- Email: [email protected]
Neutron radiographic in operando investigation of water transport in polymer electrolyte membrane fuel cells with channel barriers
Energy Conversion and Management 148 (2017) 604–610
Contents lists available at ScienceDirect
Energy Conversion and Management journal homepage: www.elsevier.com/locate/enconman
Neutron radiographic in operando investigation of water transport in polymer electrolyte membrane fuel cells with channel barriers Saad S. Alrwashdeh a,b,c,⇑, Ingo Manke a, Henning Markötter a,c, Jan Haußmann d, Nikolay Kardjilov a, André Hilger a, Mohammad J. Kermani e, Merle Klages d, A.M. Al-Falahat a,b,c, Joachim Scholta d, John Banhart a,c a
Helmholtz-Zentrum Berlin für Materialien und Energie, Hahn-Meitner-Platz 1, 14109 Berlin, Germany Mechanical Engineering Department, Faculty of Engineering, Mu’tah University, P.O Box 7, Al-Karak 61710, Jordan Technische Universität Berlin, Straße des 17. Juni 135, 10623 Berlin, Germany d Zentrum für Sonnenenergie- und Wasserstoff-Forschung Baden Württemberg (ZSW), Helmholtzstraße 8, 89081 Ulm, Germany e Amirkabir University of Technology (AUT), Tehran Polytechnic, 424 Hafez Ave., Tehran 15875-4413, Iran b c
a r t i c l e
i n f o
Article history: Received 5 February 2017 Received in revised form 8 May 2017 Accepted 11 June 2017
Keywords: Polymer electrolyte membrane fuel cell Water transport Neutron radiography Flow field design
a b s t r a c t We present a study on a new type of flow field channel design for polymer electrolyte membrane fuel cells (PEMFCs). Small barriers have been implemented into the flow field channels that force the gas flow to move through the gas diffusion layers in order to improve the supply of the catalyst with reactant gases. We investigated the water distribution in the PEMFC with neutron imaging during operation and compared the results with a comparable reference cell without barriers. We found strong hints for an increased mechanical gas flow resistance by the barriers caused by additional liquid water agglomerations. Furthermore water distribution in the barrier flow field is much more homogenous compared to the reference cell. We assume that both effects, namely the gas flow through the GDL and the homogenous water distribution are responsible for the found performance increase of up to 10%. Ó 2017 Elsevier Ltd. All rights reserved.
1. Introduction Fuel cell technology plays a major role in offering alternative energy supplies for mobile and stationary energy users [1,2]. A common example for mobile systems is the automotive sector, where polymer electrolyte membrane fuel cells (PEMFC) are considered the most favorable fuel cell type because of their high power density and flexible operating conditions [3,4]. Fuel cells can also provide combined heat and electric power in stationary energy systems with a good efficiency [5,6]. However, there are some practical problems that still limit fuel cell use. Especially under critical operating conditions that cause flooding of a fuel cell by product water an optimization of water transport in a PEMFC can lead to an improved efficiency. Such conditions include temperatures below 60 °C as well as high currents that both can cause increased water agglomerations [7–9]. The different materials used in PEMFCs can affect water transport in the cells. Through the material properties the interaction
⇑ Corresponding author at: Helmholtz-Zentrum Berlin für Materialien und Energie, Hahn-Meitner-Platz 1, 14109 Berlin, Germany. E-mail address: [email protected] (S.S. Alrwashdeh). http://dx.doi.org/10.1016/j.enconman.2017.06.032 0196-8904/Ó 2017 Elsevier Ltd. All rights reserved.
with liquid water during operation influences cell performance. A more detailed understanding of the influence of structural and chemical properties was obtained by analyzing several different materials [10–16]. Flow field patterns machined on bipolar plates are considered one of the most important components affecting the PEMFCs performance. Many researchers have concentrated on improving the performance of PEMFCs through optimization of the channel patterns and dimensions [17–22]. One way to improve both water gas flow and water distribution in a fuel cell is to apply partial blockages to the flow field channels in order to ensure a better supply of the catalysts with reaction gases. Previous studies [23,24] have shown that such blockages spanning a part of the entire cross section of channels can affect the performance of a cell and change water transport within. Different groups have demonstrated the strong influence of changes in the flow field structure on cell performance. Wang et al. used a flow field structure that is a mixture of serpentine and interdigitated flow fields [17]. They found that interdigitated flow fields show better performance due to convection above the ribs. Perng et al. investigated the influence of obstacles in the gas flow channels on the cathode side on the performance of PEM fuel cells in a computational study [25]. They found an increase in the
S.S. Alrwashdeh et al. / Energy Conversion and Management 148 (2017) 604–610
overall cell performance at a cell voltage of 0.2 V ranging between 7.32% and 14.22%. Perng and Wu applied a tapered flow field channel with a baffle plate on the cathode side of a PEMFC for increasing the cell performance [26]. They investigated the cell performance with various gap ratios and taper ratios. They found that the tapered flow channel with a baffle blockage enhances convective heat transfer and flow velocity, which in turn improves the overall cell performance. The maximum enhancement in overall cell performance for various taper ratios while choosing a narrow gap size was 15.5%. Up to now, the water distribution in such obstacles flow fields has not yet been investigated. Neutron imaging is a measurement technique that provides unique insights into the water distribution in operating PEMFCs [27–30]. Neutron sources such as research reactors and accelerator-based spallation neutron sources produce the intense neutron beams required for efficient and practical neutron imaging. The interactions of neutrons with materials are different compared to X-rays and therefore they can be used in unique ways for nondestructive material testing [27,31]. During the past decades, neutron imaging has been successfully used in many different research fields such as, e.g., engineering, geoscience, soil physics, cultural heritage, magnetism research [32– 35]. One of the most important application fields is research on materials for energy conversion or storage [28,36–42]. Aluminium and carbon are nearly transparent for neutrons, while water strongly attenuates a neutron beam. Thus, one can measure the water content in a model PEMFC constructed from the same materials as a real working component. Because of this, many questions related to water management can be studied applying neutron imaging, ranging from the influence of flow field geometry and channel design, freeze phenomena, and properties of gas diffusion media [43–52]. In this study, neutron radiography was used for in-operando investigations of water distributions in the channel system of PEMFCs. A novel approach with partial barriers in the flow field channels that force the reactant gas to flow through the GDL allows for a better supply of the catalyst layer with the reactants. The influence of the flow field channels on the water distribution was studied. 2. Experiments 2.1. Neutron radiography Fig. 1 shows the imaging instrument CONRAD (COld Neutron RADiography) used for our investigations [53–56]. Neutrons are transported from the 10 MW research reactor BER II to the experiment through neutron guides. The neutron beam at CONRAD is polychromatic with wavelengths mainly between 2 and 6 Å and a maximum intensity at about 3.0 Å. The radiographic measurements of the fuel cells discussed here are performed 10 m behind
605
the end of the neutron guide. The large distance allows for a good spatial resolution of around 200 mm because of the high L/D ratio of 330 when a pinhole of 3 cm diameter is used. Behind the sample, a detector system consisting of a scintillator, a mirror, a lens system and a CCD camera is positioned as close as possible to the sample [53,54,57]. When the neutrons hit the scintillator, photons in the visible spectrum are emitted. The scintillator used for the radiography measurements was a 200-mm thick lithium fluoride crystal with silver-doped zinc sulfide (6LiF/ZnS (Ag)). The photons are projected onto the camera by a mirror/lens system. The Andor DW436 camera used contains a 16 bit chip with (2048 2048) pixels, each with a size of (13.5 13.5) mm2. The CCD sensor is continuously cooled to below 50 °C to ensure a thermal noise as low as possible. With the optics used, an imaging field of view of (108 108) mm2 with a pixel size of 56 mm was achieved. The pixel size applied is enough to resolve the water distribution inside the cells in sufficient detail. Each radiographic projection is acquired with an exposure time of 2 s. 2.2. Fuel cell setup Two different cell designs were chosen. In both cases, control of the cell temperature was carried out by using a cooling circuit filled with deuterium oxide (D₂O). Compared to hydrogen, the attenuation coefficient of deuterium is much smaller [27,58–61]. As a result, D₂O hardly attenuates the neutron beam and can only be seen faintly in radiographs. The flow field of the cooling circuit is embedded in the backside of the bipolar plates and is connected with a secondary water coolant circuit via a heat exchanger. A modified PEMFC was used with a 100 cm2 large meandershaped flow field design (two U-turns), comprising a group of 23 channels provided with repeated barriers, which are small regions of the channels of reduced cross sectional areas. The shape and location of the barriers within a flow field channel is schematically shown in Fig. 2A. The channels were 0.8 mm deep and 0.6 mm wide. Between neighboring channels the positions of the barriers are shifted with respect to each other (Fig. 3F), thus enforcing cross rib diffusion, as described e.g. in [62,63]. The modified cell is compared with a cell containing flow field channels of a cross-sectional area of (0.6 0.6) mm2 without any barriers (see Fig. 2B). Cathode utilization curves were performed to investigate the influence of a low gas flow rate on the water content inside the channels and thus on fuel cell performance. The relative humidity was set to a value of 20% at the inlets of anode and cathode for both cells. The product water enriches the humidity to being able to observe liquid water in the channel system. The pressure drop in the cell with barriers has increased from 0.14 to 0.2 bar compared to the cell without barriers. The operating conditions are given in Table 1. Both reference and modified fuel cell showed an almost stable performance down to cathode gas flow stoichiometries as low as 1.1.
Fig. 1. Layout of the cold neutron radiography beamline ‘CONRAD’.
606
S.S. Alrwashdeh et al. / Energy Conversion and Management 148 (2017) 604–610
Fig. 2. Schematic drawing of the two different investigated flow fields. A. with barriers, B. without barriers.
Fig. 3. Water distribution in the PEM fuel cell without barriers (A), cutout from A (B), flow field design of cell used in A (C). Water distribution in the PEM fuel cell with flow field barriers (D), cutout from D (E), flow field design of cell used in D (F).
S.S. Alrwashdeh et al. / Energy Conversion and Management 148 (2017) 604–610 Table 1 Operating conditions of the investigated cells. Operating parameter
Value
Current density Stack temperature Back pressure Anode gas supply Cathode gas supply Relative gases humidity Anode gas stoichiometry Cathode gas stoichiometry
1.2 A/cm2 60 °C 0 bar 100% H2 100% air 20% 2.0 3.0
3. Results and discussion Fig. 3 displays normalized neutron radiographic images of the two investigated cells. The images show the water distributions in the cells without and with barriers (Fig. 3A and D) during operation and additionally cutouts of both images (B and E). The schematic drawings in Fig. 3C and F shows the layout of the flow field channels systems and the locations of the barriers in the corresponding cutouts in Fig. 3B and E.
607
Water agglomerations around many of the barriers can be clearly seen in Fig. 3D and E, when Fig. 3E is compared to the layout in Fig. 3F. Most of the barriers become (indirectly) visible due to accumulated (liquid) water. For better visualization, some of these areas with water around the barriers are highlighted in red in Fig. 3D–F. In Fig. 4, the water agglomerations around channel barriers and their corresponding time dependent behavior are analyzed. The figure shows normalized neutron radiographic images of the fuel cells with and without barriers (Fig. 4A and H) and water accumulation around some selected flow fields with and without barriers (see Fig. 4A, B and H, red squares) at different times during operation (Fig. 4C–G). (Fig. 4B is only used for illustration). It shows the same area marked in red in Fig. 4A in a non-normalized neutron radiographic image in which the flow field channels can be clearly seen. Fig. 4C–G reveals that the water distribution around the barriers fluctuates during operation. The graph in Fig. 4I shows the water volume versus operation time for two positions inside the cut out shown in Fig. 4C, one in front of and one behind a selected barrier with respect to the flow direction, both marked with numbers 1 and 2 respectively. A minimum quantity of water is always
Fig. 4. Water distribution in the fuel cell equipped with barrier in the channels (A), non-normalized neutron radiographic images that shows the flow field channels around the red marked location in A (B), cutout showing water evolution at the barriers (C–G), water distribution in the fuel cell without barriers (H), and a graph representing the water volume versus time at positions 1 and 2 as marked in C and at the location marked by a green square in H (I). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
608
S.S. Alrwashdeh et al. / Energy Conversion and Management 148 (2017) 604–610
Fig. 5. Schematic drawing of the water agglomerations created around the flow field barriers during operation. Schematic 3D view of a barrier in the channel with agglomerated water (A) and a corresponding cross sectional view (B). For comparison, a neutron radiographic image from such water agglomerations is shown in (C). Cross section taken from the image in C that shows the water thickness around a selected barrier (D).
found in front of and behind the barriers in the channel but a bigger amount accumulates behind the barriers (location 2 in Fig. 4C). To illustrate this, a scheme of the flow field design with the barriers and the water agglomerations found in front of and behind the barriers is shown in Fig. 5B. The gases need to pass the barriers through the GDL, which ensures a better gas supply for the catalyst in these regions. However, at the step in front of as well as behind the barrier water tends to accumulate, which enlarges the area in which the gases need to pass through the GDL. The plots of the water volume in front of and behind the barriers (Fig. 4I) as a function of time exhibit a quasi-periodic behavior: About every 250 s, a new peak of water agglomeration appears (see dotted blue vertical lines in Fig. 4I), after which a large part of the accumulated water is carried away again. Furthermore, the accumulated product water in front of and behind the barriers is build up and carried away at the same time. This process is reflected by synchronous sharp drops of the plotted water volumes (compare red and black lines in Fig. 4I). The water volume at the locations with flow field barriers is about twice as high as at a typical location within the cell without barriers as noted in Fig. 4I (green line in blue rectangle). Furthermore, after turning the cell off, the water accumulations disappear. In the following, the homogeneity of the overall water distribution in both investigated fuel cells is analyzed. Fig. 6 shows calculated water thickness images of the fuel cells without and with barriers (A and B) and the overall water thickness summed up along the y axis (see coordinate system in Fig. 6A) for both investigated cells (black line and smoothed red1 in Fig. 6C and D). Here,
1 For interpretation of color in Figs. 3 and 6, the reader is referred to the web version of this article.
the averaged water thickness W.Th. at a given horizontal position x is defined as
Z W:Th:ðyÞ ¼
xmax
Dðx; yÞdx x¼0
where D(x,y) is the local water thickness (or depth). The average water amount of the cell with barriers is 0.063 mm which is much more (about 165%) than in the cell without barriers, which has an average water amount of about 0.024 mm (Fig. 6C and D). As can already be seen in Fig. 6A, C and B, D, the homogeneity of the water distribution for both cells is very different. To assess this, the local mean water amount (along the y axis) is expressed as a fraction of the overall mean water amount in the corresponding cell by dividing the values in Fig. 6C and D by the corresponding mean water amount, i.e. the whole water amount for the corresponding cell. In the cell without barriers, the difference between the lowest and the highest value is about 270% of the average value (Fig. 6E) while the same value is only about 60% for the flow field containing barriers. Much more water could be found in the flow field cell containing barriers than the flow field cell without barriers (see Fig. 4I and Fig. 6C and D) but this water is homogenously distributed. No large local water agglomerations (except directly at the barriers, see below) or locations with very small water amount could be found that may yield as bottle necks either affecting the gas stream (e.g. flooding) or the membrane humidification (dehydration). In contrast, in the reference cell without barriers much less water is found in the top area compared to the bottom area (Fig. 6A) (This effect is mostly caused by gas flow directions and gravity). It can be assumed that this may reduce humidification of the membrane in the top area and cause a dehydration effect. On the other hand, the larger overall water amount in the flow field containing
S.S. Alrwashdeh et al. / Energy Conversion and Management 148 (2017) 604–610
609
Fig. 6. Neutron radiographic water depth image of the cells without and with barriers (A and B) representing the mean water distribution over about 20 min of operation, accumulated water amount along the Y axis (C and D) and water distribution as in C and D, but as a percentage of the mean water amount in the entire cell (E and F).
barriers might also improve membrane humidity and thus ion conductivity at most locations in the cell. Small water agglomerations were predominantly found around the barriers, which supports the idea that the function of the barriers is to force the gas to flow through the GDLs. Water agglomerations at the barriers enlarge the barriers, i.e. they may then just act like larger barriers. Barrier together with the surrounding water agglomeration forces the gas flow through the GDL and therefore increase the content of oxygen at the catalyst layer. This is due to a synergy effect between diffusion and convection mechanisms that transports oxygen from channels to the catalyst layer. The water distribution at both sides of the barriers is different. Behind the barriers (seen in the flow direction) more water can be found than in front of them, which was expected because the barriers partially screen the gas flow [63]. The water is continuously exchanged. Once the water droplets in front of and behind the barriers reach certain sizes they are taken away by the gas stream at the same time. After this, new agglomerations start to grow on both sides again. This dynamic and almost periodic behavior (Fig. 4I) is very similar to effects found for water agglomerations in GDL holes or MPL cracks [64–66].
4. Conclusions/summary Liquid water dynamics inside a specially adapted PEMFC with a flow field channel system containing barriers was studied. We compared this cell with a reference cell with a comparable flow field design but without barriers. We found that such flow field barriers have a strong influence on the overall water distribution and transport dynamics. The barriers have a considerable effect on cell performance that was up to 10% better compared to the reference cell. The following points can be assumed to be among the main reasons for this performance increase: 1. Barriers together with water agglomerations around them force the gas stream through the GDL and thereby improve catalyst gas supply without closing the entire channel ensuring still sufficient liquid water and gas transport through the channels at the same time 2. The homogenous water distribution in the flow field containing barriers avoids functional loss, such as flooded or dehydrated areas
610
S.S. Alrwashdeh et al. / Energy Conversion and Management 148 (2017) 604–610
3. The overall higher amount of liquid water in the cell containing barriers may help to improve membrane ion conductivity at most locations without the risk of flooding. This study provides first insights into the behavior of barriers in flow field channels in an operating fuel cell and will serve as a base for future calculations of water distributions in such flow fields. Especially the cyclic dynamic behavior of water agglomeration around barriers and its impact on the gas stream could be investigated in more detail in the future. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33]
Hoogers G. in, Boca Raton, FL: CRC Press LLC; 2003. Vielstich W,Lamm A, Gasteiger HA. in, Chichester: John Wiley & Sons; 2003. Mehta V, Cooper JS. J Power Sources 2003;114:32–53. Andreas Friedrich K, Carrette L, Stimming U, Fuel cells: principles, types, fuels, and applications. Weinheim: WILEY-VCH Verlag GmbH; 2000. Barelli L, Bidini G, Gallorini F, Ottaviano A. Appl Energy 2012;91:13–28. Gigliucci G, Petruzzi L, Cerelli E, Garzisi A, La Mendola A. J Power Sources 2004;131:62–8. Bazylak A. Int J Hydrogen Energy 2009;34:3845–57. Krüger P, Markötter H, Haussmann J, Klages M, Arlt T, Banhart J, et al. J Power Sources 2011;196:5250–5. Markotter H, Alink R, Haussmann J, Dittmann K, Arlt T, Wieder F, et al. Int J Hydrogen Energy 2012;37:7757–61. Cindrella L, Kannan AM. J Power Sources 2009;193:447–53. Cindrella L, Kannan AM, Ahmad R, Thommes M. Int J Hydrogen Energy 2009;34:6377–83. Mohanraju K, Sreejith V, Ananth R, Cindrella L. J Power Sources 2015;284:383–91. Cindrella L, Kannan AM, Lin JF, Saminathan K, Ho Y, Lin CW, et al. J Power Sources 2009;194:146–60. Alrwashdeh SS, Markötter H, Haußmann J, Arlt T, Klages M, Scholta J, et al. Energy 2016;102:161–5. Alink R, Haußmann J, Markötter H, Schwager M, Manke I, Gerteisen D. J Power Sources 2013;233:358–68. Haußmann J, Markötter H, Alink R, Bauder A, Dittmann K, Manke I, et al. J Power Sources 2013;239:611–22. Wang X-D, Duan Y-Y, Yan W-M, Peng X-F. Electrochim Acta 2008;53:5334–43. Boddu R, Marupakula UK, Summers B, Majumdar P. J Power Sources 2009;189:1083–92. Yan W-M, Liu H-C, Soong C-Y, Chen F, Cheng CH. J Power Sources 2006;161:907–19. Wang L, Liu H. J Power Sources 2004;134:185–96. Kumar A, Reddy RG. J Power Sources 2003;113:11–8. Kumar A, Reddy RG. J Power Sources 2006;155:264–71. Liu H-C, Yan W-M, Soong C-Y, Chen F. J Power Sources 2005;142:125–33. Wang X-D, Duan Y-Y, Yan W-M. J Power Sources 2007;173:210–21. Perng S-W, Wu H-W, Jue T-C, Cheng K-C. Appl Energy 2009;86:1541–54. Perng S-W, Wu H-W. Appl Energy 2011;88:52–67. Banhart J. in, Oxford, UK: Oxford University Press; 2008. Bellows RJ, Lin MY, Arif M, Thompson AK, Jacobson D. J Electrochem Soc 1999;146:1099–103. Arif M, Jacobson DL, Hussey DS. Neutron imaging study of the water transport in operating fuel cells. In: FY 2006 Annual Progress Report; 2006. p. 875–877. Manke I, Hartnig C, Grunerbel M, Kaczerowski J, Lehnert W, Kardjilov N, et al. Appl Phys Lett 2007;90:184101. Vontobel P, Lehmann EH, Hassanein R, Frei G. Phys B 2006;385–386:475–80. Kardjilov N, Manke I, Hilger A, Strobl M, Banhart J. Mater Today 2011;14:248–56. Griesche A, Dabah E, Kannengiesser T, Kardjilov N, Hilger A, Manke I. Acta Mater 2014;78:14–22.
[34] Dawson M, Manke I, Kardjilov N, Hilger A, Strobl M, Banhart J. New J Phys 2009;11:043013. [35] Woracek R, Penumadu D, Kardjilov N, Hilger A, Boin M, Banhart J, et al. Adv Mater 2014;26:4069–73. [36] Boillat P, Frei G, Lehmann EH, Scherer GG, Wokaun A. Electrochem Solid-State Lett 2010;13:B25–7. [37] Boillat P, Kramer D, Seyfang BC, Frei G, Lehmann E, Scherer GG, et al. Electrochem Commun 2008;10:546–50. [38] Chen Y-S, Peng H, Hussey DS, Jacobson DL, Tran DT, Abdel-Baset T, et al. J Power Sources 2007;170:376–86. [39] Gagliardo JJ, Owejan JP, Trabold TA, Tighe TW. Nucl Instrum Methods Phys Res Sect A: Accel, Spectrometers, Detectors Assoc Equip 2009;605:115–8. [40] Gondek L, Selvaraj NB, Czub J, Figiel H, Chapelle D, Kardjilov N, et al. Int J Hydrogen Energy 2011;36:9751–7. [41] Herbrig K, Pohlmann C, Gondek L, Figiel H, Kardjilov N, Hilger A, et al. J Power Sources 2015;293:109–18. [42] Pohlmann C, Herbrig K, Gondek L, Kardjilov N, Hilger A, Figiel H, et al. J Power Sources 2015;277:360–9. [43] Zhang J, Kramer D, Shimoi R, Ono Y, Lehmann E, Wokaun A, et al. Electrochim Acta 2006;51:2715–27. [44] Zhu X, Sui PC, Djilali N. J Power Sources 2007;172:287–95. [45] Turhan A, Heller K, Brenizer JS, Mench MM. J Power Sources 2006;160:1195–203. [46] Schröder A, Wippermann K, Lehnert W, Stolten D, Sanders T, Baumhöfer T, et al. J Power Sources 2010;195:4765–71. [47] Hickner MA, Siegel NP, Chen KS, Hussey DS, Jacobson DL, Arif M. J Electrochem Soc 2008;155:B427–34. [48] Boillat P, Biesdorf J, Oberholzer P, Kaestner A, Schmidt TJ. J Electrochem Soc 2014;161:F192–8. [49] Santamaria A, Tang H-Y, Park JW, Park G-G, Sohn Y-J. Int J Hydrogen Energy 2012;37:10836–43. [50] Lange A, Kupsch A, Hentschel MP, Manke I, Kardjilov N, Arlt T, et al. J Power Sources 2010;196:5293–8. [51] Klages M, Enz S, Markotter H, Manke I, Kardjilov N, Scholta J. J Power Sources 2013;239:596–603. [52] Cho KT, Turhan A, Lee JH, Brenizer JS, Heller AK, Shi L, et al. Nucl Instrum Methods Phys Res Sect A: Accel, Spectrometers, Detectors Assoc Equip 2009;605:119–22. [53] Kardjilov N,Dawson M,Hilger A,Manke I,Strobl M,Penumadu D, Kim KH,GarciaMoreno F,Banhart J. Nucl Instrum Methods Phys Res Sect A: Accel, Spectrometers, Detectors Assoc Equip, 2011;651:95–99. [54] Kardjilov N, Hilger A, Manke I, Woracek R, Banhart J. J Appl Crystallogr 2016;49:195–202. [55] Kardjilov N, Hilger A, Manke I, Strobl M, Dawson M, Williams S, et al. Nucl Instrum Methods Phys Res Sect A: Accel, Spectrometers, Detectors Assoc Equip 2011;651:47–52. [56] Totzke C, Manke I, Hilger A, Choinka G, Kardjilov N, Arlt T, et al. J Power Sources 2011;196:4631–7. [57] Strobl M, Kardjilov N, Hilger A, Kühne G, Frei G, Manke I. Nucl Instrum Methods Phys Res Sect A: Accel, Spectrometers, Detectors Assoc Equip 2009;604:640–5. [58] Manke I, Hartnig C, Kardjilov N, Messerschmidt M, Hilger A, Strobl M, et al. Appl Phys Lett 2008;92:244101. [59] Cho KT, Mench MM. Phys Chem Chem Phys 2012;14:4296–302. [60] Arlt T, Lueke W, Kardjilov N, Banhart J, Lehnert W, Manke I. J Power Sources 2015;299:125–9. [61] Hartnig C, Manke I, Kuhn R, Kleinau S, Goebbels J, Banhart J. J Power Sources 2009;188:468–74. [62] Ghanbarian A, Kermani MJ. Energy Convers Manage 2016;110:356–66. [63] Dehsara M, Kermani MJ. J Mech Sci Technol 2014;28:365–76. [64] Markötter H, Haußmann J, Alink R, Tötzke C, Arlt T, Klages M, et al. Electrochem Commun 2013;34:22–4. [65] Manke I, Hartnig C, Grünerbel M, Lehnert W, Kardjilov N, Haibel A, et al. Appl Phys Lett 2007;90:1–3. [66] Sasabe T, Deevanhxay P, Tsushima S, Hirai S. Electrochem Commun 2011;13:638–41.
Contents lists available at ScienceDirect
Energy Conversion and Management journal homepage: www.elsevier.com/locate/enconman
Neutron radiographic in operando investigation of water transport in polymer electrolyte membrane fuel cells with channel barriers Saad S. Alrwashdeh a,b,c,⇑, Ingo Manke a, Henning Markötter a,c, Jan Haußmann d, Nikolay Kardjilov a, André Hilger a, Mohammad J. Kermani e, Merle Klages d, A.M. Al-Falahat a,b,c, Joachim Scholta d, John Banhart a,c a
Helmholtz-Zentrum Berlin für Materialien und Energie, Hahn-Meitner-Platz 1, 14109 Berlin, Germany Mechanical Engineering Department, Faculty of Engineering, Mu’tah University, P.O Box 7, Al-Karak 61710, Jordan Technische Universität Berlin, Straße des 17. Juni 135, 10623 Berlin, Germany d Zentrum für Sonnenenergie- und Wasserstoff-Forschung Baden Württemberg (ZSW), Helmholtzstraße 8, 89081 Ulm, Germany e Amirkabir University of Technology (AUT), Tehran Polytechnic, 424 Hafez Ave., Tehran 15875-4413, Iran b c
a r t i c l e
i n f o
Article history: Received 5 February 2017 Received in revised form 8 May 2017 Accepted 11 June 2017
Keywords: Polymer electrolyte membrane fuel cell Water transport Neutron radiography Flow field design
a b s t r a c t We present a study on a new type of flow field channel design for polymer electrolyte membrane fuel cells (PEMFCs). Small barriers have been implemented into the flow field channels that force the gas flow to move through the gas diffusion layers in order to improve the supply of the catalyst with reactant gases. We investigated the water distribution in the PEMFC with neutron imaging during operation and compared the results with a comparable reference cell without barriers. We found strong hints for an increased mechanical gas flow resistance by the barriers caused by additional liquid water agglomerations. Furthermore water distribution in the barrier flow field is much more homogenous compared to the reference cell. We assume that both effects, namely the gas flow through the GDL and the homogenous water distribution are responsible for the found performance increase of up to 10%. Ó 2017 Elsevier Ltd. All rights reserved.
1. Introduction Fuel cell technology plays a major role in offering alternative energy supplies for mobile and stationary energy users [1,2]. A common example for mobile systems is the automotive sector, where polymer electrolyte membrane fuel cells (PEMFC) are considered the most favorable fuel cell type because of their high power density and flexible operating conditions [3,4]. Fuel cells can also provide combined heat and electric power in stationary energy systems with a good efficiency [5,6]. However, there are some practical problems that still limit fuel cell use. Especially under critical operating conditions that cause flooding of a fuel cell by product water an optimization of water transport in a PEMFC can lead to an improved efficiency. Such conditions include temperatures below 60 °C as well as high currents that both can cause increased water agglomerations [7–9]. The different materials used in PEMFCs can affect water transport in the cells. Through the material properties the interaction
⇑ Corresponding author at: Helmholtz-Zentrum Berlin für Materialien und Energie, Hahn-Meitner-Platz 1, 14109 Berlin, Germany. E-mail address: [email protected] (S.S. Alrwashdeh). http://dx.doi.org/10.1016/j.enconman.2017.06.032 0196-8904/Ó 2017 Elsevier Ltd. All rights reserved.
with liquid water during operation influences cell performance. A more detailed understanding of the influence of structural and chemical properties was obtained by analyzing several different materials [10–16]. Flow field patterns machined on bipolar plates are considered one of the most important components affecting the PEMFCs performance. Many researchers have concentrated on improving the performance of PEMFCs through optimization of the channel patterns and dimensions [17–22]. One way to improve both water gas flow and water distribution in a fuel cell is to apply partial blockages to the flow field channels in order to ensure a better supply of the catalysts with reaction gases. Previous studies [23,24] have shown that such blockages spanning a part of the entire cross section of channels can affect the performance of a cell and change water transport within. Different groups have demonstrated the strong influence of changes in the flow field structure on cell performance. Wang et al. used a flow field structure that is a mixture of serpentine and interdigitated flow fields [17]. They found that interdigitated flow fields show better performance due to convection above the ribs. Perng et al. investigated the influence of obstacles in the gas flow channels on the cathode side on the performance of PEM fuel cells in a computational study [25]. They found an increase in the
S.S. Alrwashdeh et al. / Energy Conversion and Management 148 (2017) 604–610
overall cell performance at a cell voltage of 0.2 V ranging between 7.32% and 14.22%. Perng and Wu applied a tapered flow field channel with a baffle plate on the cathode side of a PEMFC for increasing the cell performance [26]. They investigated the cell performance with various gap ratios and taper ratios. They found that the tapered flow channel with a baffle blockage enhances convective heat transfer and flow velocity, which in turn improves the overall cell performance. The maximum enhancement in overall cell performance for various taper ratios while choosing a narrow gap size was 15.5%. Up to now, the water distribution in such obstacles flow fields has not yet been investigated. Neutron imaging is a measurement technique that provides unique insights into the water distribution in operating PEMFCs [27–30]. Neutron sources such as research reactors and accelerator-based spallation neutron sources produce the intense neutron beams required for efficient and practical neutron imaging. The interactions of neutrons with materials are different compared to X-rays and therefore they can be used in unique ways for nondestructive material testing [27,31]. During the past decades, neutron imaging has been successfully used in many different research fields such as, e.g., engineering, geoscience, soil physics, cultural heritage, magnetism research [32– 35]. One of the most important application fields is research on materials for energy conversion or storage [28,36–42]. Aluminium and carbon are nearly transparent for neutrons, while water strongly attenuates a neutron beam. Thus, one can measure the water content in a model PEMFC constructed from the same materials as a real working component. Because of this, many questions related to water management can be studied applying neutron imaging, ranging from the influence of flow field geometry and channel design, freeze phenomena, and properties of gas diffusion media [43–52]. In this study, neutron radiography was used for in-operando investigations of water distributions in the channel system of PEMFCs. A novel approach with partial barriers in the flow field channels that force the reactant gas to flow through the GDL allows for a better supply of the catalyst layer with the reactants. The influence of the flow field channels on the water distribution was studied. 2. Experiments 2.1. Neutron radiography Fig. 1 shows the imaging instrument CONRAD (COld Neutron RADiography) used for our investigations [53–56]. Neutrons are transported from the 10 MW research reactor BER II to the experiment through neutron guides. The neutron beam at CONRAD is polychromatic with wavelengths mainly between 2 and 6 Å and a maximum intensity at about 3.0 Å. The radiographic measurements of the fuel cells discussed here are performed 10 m behind
605
the end of the neutron guide. The large distance allows for a good spatial resolution of around 200 mm because of the high L/D ratio of 330 when a pinhole of 3 cm diameter is used. Behind the sample, a detector system consisting of a scintillator, a mirror, a lens system and a CCD camera is positioned as close as possible to the sample [53,54,57]. When the neutrons hit the scintillator, photons in the visible spectrum are emitted. The scintillator used for the radiography measurements was a 200-mm thick lithium fluoride crystal with silver-doped zinc sulfide (6LiF/ZnS (Ag)). The photons are projected onto the camera by a mirror/lens system. The Andor DW436 camera used contains a 16 bit chip with (2048 2048) pixels, each with a size of (13.5 13.5) mm2. The CCD sensor is continuously cooled to below 50 °C to ensure a thermal noise as low as possible. With the optics used, an imaging field of view of (108 108) mm2 with a pixel size of 56 mm was achieved. The pixel size applied is enough to resolve the water distribution inside the cells in sufficient detail. Each radiographic projection is acquired with an exposure time of 2 s. 2.2. Fuel cell setup Two different cell designs were chosen. In both cases, control of the cell temperature was carried out by using a cooling circuit filled with deuterium oxide (D₂O). Compared to hydrogen, the attenuation coefficient of deuterium is much smaller [27,58–61]. As a result, D₂O hardly attenuates the neutron beam and can only be seen faintly in radiographs. The flow field of the cooling circuit is embedded in the backside of the bipolar plates and is connected with a secondary water coolant circuit via a heat exchanger. A modified PEMFC was used with a 100 cm2 large meandershaped flow field design (two U-turns), comprising a group of 23 channels provided with repeated barriers, which are small regions of the channels of reduced cross sectional areas. The shape and location of the barriers within a flow field channel is schematically shown in Fig. 2A. The channels were 0.8 mm deep and 0.6 mm wide. Between neighboring channels the positions of the barriers are shifted with respect to each other (Fig. 3F), thus enforcing cross rib diffusion, as described e.g. in [62,63]. The modified cell is compared with a cell containing flow field channels of a cross-sectional area of (0.6 0.6) mm2 without any barriers (see Fig. 2B). Cathode utilization curves were performed to investigate the influence of a low gas flow rate on the water content inside the channels and thus on fuel cell performance. The relative humidity was set to a value of 20% at the inlets of anode and cathode for both cells. The product water enriches the humidity to being able to observe liquid water in the channel system. The pressure drop in the cell with barriers has increased from 0.14 to 0.2 bar compared to the cell without barriers. The operating conditions are given in Table 1. Both reference and modified fuel cell showed an almost stable performance down to cathode gas flow stoichiometries as low as 1.1.
Fig. 1. Layout of the cold neutron radiography beamline ‘CONRAD’.
606
S.S. Alrwashdeh et al. / Energy Conversion and Management 148 (2017) 604–610
Fig. 2. Schematic drawing of the two different investigated flow fields. A. with barriers, B. without barriers.
Fig. 3. Water distribution in the PEM fuel cell without barriers (A), cutout from A (B), flow field design of cell used in A (C). Water distribution in the PEM fuel cell with flow field barriers (D), cutout from D (E), flow field design of cell used in D (F).
S.S. Alrwashdeh et al. / Energy Conversion and Management 148 (2017) 604–610 Table 1 Operating conditions of the investigated cells. Operating parameter
Value
Current density Stack temperature Back pressure Anode gas supply Cathode gas supply Relative gases humidity Anode gas stoichiometry Cathode gas stoichiometry
1.2 A/cm2 60 °C 0 bar 100% H2 100% air 20% 2.0 3.0
3. Results and discussion Fig. 3 displays normalized neutron radiographic images of the two investigated cells. The images show the water distributions in the cells without and with barriers (Fig. 3A and D) during operation and additionally cutouts of both images (B and E). The schematic drawings in Fig. 3C and F shows the layout of the flow field channels systems and the locations of the barriers in the corresponding cutouts in Fig. 3B and E.
607
Water agglomerations around many of the barriers can be clearly seen in Fig. 3D and E, when Fig. 3E is compared to the layout in Fig. 3F. Most of the barriers become (indirectly) visible due to accumulated (liquid) water. For better visualization, some of these areas with water around the barriers are highlighted in red in Fig. 3D–F. In Fig. 4, the water agglomerations around channel barriers and their corresponding time dependent behavior are analyzed. The figure shows normalized neutron radiographic images of the fuel cells with and without barriers (Fig. 4A and H) and water accumulation around some selected flow fields with and without barriers (see Fig. 4A, B and H, red squares) at different times during operation (Fig. 4C–G). (Fig. 4B is only used for illustration). It shows the same area marked in red in Fig. 4A in a non-normalized neutron radiographic image in which the flow field channels can be clearly seen. Fig. 4C–G reveals that the water distribution around the barriers fluctuates during operation. The graph in Fig. 4I shows the water volume versus operation time for two positions inside the cut out shown in Fig. 4C, one in front of and one behind a selected barrier with respect to the flow direction, both marked with numbers 1 and 2 respectively. A minimum quantity of water is always
Fig. 4. Water distribution in the fuel cell equipped with barrier in the channels (A), non-normalized neutron radiographic images that shows the flow field channels around the red marked location in A (B), cutout showing water evolution at the barriers (C–G), water distribution in the fuel cell without barriers (H), and a graph representing the water volume versus time at positions 1 and 2 as marked in C and at the location marked by a green square in H (I). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
608
S.S. Alrwashdeh et al. / Energy Conversion and Management 148 (2017) 604–610
Fig. 5. Schematic drawing of the water agglomerations created around the flow field barriers during operation. Schematic 3D view of a barrier in the channel with agglomerated water (A) and a corresponding cross sectional view (B). For comparison, a neutron radiographic image from such water agglomerations is shown in (C). Cross section taken from the image in C that shows the water thickness around a selected barrier (D).
found in front of and behind the barriers in the channel but a bigger amount accumulates behind the barriers (location 2 in Fig. 4C). To illustrate this, a scheme of the flow field design with the barriers and the water agglomerations found in front of and behind the barriers is shown in Fig. 5B. The gases need to pass the barriers through the GDL, which ensures a better gas supply for the catalyst in these regions. However, at the step in front of as well as behind the barrier water tends to accumulate, which enlarges the area in which the gases need to pass through the GDL. The plots of the water volume in front of and behind the barriers (Fig. 4I) as a function of time exhibit a quasi-periodic behavior: About every 250 s, a new peak of water agglomeration appears (see dotted blue vertical lines in Fig. 4I), after which a large part of the accumulated water is carried away again. Furthermore, the accumulated product water in front of and behind the barriers is build up and carried away at the same time. This process is reflected by synchronous sharp drops of the plotted water volumes (compare red and black lines in Fig. 4I). The water volume at the locations with flow field barriers is about twice as high as at a typical location within the cell without barriers as noted in Fig. 4I (green line in blue rectangle). Furthermore, after turning the cell off, the water accumulations disappear. In the following, the homogeneity of the overall water distribution in both investigated fuel cells is analyzed. Fig. 6 shows calculated water thickness images of the fuel cells without and with barriers (A and B) and the overall water thickness summed up along the y axis (see coordinate system in Fig. 6A) for both investigated cells (black line and smoothed red1 in Fig. 6C and D). Here,
1 For interpretation of color in Figs. 3 and 6, the reader is referred to the web version of this article.
the averaged water thickness W.Th. at a given horizontal position x is defined as
Z W:Th:ðyÞ ¼
xmax
Dðx; yÞdx x¼0
where D(x,y) is the local water thickness (or depth). The average water amount of the cell with barriers is 0.063 mm which is much more (about 165%) than in the cell without barriers, which has an average water amount of about 0.024 mm (Fig. 6C and D). As can already be seen in Fig. 6A, C and B, D, the homogeneity of the water distribution for both cells is very different. To assess this, the local mean water amount (along the y axis) is expressed as a fraction of the overall mean water amount in the corresponding cell by dividing the values in Fig. 6C and D by the corresponding mean water amount, i.e. the whole water amount for the corresponding cell. In the cell without barriers, the difference between the lowest and the highest value is about 270% of the average value (Fig. 6E) while the same value is only about 60% for the flow field containing barriers. Much more water could be found in the flow field cell containing barriers than the flow field cell without barriers (see Fig. 4I and Fig. 6C and D) but this water is homogenously distributed. No large local water agglomerations (except directly at the barriers, see below) or locations with very small water amount could be found that may yield as bottle necks either affecting the gas stream (e.g. flooding) or the membrane humidification (dehydration). In contrast, in the reference cell without barriers much less water is found in the top area compared to the bottom area (Fig. 6A) (This effect is mostly caused by gas flow directions and gravity). It can be assumed that this may reduce humidification of the membrane in the top area and cause a dehydration effect. On the other hand, the larger overall water amount in the flow field containing
S.S. Alrwashdeh et al. / Energy Conversion and Management 148 (2017) 604–610
609
Fig. 6. Neutron radiographic water depth image of the cells without and with barriers (A and B) representing the mean water distribution over about 20 min of operation, accumulated water amount along the Y axis (C and D) and water distribution as in C and D, but as a percentage of the mean water amount in the entire cell (E and F).
barriers might also improve membrane humidity and thus ion conductivity at most locations in the cell. Small water agglomerations were predominantly found around the barriers, which supports the idea that the function of the barriers is to force the gas to flow through the GDLs. Water agglomerations at the barriers enlarge the barriers, i.e. they may then just act like larger barriers. Barrier together with the surrounding water agglomeration forces the gas flow through the GDL and therefore increase the content of oxygen at the catalyst layer. This is due to a synergy effect between diffusion and convection mechanisms that transports oxygen from channels to the catalyst layer. The water distribution at both sides of the barriers is different. Behind the barriers (seen in the flow direction) more water can be found than in front of them, which was expected because the barriers partially screen the gas flow [63]. The water is continuously exchanged. Once the water droplets in front of and behind the barriers reach certain sizes they are taken away by the gas stream at the same time. After this, new agglomerations start to grow on both sides again. This dynamic and almost periodic behavior (Fig. 4I) is very similar to effects found for water agglomerations in GDL holes or MPL cracks [64–66].
4. Conclusions/summary Liquid water dynamics inside a specially adapted PEMFC with a flow field channel system containing barriers was studied. We compared this cell with a reference cell with a comparable flow field design but without barriers. We found that such flow field barriers have a strong influence on the overall water distribution and transport dynamics. The barriers have a considerable effect on cell performance that was up to 10% better compared to the reference cell. The following points can be assumed to be among the main reasons for this performance increase: 1. Barriers together with water agglomerations around them force the gas stream through the GDL and thereby improve catalyst gas supply without closing the entire channel ensuring still sufficient liquid water and gas transport through the channels at the same time 2. The homogenous water distribution in the flow field containing barriers avoids functional loss, such as flooded or dehydrated areas
610
S.S. Alrwashdeh et al. / Energy Conversion and Management 148 (2017) 604–610
3. The overall higher amount of liquid water in the cell containing barriers may help to improve membrane ion conductivity at most locations without the risk of flooding. This study provides first insights into the behavior of barriers in flow field channels in an operating fuel cell and will serve as a base for future calculations of water distributions in such flow fields. Especially the cyclic dynamic behavior of water agglomeration around barriers and its impact on the gas stream could be investigated in more detail in the future. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33]
Hoogers G. in, Boca Raton, FL: CRC Press LLC; 2003. Vielstich W,Lamm A, Gasteiger HA. in, Chichester: John Wiley & Sons; 2003. Mehta V, Cooper JS. J Power Sources 2003;114:32–53. Andreas Friedrich K, Carrette L, Stimming U, Fuel cells: principles, types, fuels, and applications. Weinheim: WILEY-VCH Verlag GmbH; 2000. Barelli L, Bidini G, Gallorini F, Ottaviano A. Appl Energy 2012;91:13–28. Gigliucci G, Petruzzi L, Cerelli E, Garzisi A, La Mendola A. J Power Sources 2004;131:62–8. Bazylak A. Int J Hydrogen Energy 2009;34:3845–57. Krüger P, Markötter H, Haussmann J, Klages M, Arlt T, Banhart J, et al. J Power Sources 2011;196:5250–5. Markotter H, Alink R, Haussmann J, Dittmann K, Arlt T, Wieder F, et al. Int J Hydrogen Energy 2012;37:7757–61. Cindrella L, Kannan AM. J Power Sources 2009;193:447–53. Cindrella L, Kannan AM, Ahmad R, Thommes M. Int J Hydrogen Energy 2009;34:6377–83. Mohanraju K, Sreejith V, Ananth R, Cindrella L. J Power Sources 2015;284:383–91. Cindrella L, Kannan AM, Lin JF, Saminathan K, Ho Y, Lin CW, et al. J Power Sources 2009;194:146–60. Alrwashdeh SS, Markötter H, Haußmann J, Arlt T, Klages M, Scholta J, et al. Energy 2016;102:161–5. Alink R, Haußmann J, Markötter H, Schwager M, Manke I, Gerteisen D. J Power Sources 2013;233:358–68. Haußmann J, Markötter H, Alink R, Bauder A, Dittmann K, Manke I, et al. J Power Sources 2013;239:611–22. Wang X-D, Duan Y-Y, Yan W-M, Peng X-F. Electrochim Acta 2008;53:5334–43. Boddu R, Marupakula UK, Summers B, Majumdar P. J Power Sources 2009;189:1083–92. Yan W-M, Liu H-C, Soong C-Y, Chen F, Cheng CH. J Power Sources 2006;161:907–19. Wang L, Liu H. J Power Sources 2004;134:185–96. Kumar A, Reddy RG. J Power Sources 2003;113:11–8. Kumar A, Reddy RG. J Power Sources 2006;155:264–71. Liu H-C, Yan W-M, Soong C-Y, Chen F. J Power Sources 2005;142:125–33. Wang X-D, Duan Y-Y, Yan W-M. J Power Sources 2007;173:210–21. Perng S-W, Wu H-W, Jue T-C, Cheng K-C. Appl Energy 2009;86:1541–54. Perng S-W, Wu H-W. Appl Energy 2011;88:52–67. Banhart J. in, Oxford, UK: Oxford University Press; 2008. Bellows RJ, Lin MY, Arif M, Thompson AK, Jacobson D. J Electrochem Soc 1999;146:1099–103. Arif M, Jacobson DL, Hussey DS. Neutron imaging study of the water transport in operating fuel cells. In: FY 2006 Annual Progress Report; 2006. p. 875–877. Manke I, Hartnig C, Grunerbel M, Kaczerowski J, Lehnert W, Kardjilov N, et al. Appl Phys Lett 2007;90:184101. Vontobel P, Lehmann EH, Hassanein R, Frei G. Phys B 2006;385–386:475–80. Kardjilov N, Manke I, Hilger A, Strobl M, Banhart J. Mater Today 2011;14:248–56. Griesche A, Dabah E, Kannengiesser T, Kardjilov N, Hilger A, Manke I. Acta Mater 2014;78:14–22.
[34] Dawson M, Manke I, Kardjilov N, Hilger A, Strobl M, Banhart J. New J Phys 2009;11:043013. [35] Woracek R, Penumadu D, Kardjilov N, Hilger A, Boin M, Banhart J, et al. Adv Mater 2014;26:4069–73. [36] Boillat P, Frei G, Lehmann EH, Scherer GG, Wokaun A. Electrochem Solid-State Lett 2010;13:B25–7. [37] Boillat P, Kramer D, Seyfang BC, Frei G, Lehmann E, Scherer GG, et al. Electrochem Commun 2008;10:546–50. [38] Chen Y-S, Peng H, Hussey DS, Jacobson DL, Tran DT, Abdel-Baset T, et al. J Power Sources 2007;170:376–86. [39] Gagliardo JJ, Owejan JP, Trabold TA, Tighe TW. Nucl Instrum Methods Phys Res Sect A: Accel, Spectrometers, Detectors Assoc Equip 2009;605:115–8. [40] Gondek L, Selvaraj NB, Czub J, Figiel H, Chapelle D, Kardjilov N, et al. Int J Hydrogen Energy 2011;36:9751–7. [41] Herbrig K, Pohlmann C, Gondek L, Figiel H, Kardjilov N, Hilger A, et al. J Power Sources 2015;293:109–18. [42] Pohlmann C, Herbrig K, Gondek L, Kardjilov N, Hilger A, Figiel H, et al. J Power Sources 2015;277:360–9. [43] Zhang J, Kramer D, Shimoi R, Ono Y, Lehmann E, Wokaun A, et al. Electrochim Acta 2006;51:2715–27. [44] Zhu X, Sui PC, Djilali N. J Power Sources 2007;172:287–95. [45] Turhan A, Heller K, Brenizer JS, Mench MM. J Power Sources 2006;160:1195–203. [46] Schröder A, Wippermann K, Lehnert W, Stolten D, Sanders T, Baumhöfer T, et al. J Power Sources 2010;195:4765–71. [47] Hickner MA, Siegel NP, Chen KS, Hussey DS, Jacobson DL, Arif M. J Electrochem Soc 2008;155:B427–34. [48] Boillat P, Biesdorf J, Oberholzer P, Kaestner A, Schmidt TJ. J Electrochem Soc 2014;161:F192–8. [49] Santamaria A, Tang H-Y, Park JW, Park G-G, Sohn Y-J. Int J Hydrogen Energy 2012;37:10836–43. [50] Lange A, Kupsch A, Hentschel MP, Manke I, Kardjilov N, Arlt T, et al. J Power Sources 2010;196:5293–8. [51] Klages M, Enz S, Markotter H, Manke I, Kardjilov N, Scholta J. J Power Sources 2013;239:596–603. [52] Cho KT, Turhan A, Lee JH, Brenizer JS, Heller AK, Shi L, et al. Nucl Instrum Methods Phys Res Sect A: Accel, Spectrometers, Detectors Assoc Equip 2009;605:119–22. [53] Kardjilov N,Dawson M,Hilger A,Manke I,Strobl M,Penumadu D, Kim KH,GarciaMoreno F,Banhart J. Nucl Instrum Methods Phys Res Sect A: Accel, Spectrometers, Detectors Assoc Equip, 2011;651:95–99. [54] Kardjilov N, Hilger A, Manke I, Woracek R, Banhart J. J Appl Crystallogr 2016;49:195–202. [55] Kardjilov N, Hilger A, Manke I, Strobl M, Dawson M, Williams S, et al. Nucl Instrum Methods Phys Res Sect A: Accel, Spectrometers, Detectors Assoc Equip 2011;651:47–52. [56] Totzke C, Manke I, Hilger A, Choinka G, Kardjilov N, Arlt T, et al. J Power Sources 2011;196:4631–7. [57] Strobl M, Kardjilov N, Hilger A, Kühne G, Frei G, Manke I. Nucl Instrum Methods Phys Res Sect A: Accel, Spectrometers, Detectors Assoc Equip 2009;604:640–5. [58] Manke I, Hartnig C, Kardjilov N, Messerschmidt M, Hilger A, Strobl M, et al. Appl Phys Lett 2008;92:244101. [59] Cho KT, Mench MM. Phys Chem Chem Phys 2012;14:4296–302. [60] Arlt T, Lueke W, Kardjilov N, Banhart J, Lehnert W, Manke I. J Power Sources 2015;299:125–9. [61] Hartnig C, Manke I, Kuhn R, Kleinau S, Goebbels J, Banhart J. J Power Sources 2009;188:468–74. [62] Ghanbarian A, Kermani MJ. Energy Convers Manage 2016;110:356–66. [63] Dehsara M, Kermani MJ. J Mech Sci Technol 2014;28:365–76. [64] Markötter H, Haußmann J, Alink R, Tötzke C, Arlt T, Klages M, et al. Electrochem Commun 2013;34:22–4. [65] Manke I, Hartnig C, Grünerbel M, Lehnert W, Kardjilov N, Haibel A, et al. Appl Phys Lett 2007;90:1–3. [66] Sasabe T, Deevanhxay P, Tsushima S, Hirai S. Electrochem Commun 2011;13:638–41.