- Email: [email protected]
In-situ synchrotron X-ray radiography on high temperature polymer electrolyte fuel cells
Electrochemistry Communications 12 (2010) 1436–1438
Contents lists available at ScienceDirect
Electrochemistry Communications j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / e l e c o m
In-situ synchrotron X-ray radiography on high temperature polymer electrolyte fuel cells Wiebke Maier a,⁎, Tobias Arlt c, Christoph Wannek a, Ingo Manke c, Heinrich Riesemeier d, Philipp Krüger b, Joachim Scholta b, Werner Lehnert a, John Banhart c, Detlef Stolten a a
Forschungszentrum Jülich GmbH, Institute of Energy Research, IEF-3: Fuel Cells, 52425 Jülich, Germany Zentrum für Sonnenenergie- und Wasserstoff-Forschung Baden-Württemberg, Helmholtzstr. 8, 89081 Ulm, Germany Helmholtz-Centre Berlin for Materials and Energy, Hahn-Meitner-Platz 1, 14109 Berlin, Germany d Bundesanstalt für Materialforschung und -prüfung, Richard-Willstätter-Str. 11, 12489 Berlin, Germany b c
a r t i c l e
i n f o
Article history: Received 28 June 2010 Received in revised form 18 July 2010 Accepted 3 August 2010 Available online 10 August 2010 Keywords: HT-PEFC Synchrotron X-ray radiography Membrane electrode assembly (MEA) Phosphoric acid
a b s t r a c t In contrast to classical low temperature polymer electrolyte fuel cells (LT-PEFCs), the membrane conductivity in high temperature polymer electrolyte fuel cells (HT-PEFCs) (operating temperature ~ 160 °C) is based on proton transport within phosphorus-oxygen acids at different levels of hydration, orthophosphoric acid (H3PO4) being the simplest example. We present for the first time in-situ synchrotron X-ray radiography measurements applied to a HT-PEFC to gain insight into the local composition of the membrane electrode assembly (MEA) under dynamic operating conditions. Transmission changes during the radiographic measurements exhibit a clear influence of the formation of product water on the membrane composition. © 2010 Elsevier B.V. All rights reserved.
1. Introduction Phosphoric acid doped polybenzimidazole as electrolyte for high temperature polymer electrolyte fuel cells (HT-PEFCs) with an operating temperature of about 160 °C exhibits good proton conductivity, low gas permeability and good mechanical stability at elevated temperatures [1]. To maximize the performance and lifetime of HTPEFCs, it is necessary to design membrane electrode assemblies (MEAs) in which the distribution of the phosphoric acid between the membrane, the catalyst layers and the gas diffusion layers can be carefully adjusted [2]. The performance of a HT-PEFC with phosphoric acid doped polybenzimidazole as electrolyte is nearly independent of the way of acid introduction into the MEA but strongly depends on the amount inserted into it [3,4]. In order to achieve an optimum performance, high power densities and long lifetimes, it becomes necessary to provide excellent proton conductivity through the membrane. The composition of the membrane at 160 °C consists of a polybenzimidazole type membrane, in the present case poly(2,5benzimidazole) (ABPBI), and phosphoric acid (H3PO4) which forms an equilibrium with its dehydration products (mainly pyrophosphoric acid H4P2O7) Eqs. (1) [5,6]. In the following we only will use the term phosphoric acid when talking about this equilibrium. The redistribution of the phosphoric acid and the hydration of the pyrophosphoric
⁎ Corresponding author. Tel.: + 49 2461619073; fax: + 49 2461616695. E-mail address: [email protected] (W. Maier). 1388-2481/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.elecom.2010.08.002
acid in the MEA are fast processes, accelerated by product water from the fuel cell operation. When produced in larger amounts, the product water leads to a dilution of the phosphoric acid Eqs. (2): H4 P2 O7 þ H2 O↔2H3 PO4 −
ð1Þ þ
H3 PO4 þ H2 O↔H2 PO4 þ H3 O
ð2Þ
These processes are believed to influence the local density of the different phosphoric acid type species in the MEA, the overall conductivity and viscosity of the electrolyte as well as the membrane composition and thus local physico-chemical properties of the HTPEFC, such as proton conductivity and oxygen gas solubility. Up to now only averaging and ex-situ methods have been used to analyze the acid distribution [4,7–9]. Therefore, highly temporally and spatially resolved in-situ measurements with synchrotron X-ray radiation were performed in order to gain insight into the local distribution of phosphoric acid in the MEA without the need for intermitting the dynamic operating conditions. Until now synchrotron X-ray radiography was only applied to low temperature fuel cells for the investigation of the water content in LT-PEFCs [10,11] and the CO2 evolution in direct methanol fuel cells [12]. As an alternative to the synchrotron X-ray radiography measurements, neutron radiography was already applied to LT-PEFCs in order to visualize liquid water profiles of the cell at different operating conditions [13,14]. The advantage of the neutron radiography is the strong attenuation of the neutrons due to water compared to other
W. Maier et al. / Electrochemistry Communications 12 (2010) 1436–1438
1437
cell materials like graphite flow fields and aluminum endplates. Neutron radiography was not applied to HT-PEFCs, yet. 2. Experimental 2.1. MEA preparation and single cell performance A commercial gas diffusion layer (GDL) composed of a carbon cloth with a microporous layer on one side (BASF Fuel Cell) was coated with a dispersion of carbon-supported catalyst (20% HP Pt on Vucan XC-72, BASF Fuel Cell), PTFE (Dyneon, 40 wt.% in the final catalyst layer) and dispersants by means of a doctor blade technique. The gas diffusion electrodes (GDE) were produced with a platinum loading of 1.1 mg cm-2. After drying, the GDEs were doped with 16.0 mg cm−2 phosphoric acid and assembled with an undoped 30 μm thick poly(2,5-benzimidazole) (ABPBI) membrane (FuMA-Tech) in a single test cell. The cell was operated at 160 °C and ambient pressure. The mass flows of hydrogen and air have been adjusted to λ = 2/2 (anode/cathode). 2.2. Synchrotron X-ray radiography The radiographic measurements were performed at the synchrotron tomography station of the Helmholtz-Centre Berlin (HZB, BAMline, BESSY Germany). A monochromatic 30 keV X-ray beam was used to ensure a sufficiently high transmission through the cell. An optical setup with a 4008 × 2672 pixel CCD camera (PCO 4000 with a CdWO scintillator screen) was used to capture images with area sizes of up to 8.7 mm × 5.8 mm with pixel sizes between 0.438 μm and 2.190 μm and an optical spatial resolution of 1-5 μm. Typical exposure times per image were between 4 and 6 s, including a readout time of about 1.5 s. A cell set-up with in-plane viewing direction was chosen for all experiments which allows to distinguish clearly between the GDLs, the electrodes and the membrane. 3. Results and discussion In Fig. 1a–e so-called normalized images of the cross-section of a HT-PEFC at different operating conditions are displayed. These figures have been generated by dividing the respective radiographs (recorded at OCV just before applying current, exactly 15 min after adjusting the current densities to 140, 300 and 550 mA cm−2 respectively and after 1 h at OCV after the load changes) by the same reference radiograph recorded at OCV at steady state conditions after the initial break-in of the cell. Due to the fact that the cell moves slightly in the beam at this high operation temperature, the borders between the MEA components as well as the flow field edges of the cell can be seen even in Fig. 1a. Furthermore, in Fig. 1e some features are observed due to the fact that the redistribution of phosphoric acid within the MEA to steady-state conditions needs more than 1 h when switching from current density to OCV (cf. [9]). The membrane can be seen in the centre of the image. In the adjacent layers to the left and right side of the membrane the anode and cathode, respectively, on the GDLs are visible. The dry, undoped ABPBI membrane had a thickness of 30 ± 2 μm (not shown here). After the assembly of the membrane with the doped electrodes and after the break-in procedure the thickness of the membrane increased due to the phosphoric acid uptake from the electrodes (see also [4]). Accordingly, at OCV a thickness of 55 ± 3 μm was measured from the original radiograph (not normalized; inset in Fig. 1) by averaging the thickness over 1000 lines of pixel in vertical direction. The membrane exhibits a slight non-planarity in beam direction due to a somewhat undulated membrane/electrode interface under the land and the channels of the flow field. When
Fig. 1. Normalized radiographs of the cross section of the MEA at different current densities j: a) 0 mA cm−2 (OCV before), b) 140 mA cm−2, c) 300 mA cm−2, d) 550 mA cm−2 and e) 0 mA cm−2 (OCV after). Inset: Non-normalized enlarged radiographs of the membrane and parts of the catalyst layers at OCV and j=140 mA cm−2 (A, anode; M, membrane; C, cathode).
switching from OCV to a current density of j = 140 mA cm−2 (Fig. 1b; typical cell voltage: 600 mV), a swelling of the membrane of about 10 μm to a final thickness of 65 ± 3 μm was observed. A further increase in current density to 300 mA cm−2 (Fig. 1c) and 550 mA cm−2 (Fig. 1d; U = 330 mV) did not lead to a significant additional swelling of the membrane; if at all, the resulting effects are too small to be visualized with this optical spatial resolution. This swelling process is reversible as can be seen from Fig. 1e. After 1 h at OCV following the load-cycle changes, the membrane exhibited the same thickness as before the load-cycle changes. There are three possible explanations for the increase in membrane thickness: (i) the dilution/hydration of phosphoric acid in the membrane by product water and (ii) a transport of phosphoric acid from pores of the electrodes into the membrane or (iii) a combination of (i) and (ii). In order to determine which of the aforementioned effects has the strongest impact on the increase in membrane thickness during loadcycle changes, the changes of the transmission signals (i.e. grey values) in the MEA compared to the steady-state OCV condition were analyzed quantitatively. Therefore, different line scans with a vertical width of about 1000 pixel have been evaluated for OCV, 140 mA cm−2, 300 mA cm−2 and 550 mA cm−2 (Fig. 2). Additionally, these scans were filtered by a Gaussian blur. The grey values of these normalized radiographs directly correspond to a change in transmission: a grey value N 1 represents an increase in transmission, while a grey value b 1 refers to a decrease in transmission. A grey value equal to one declares that there are no differences between the reference image at OCV and the actual operating condition. Fig. 2 shows that switching the cell from OCV to a current density of j = 140 mA cm−2 results in an increased transmission in the membrane (M). This result can be explained by the formation of product water, which leads to a hydration of the pyrophosphoric/ phosphoric acid and thus to a swelling of the membrane (conf. Eqs. (1) and (2)). The attenuation coefficient of water (0.1572 cm−1 at 30 keV) is one order of magnitude smaller than that of phosphoric
1438
W. Maier et al. / Electrochemistry Communications 12 (2010) 1436–1438
lead to a grey value smaller than one. The electrodes at the catalyst layer/GDL interface have probably not been completely saturated before, because a significant amount of the phosphoric acid diffused into the pores of the electrodes close to the membrane and into the membrane itself during the break-in procedure (see also [4]), thus a huge part of the pores at the GDL side remained nearly empty. Filling of these pores with phosphoric acid due the fact that the acid phase expands in volume as a consequence of its hydration by product water leads to an increase in absorption in the beam direction and thus to the observed decrease in transmission. 4. Conclusion
Fig. 2. Changes of the transmission through the MEA (grey values) compared to steadystate conditions at OCV for different operating conditions: at OCV (solid line), 140 mA cm−2 (dashed line), 300 mA cm−2 (dotted line) and 550 mA cm−2 (dot and dash line). The thicknesses of the membrane and the catalyst layers when current is drawn from the cell (taken from the original radiographs, cf. inset in Fig. 1) are indicated by the grey shaded areas.
acid (1.0198 cm−1 at 30 keV) at the same X-ray energy [15]. Thus, an increase of water in the membrane as well as a build-up of hydration products of phosphoric acid leads to a higher transmission due to a change in membrane/electrolyte composition. Recently, a similar mechanism has been proposed to describe changes in the impedance of a HT-PEFC (operated at identical conditions as used here) during load-cycle changes [9]. But likewise the membrane thickness does not increase with a further rising of the current density as described before, the transmission within the membrane does not further increase with increasing current density. Hence, it is clearly visible that the biggest effect in membrane swelling as well as in transmission change can be seen when switching the cell from a currentless operating condition to one where current is drawn and thus the formation of product water begins. The further increase of the current does not lead to any further visible effects. This may be due to the fast hydration process at the beginning of the water production because of the hygroscopic nature of phosphoric acid. Fig. 2 shows also an effect within the catalyst layers and the GDLs. Due to a flux of product water from the cathode to the anode about 20% of the product water of a HT-PEFC exits the cell on the anode side under these operating conditions. Accordingly, hydration processes are visible both on the cathode and on the anode side. In the vicinity of the membrane an increase in X-ray transmission can be seen, whereas further away from the membrane, in the vicinity of the GDLs, the transmission within the anode A and the cathode C decreases when switching the operation mode from OCV to a current density of j = 140 mA cm−2. A possible reason could be the hydration of phosphoric acid (and its dehydration products) within the pores of the electrodes in the vicinity of the membrane which leads to an increase in transmission, similar to the effects within the membrane described before. At the same time at the electrode/GDL interface an increased degree of filling of the pores with phosphoric acid would
In-situ synchrotron X-ray radiography was applied for the first time to study (a dynamically) operated HT-PEFC. Analyses of the Xray transmission through the cell revealed significant changes of the intensity distribution during load-cycle changes. The formation of product water after switching from currentless operation to j = 140 mA cm−2 led to a (reversible) ~20% swelling of the membrane which is explained by a hydration/dilution of the phosphoric acidbased electrolyte. The same process is found to take place in that part of the electrodes being close to the membrane, furthermore a filling of pores is assumed to happen in outer parts of the catalyst layers. These effects are fully reversible, whereas a further increase of the current density (from j = 140 to 550 mA cm−2) did not cause any significant detectable changes. Generally, synchrotron X-ray radiography can be used as novel and efficient analytical tool to gain information about the local composition of HT-PEFCs in operando at all relevant operating conditions. It can thus help to optimize the MEA design to increase the performance and to understate degradation mechanisms. To further verify these conclusions, a combination of the radiographic measurements with in-situ performed time-dependent impedance spectroscopy is currently under way. In the future, the impact of GDL materials on the acid distribution within the MEA as well as different acid management strategies will be studied. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15]
Y.L. Ma, J.S. Wainright, M.H. Litt, R.F. Savinell, J. Electrochem. Soc. 151 (2004) A8. J. Mader, L. Xiao, T.J. Schmidt, B.C. Benicewicz, Adv. Polym. Sci. 216 (2008) 63. C. Wannek, B. Kohnen, H.F. Oetjen, H. Lippert, J. Mergel, Fuel Cells 8 (2008) 87. C. Wannek, I. Konradi, J. Mergel, W. Lehnert, Int. J. Hydrogen Energy 34 (2009) 9479. R.F. Jameson, J. Chem. Soc. (1959) 752. D.-T. Chin, H. Chang, J. Appl. Electrochem. 19 (1989) 95. Y. Oono, A. Sounai, M. Hori, J. Power Sources 189 (2009) 943. K. Kwon, J.O. Park, D.Y. Yoo, J.S. Yi, Electrochim. Acta 54 (2009) 6570. K. Wippermann, C. Wannek, H.-F. Oetjen, J. Mergel, W. Lehnert, J. Power Sources 195 (2010) 2806. C. Hartnig, I. Manke, R. Kuhn, N. Kardjilov, J. Banhart, W. Lehnert, Appl. Phys. Lett. 92 (2008) 134106-1. I. Manke, C. Hartnig, M. Grünerbel, W. Lehnert, N. Kardjilov, A. Haibel, A. Hilger, J. Banhart, H. Riesemeier, Appl. Phys. Lett. 90 (2007) 174105-1. C. Hartnig, I. Manke, J. Schloesser, P. Krüger, R. Kuhn, H. Riesemeier, K. Wippermann, J. Banhart, Electrochem. Commun. 11 (2009) 1559. R. Satija, D. Jacobson, M. Arif, S. Werner, J. Power Sources 129 (2004) 238. P. Boillat, D. Kramer, B. Seyfang, G. Frei, E. Lehmann, G. Scherer, A. Wokaun, Y. Ichikawa, Y. Tasaki, K. Shinohara, Electrochem. Commun. 10 (2008) 546. Data calculated with the “X-ray transmission of a solid” program, available on the homepage of the Lawrence Berkeley National Laboratory, Centre of X-ray Optics, http://henke.lbl.gov/optical_constants/filter2.html, 18.06.2010.
Contents lists available at ScienceDirect
Electrochemistry Communications j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / e l e c o m
In-situ synchrotron X-ray radiography on high temperature polymer electrolyte fuel cells Wiebke Maier a,⁎, Tobias Arlt c, Christoph Wannek a, Ingo Manke c, Heinrich Riesemeier d, Philipp Krüger b, Joachim Scholta b, Werner Lehnert a, John Banhart c, Detlef Stolten a a
Forschungszentrum Jülich GmbH, Institute of Energy Research, IEF-3: Fuel Cells, 52425 Jülich, Germany Zentrum für Sonnenenergie- und Wasserstoff-Forschung Baden-Württemberg, Helmholtzstr. 8, 89081 Ulm, Germany Helmholtz-Centre Berlin for Materials and Energy, Hahn-Meitner-Platz 1, 14109 Berlin, Germany d Bundesanstalt für Materialforschung und -prüfung, Richard-Willstätter-Str. 11, 12489 Berlin, Germany b c
a r t i c l e
i n f o
Article history: Received 28 June 2010 Received in revised form 18 July 2010 Accepted 3 August 2010 Available online 10 August 2010 Keywords: HT-PEFC Synchrotron X-ray radiography Membrane electrode assembly (MEA) Phosphoric acid
a b s t r a c t In contrast to classical low temperature polymer electrolyte fuel cells (LT-PEFCs), the membrane conductivity in high temperature polymer electrolyte fuel cells (HT-PEFCs) (operating temperature ~ 160 °C) is based on proton transport within phosphorus-oxygen acids at different levels of hydration, orthophosphoric acid (H3PO4) being the simplest example. We present for the first time in-situ synchrotron X-ray radiography measurements applied to a HT-PEFC to gain insight into the local composition of the membrane electrode assembly (MEA) under dynamic operating conditions. Transmission changes during the radiographic measurements exhibit a clear influence of the formation of product water on the membrane composition. © 2010 Elsevier B.V. All rights reserved.
1. Introduction Phosphoric acid doped polybenzimidazole as electrolyte for high temperature polymer electrolyte fuel cells (HT-PEFCs) with an operating temperature of about 160 °C exhibits good proton conductivity, low gas permeability and good mechanical stability at elevated temperatures [1]. To maximize the performance and lifetime of HTPEFCs, it is necessary to design membrane electrode assemblies (MEAs) in which the distribution of the phosphoric acid between the membrane, the catalyst layers and the gas diffusion layers can be carefully adjusted [2]. The performance of a HT-PEFC with phosphoric acid doped polybenzimidazole as electrolyte is nearly independent of the way of acid introduction into the MEA but strongly depends on the amount inserted into it [3,4]. In order to achieve an optimum performance, high power densities and long lifetimes, it becomes necessary to provide excellent proton conductivity through the membrane. The composition of the membrane at 160 °C consists of a polybenzimidazole type membrane, in the present case poly(2,5benzimidazole) (ABPBI), and phosphoric acid (H3PO4) which forms an equilibrium with its dehydration products (mainly pyrophosphoric acid H4P2O7) Eqs. (1) [5,6]. In the following we only will use the term phosphoric acid when talking about this equilibrium. The redistribution of the phosphoric acid and the hydration of the pyrophosphoric
⁎ Corresponding author. Tel.: + 49 2461619073; fax: + 49 2461616695. E-mail address: [email protected] (W. Maier). 1388-2481/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.elecom.2010.08.002
acid in the MEA are fast processes, accelerated by product water from the fuel cell operation. When produced in larger amounts, the product water leads to a dilution of the phosphoric acid Eqs. (2): H4 P2 O7 þ H2 O↔2H3 PO4 −
ð1Þ þ
H3 PO4 þ H2 O↔H2 PO4 þ H3 O
ð2Þ
These processes are believed to influence the local density of the different phosphoric acid type species in the MEA, the overall conductivity and viscosity of the electrolyte as well as the membrane composition and thus local physico-chemical properties of the HTPEFC, such as proton conductivity and oxygen gas solubility. Up to now only averaging and ex-situ methods have been used to analyze the acid distribution [4,7–9]. Therefore, highly temporally and spatially resolved in-situ measurements with synchrotron X-ray radiation were performed in order to gain insight into the local distribution of phosphoric acid in the MEA without the need for intermitting the dynamic operating conditions. Until now synchrotron X-ray radiography was only applied to low temperature fuel cells for the investigation of the water content in LT-PEFCs [10,11] and the CO2 evolution in direct methanol fuel cells [12]. As an alternative to the synchrotron X-ray radiography measurements, neutron radiography was already applied to LT-PEFCs in order to visualize liquid water profiles of the cell at different operating conditions [13,14]. The advantage of the neutron radiography is the strong attenuation of the neutrons due to water compared to other
W. Maier et al. / Electrochemistry Communications 12 (2010) 1436–1438
1437
cell materials like graphite flow fields and aluminum endplates. Neutron radiography was not applied to HT-PEFCs, yet. 2. Experimental 2.1. MEA preparation and single cell performance A commercial gas diffusion layer (GDL) composed of a carbon cloth with a microporous layer on one side (BASF Fuel Cell) was coated with a dispersion of carbon-supported catalyst (20% HP Pt on Vucan XC-72, BASF Fuel Cell), PTFE (Dyneon, 40 wt.% in the final catalyst layer) and dispersants by means of a doctor blade technique. The gas diffusion electrodes (GDE) were produced with a platinum loading of 1.1 mg cm-2. After drying, the GDEs were doped with 16.0 mg cm−2 phosphoric acid and assembled with an undoped 30 μm thick poly(2,5-benzimidazole) (ABPBI) membrane (FuMA-Tech) in a single test cell. The cell was operated at 160 °C and ambient pressure. The mass flows of hydrogen and air have been adjusted to λ = 2/2 (anode/cathode). 2.2. Synchrotron X-ray radiography The radiographic measurements were performed at the synchrotron tomography station of the Helmholtz-Centre Berlin (HZB, BAMline, BESSY Germany). A monochromatic 30 keV X-ray beam was used to ensure a sufficiently high transmission through the cell. An optical setup with a 4008 × 2672 pixel CCD camera (PCO 4000 with a CdWO scintillator screen) was used to capture images with area sizes of up to 8.7 mm × 5.8 mm with pixel sizes between 0.438 μm and 2.190 μm and an optical spatial resolution of 1-5 μm. Typical exposure times per image were between 4 and 6 s, including a readout time of about 1.5 s. A cell set-up with in-plane viewing direction was chosen for all experiments which allows to distinguish clearly between the GDLs, the electrodes and the membrane. 3. Results and discussion In Fig. 1a–e so-called normalized images of the cross-section of a HT-PEFC at different operating conditions are displayed. These figures have been generated by dividing the respective radiographs (recorded at OCV just before applying current, exactly 15 min after adjusting the current densities to 140, 300 and 550 mA cm−2 respectively and after 1 h at OCV after the load changes) by the same reference radiograph recorded at OCV at steady state conditions after the initial break-in of the cell. Due to the fact that the cell moves slightly in the beam at this high operation temperature, the borders between the MEA components as well as the flow field edges of the cell can be seen even in Fig. 1a. Furthermore, in Fig. 1e some features are observed due to the fact that the redistribution of phosphoric acid within the MEA to steady-state conditions needs more than 1 h when switching from current density to OCV (cf. [9]). The membrane can be seen in the centre of the image. In the adjacent layers to the left and right side of the membrane the anode and cathode, respectively, on the GDLs are visible. The dry, undoped ABPBI membrane had a thickness of 30 ± 2 μm (not shown here). After the assembly of the membrane with the doped electrodes and after the break-in procedure the thickness of the membrane increased due to the phosphoric acid uptake from the electrodes (see also [4]). Accordingly, at OCV a thickness of 55 ± 3 μm was measured from the original radiograph (not normalized; inset in Fig. 1) by averaging the thickness over 1000 lines of pixel in vertical direction. The membrane exhibits a slight non-planarity in beam direction due to a somewhat undulated membrane/electrode interface under the land and the channels of the flow field. When
Fig. 1. Normalized radiographs of the cross section of the MEA at different current densities j: a) 0 mA cm−2 (OCV before), b) 140 mA cm−2, c) 300 mA cm−2, d) 550 mA cm−2 and e) 0 mA cm−2 (OCV after). Inset: Non-normalized enlarged radiographs of the membrane and parts of the catalyst layers at OCV and j=140 mA cm−2 (A, anode; M, membrane; C, cathode).
switching from OCV to a current density of j = 140 mA cm−2 (Fig. 1b; typical cell voltage: 600 mV), a swelling of the membrane of about 10 μm to a final thickness of 65 ± 3 μm was observed. A further increase in current density to 300 mA cm−2 (Fig. 1c) and 550 mA cm−2 (Fig. 1d; U = 330 mV) did not lead to a significant additional swelling of the membrane; if at all, the resulting effects are too small to be visualized with this optical spatial resolution. This swelling process is reversible as can be seen from Fig. 1e. After 1 h at OCV following the load-cycle changes, the membrane exhibited the same thickness as before the load-cycle changes. There are three possible explanations for the increase in membrane thickness: (i) the dilution/hydration of phosphoric acid in the membrane by product water and (ii) a transport of phosphoric acid from pores of the electrodes into the membrane or (iii) a combination of (i) and (ii). In order to determine which of the aforementioned effects has the strongest impact on the increase in membrane thickness during loadcycle changes, the changes of the transmission signals (i.e. grey values) in the MEA compared to the steady-state OCV condition were analyzed quantitatively. Therefore, different line scans with a vertical width of about 1000 pixel have been evaluated for OCV, 140 mA cm−2, 300 mA cm−2 and 550 mA cm−2 (Fig. 2). Additionally, these scans were filtered by a Gaussian blur. The grey values of these normalized radiographs directly correspond to a change in transmission: a grey value N 1 represents an increase in transmission, while a grey value b 1 refers to a decrease in transmission. A grey value equal to one declares that there are no differences between the reference image at OCV and the actual operating condition. Fig. 2 shows that switching the cell from OCV to a current density of j = 140 mA cm−2 results in an increased transmission in the membrane (M). This result can be explained by the formation of product water, which leads to a hydration of the pyrophosphoric/ phosphoric acid and thus to a swelling of the membrane (conf. Eqs. (1) and (2)). The attenuation coefficient of water (0.1572 cm−1 at 30 keV) is one order of magnitude smaller than that of phosphoric
1438
W. Maier et al. / Electrochemistry Communications 12 (2010) 1436–1438
lead to a grey value smaller than one. The electrodes at the catalyst layer/GDL interface have probably not been completely saturated before, because a significant amount of the phosphoric acid diffused into the pores of the electrodes close to the membrane and into the membrane itself during the break-in procedure (see also [4]), thus a huge part of the pores at the GDL side remained nearly empty. Filling of these pores with phosphoric acid due the fact that the acid phase expands in volume as a consequence of its hydration by product water leads to an increase in absorption in the beam direction and thus to the observed decrease in transmission. 4. Conclusion
Fig. 2. Changes of the transmission through the MEA (grey values) compared to steadystate conditions at OCV for different operating conditions: at OCV (solid line), 140 mA cm−2 (dashed line), 300 mA cm−2 (dotted line) and 550 mA cm−2 (dot and dash line). The thicknesses of the membrane and the catalyst layers when current is drawn from the cell (taken from the original radiographs, cf. inset in Fig. 1) are indicated by the grey shaded areas.
acid (1.0198 cm−1 at 30 keV) at the same X-ray energy [15]. Thus, an increase of water in the membrane as well as a build-up of hydration products of phosphoric acid leads to a higher transmission due to a change in membrane/electrolyte composition. Recently, a similar mechanism has been proposed to describe changes in the impedance of a HT-PEFC (operated at identical conditions as used here) during load-cycle changes [9]. But likewise the membrane thickness does not increase with a further rising of the current density as described before, the transmission within the membrane does not further increase with increasing current density. Hence, it is clearly visible that the biggest effect in membrane swelling as well as in transmission change can be seen when switching the cell from a currentless operating condition to one where current is drawn and thus the formation of product water begins. The further increase of the current does not lead to any further visible effects. This may be due to the fast hydration process at the beginning of the water production because of the hygroscopic nature of phosphoric acid. Fig. 2 shows also an effect within the catalyst layers and the GDLs. Due to a flux of product water from the cathode to the anode about 20% of the product water of a HT-PEFC exits the cell on the anode side under these operating conditions. Accordingly, hydration processes are visible both on the cathode and on the anode side. In the vicinity of the membrane an increase in X-ray transmission can be seen, whereas further away from the membrane, in the vicinity of the GDLs, the transmission within the anode A and the cathode C decreases when switching the operation mode from OCV to a current density of j = 140 mA cm−2. A possible reason could be the hydration of phosphoric acid (and its dehydration products) within the pores of the electrodes in the vicinity of the membrane which leads to an increase in transmission, similar to the effects within the membrane described before. At the same time at the electrode/GDL interface an increased degree of filling of the pores with phosphoric acid would
In-situ synchrotron X-ray radiography was applied for the first time to study (a dynamically) operated HT-PEFC. Analyses of the Xray transmission through the cell revealed significant changes of the intensity distribution during load-cycle changes. The formation of product water after switching from currentless operation to j = 140 mA cm−2 led to a (reversible) ~20% swelling of the membrane which is explained by a hydration/dilution of the phosphoric acidbased electrolyte. The same process is found to take place in that part of the electrodes being close to the membrane, furthermore a filling of pores is assumed to happen in outer parts of the catalyst layers. These effects are fully reversible, whereas a further increase of the current density (from j = 140 to 550 mA cm−2) did not cause any significant detectable changes. Generally, synchrotron X-ray radiography can be used as novel and efficient analytical tool to gain information about the local composition of HT-PEFCs in operando at all relevant operating conditions. It can thus help to optimize the MEA design to increase the performance and to understate degradation mechanisms. To further verify these conclusions, a combination of the radiographic measurements with in-situ performed time-dependent impedance spectroscopy is currently under way. In the future, the impact of GDL materials on the acid distribution within the MEA as well as different acid management strategies will be studied. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15]
Y.L. Ma, J.S. Wainright, M.H. Litt, R.F. Savinell, J. Electrochem. Soc. 151 (2004) A8. J. Mader, L. Xiao, T.J. Schmidt, B.C. Benicewicz, Adv. Polym. Sci. 216 (2008) 63. C. Wannek, B. Kohnen, H.F. Oetjen, H. Lippert, J. Mergel, Fuel Cells 8 (2008) 87. C. Wannek, I. Konradi, J. Mergel, W. Lehnert, Int. J. Hydrogen Energy 34 (2009) 9479. R.F. Jameson, J. Chem. Soc. (1959) 752. D.-T. Chin, H. Chang, J. Appl. Electrochem. 19 (1989) 95. Y. Oono, A. Sounai, M. Hori, J. Power Sources 189 (2009) 943. K. Kwon, J.O. Park, D.Y. Yoo, J.S. Yi, Electrochim. Acta 54 (2009) 6570. K. Wippermann, C. Wannek, H.-F. Oetjen, J. Mergel, W. Lehnert, J. Power Sources 195 (2010) 2806. C. Hartnig, I. Manke, R. Kuhn, N. Kardjilov, J. Banhart, W. Lehnert, Appl. Phys. Lett. 92 (2008) 134106-1. I. Manke, C. Hartnig, M. Grünerbel, W. Lehnert, N. Kardjilov, A. Haibel, A. Hilger, J. Banhart, H. Riesemeier, Appl. Phys. Lett. 90 (2007) 174105-1. C. Hartnig, I. Manke, J. Schloesser, P. Krüger, R. Kuhn, H. Riesemeier, K. Wippermann, J. Banhart, Electrochem. Commun. 11 (2009) 1559. R. Satija, D. Jacobson, M. Arif, S. Werner, J. Power Sources 129 (2004) 238. P. Boillat, D. Kramer, B. Seyfang, G. Frei, E. Lehmann, G. Scherer, A. Wokaun, Y. Ichikawa, Y. Tasaki, K. Shinohara, Electrochem. Commun. 10 (2008) 546. Data calculated with the “X-ray transmission of a solid” program, available on the homepage of the Lawrence Berkeley National Laboratory, Centre of X-ray Optics, http://henke.lbl.gov/optical_constants/filter2.html, 18.06.2010.