 100 mA cmL2 with different cycles of cold start (a) with no SiO2 (b) with 5 wt.% nanosized SiO2 in the cathode CLs.
The values of R1 did not increase with the cold start cycles in Fig. 8. This indicates that the sum of membrane resistance and the contact resistance between the membrane, electrodes and bipolar plates remained constant. At 100 mA cm2, the Rct of the cell without SiO2 changed from 0.47 Ucm2 to 0.77 Ucm2 after seven cycles in Fig. 8(a). The increase rate was 43.66 mU cm2 per cycle, which was consistent with the cell performance results. With the cold start cycle, the cell with hydrophilic CCM showed an obvious increase trend in the polarization resistance [7]. The cell without SiO2 was not able to startup from 8 C at 50 mA cm2 because of the ice in the pores of the cathode CL. The pore volume expansion was caused by ice formation. Once the ice had melted, a thin water film would form and block the pores in the cathode CL, and so oxygen diffusion would be blocked by the thin water film, which would induce a decay in the cell performance. Also, the increase of Rct ascribed to the catalyst specific surface area decrease should be considered. In addition, Rct of the cell with 5 wt.% nanosized SiO2 fluctuated less than the one without SiO2, in Fig. 8(b). The increase rate was 8.1 mU cm2 per cycle. The charge transfer process related to the oxygen reduction reaction was influenced by the proton transport and oxygen infiltration to the catalyst sites. Due to the hydrophilicity of nanosized SiO2, the cell was able to startup from 8 C at 50 mA cm2 for 8 cycles with no visible performance decrease. The performance was
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 3 5 ( 2 0 1 0 ) 5 5 5 2 e5 5 5 7
not influenced by the ice formation. Thus, the charge transfer resistance changes of the cell with 5 wt.% nanosized SiO2 were not obvious with the number of the cold start cycles.
4.
Conclusion
The hydrophilic nano SiO2 was added in the cathode catalyst layer and its effect on the cold start behavior was investigated as well as the related degradation by comparative study at different current densities and temperatures. The addition of SiO2 made the cell capable of starting and running even at 100 mA cm2 from 8 C for about 25 min while the cell without the SiO2 failed very shortly after the cold start. Even at 10 C the cold start performance was dramatically improved for the cell with SiO2. Furthermore, although the catalyst layer with SiO2 could hold more water than that without SiO2, there was no evident performance loss, cell resistance increase and ECA decay observed for the cell with nano SiO2 during the 8 cold start cycles.
Acknowledgments This work was financially supported by the National High Technology Research and Development Program of China (863 Program, No. 2007AA05Z123) and the National Natural Science Foundations of China (No. 20636060, No. 20876154).
references
[1] Mao L, Wang CY. Analysis of cold start in polymer electrolyte fuel cells. J Electrochem Soc 2007;154:B139e46. [2] Tajiri K, Tabuchi Y, Kagami F, Takahashi S, Yoshizawa K, Wang CY. Effects of operating and design parameters on PEFC cold start. J Power Sources 2007;165:279e86. [3] Hishinuma Y, Chikahisa T, Kagami F, Ogawa T. The design and performance of a PEFC at temperature below freezing. JSME Int J Ser B 2004;47:235e41. [4] Yan Q, Toghiani H, Lee YW, Liang K, Causey H. Effect of subfreezing temperatures on a PEM fuel cell performance. J Power Sources 2006;160:1242e50. [5] Ahluwalia RK, Wang X. Rapid self-start of polymer electrolyte fuel cell stacks from subfreezing temperatures. J Power Sources 2006;162:502e12. [6] Oszcipok M, Riemann D, Kronenwett U, Kreideweis M, Zedda M. Statistic analysis of operational influences on the cold start behaviour of PEM fuel cells. J Power Sources 2005; 145:407e15.
5557
[7] Song W, Hou JB, Yu HM, Hao LX, Shao ZG, Yi BL. Sub-freezing endurance of PEM fuel cells with different catalyst-coated membranes. J Appl Electrochem 2009;39:609e10. [8] Tajiri K, Tabuchi Y, Wang CY. Isothermal cold start of polymer electrolyte fuel cells. J Electrochem Soc 2007;154: B147e52. [9] Mao L, Wang CY, Tabuchi Y. A multiphase model for cold start of polymer electrolyte fuel cells. J Electrochem Soc 2007; 154:B341e51. [10] Ge SH, Wang CY. In situ imaging of liquid water and ice formation in an operating PEFC during cold start. Eletrochemical Solid Lett 2006;9:A499e503. [11] Sun SC, Yu HM, Hou JB, Shao ZG, Yi BL, Ming PW, et al. Catalytic hydrogen/oxygen reaction assisted the proton exchange membrane fuel cell (PEMFC) start up at subzero temperature. J Power Sources 2008;177:137e41. [12] Hou JB, Yi BL, Yu HM, Hao LX, Song W, Fu Y, et al. Investigation of resided water effects on PEM fuel cell after cold start. Int J Hydrogen Energy 2007;32:4503e9. [13] Oszcipok M, Zedda M, Riemann D, Geckeler D. Low temperature operation and influence parameters on the cold start ability of portable PEMFCs. J Power Sources 2006;154: 404e11. [14] Jung UH, Park KT, Park EH, Kim SH. Improvement of lowhumidity performance of PEMFC by addition of hydrophilic SiO2 particles to catalyst layer addition of hydrophilic SiO2 particles to catalyst layer. J Power Sources 2006;159:529e32. [15] Vengatesan S, Kim HJ, Lee SY, Cho EA, Ha HY, Oh IH, et al. High temperature operation of PEMFC: a novel approach using MEA with silica in catalyst layer. Int J Hydrogen Energy 2008;33:171e8. [16] Miao ZL, Yu HM, Song W, Zhao D, Hao LX, Yi BL, et al. Effect of hydrophilic SiO2 additive in cathode catalyst layers on proton exchange membrane fuel cells. Electrochemical Commun 2009;11:787e90. [17] Ge SH, Wang CY. Characteristics of subzero startup and water/ice formation on the catalyst layer in a polymer electrolyte fuel cell. Eletrochim Acta 2007;52:4825e35. [18] Oszcipok M, Zedda M, Hesselmann J, Huppmann M, Wodrich M, Junghardt M, et al. Portable proton exchange membrane fuel-cell systems for outdoor applications. J Power Sources 2006;157:666e73. [19] Jiao K, Li XG. Effects of various operating and initial conditions on cold start performance of polymer electrolyte membrane fuel cells. Int J Hydrogen Energy 2009;34: 8171e84. [20] Jiao K, Li XG. Cold start analysis of polymer electrolyte membrane fuel cells. Int J Hydrogen Energy 2010;35:5077e94. [21] Ralph TR, Hards GA, Keating JE, Campbell SA, Wilkinson DP, Davis M, et al. Low cost electrodes for proton exchange membrane fuel cells. J Electrochem Soc 1997;144:3845e57. [22] Ge SH, Wang CY. Cyclic voltammetry study of ice formation in the PEFC catalyst layer during cold start. J Electrochem Soc 2007;154:B1399e406. [23] Ciureanu M, Roberge R. Eletrochemical impedance study of PEM fuel cells. Experimental diagnostics and modeling of air cathodes. J Phys Chem B 2001;105:3531e9.