0
Expert View

Multiscale Tomography-Based Analysis of Polymer Electrolyte Fuel Cells: Towards a Fully Resolved Gas Diffusion Electrode Reconstruction PUBLIC ACCESS

[+] Author and Article Information
Matthias Klingele, Riko Moroni

IMTEK,
University of Freiburg,
Georges-Koehler-Allee 103,
Freiburg 79110, Germany

Severin Vierrath

IMTEK,
University of Freiburg,
Georges-Koehler-Allee 103,
Freiburg 79110, Germany;
Hahn-Schickard,
Georges-Koehler-Allee 103,
Freiburg 79110, Germany

Simon Thiele

IMTEK,
University of Freiburg,
Georges-Koehler-Allee 103,
Freiburg 79110, Germany;
Hahn-Schickard,
Georges-Koehler-Allee 103,
Freiburg 79110, Germany;
FIT,
University of Freiburg,
Georges-Koehler-Allee 105,
Freiburg 79110, Germany

1Corresponding author.

Manuscript received May 11, 2017; final manuscript received June 29, 2017; published online September 19, 2017. Assoc. Editor: Dirk Henkensmeier.

J. Electrochem. En. Conv. Stor. 15(1), 014701 (Sep 19, 2017) (7 pages) Paper No: JEECS-17-1045; doi: 10.1115/1.4037244 History: Received May 11, 2017; Revised June 29, 2017

The microstructure of a fuel cell electrode largely determines the performance of the whole fuel cell system. In this regard, tomographic imaging is a valuable tool for the understanding and control of the electrode morphology. The distribution of pore- and feature-sizes within fuel cell electrodes covers several orders of magnitude, ranging from millimeters in the gas diffusion layer (GDL) down to few nanometers in the catalyst layer. This obligates the application of various tomographic methods for imaging every aspect of a fuel cell. This perspective evaluates the capabilities, limits, and challenges of each of these methods. Further, it highlights and suggests efforts toward the integration of multiple tomographic methods into single multiscale datasets, a venture which aims at large-scale, and morphologically fully resolved fuel cell reconstructions.

FIGURES IN THIS ARTICLE
<>

Electrochemical energy converters, such as fuel cells, are key technologies for a sustainable and clean energy economy. Fuel cells and electrolyzers enable the conversion and storage of energy obtained from renewable sources, such as solar energy or wind energy via hydrogen, without producing harmful emissions. This enables easy decentral generation of clean electrical power in both stationary and mobile devices. In this regard, the automotive industry has already concluded, that fuel cells are necessary to achieve the industry-set goal for CO2 emissions of below 95 gCO2/km in wide-range personal transport [1]. The volumetric power density of the fuel cell is (amongst other factors) strongly dependent on a high specific electrode/electrolyte interfacial area. Therefore, electrodes (including gas diffusion layers (GDLs) and microporous layers (MPLs)) have been engineered toward a high specific surface area with morphological features ranging from several micrometers down to few nanometers in size [2]. A state-of-the-art fuel cell electrode is comprised of a carbonaceous matrix consisting of 30–60 nm small (often spherical) carbon particles, typically referred to as carbon black. Those particles are “glued” together by an electrolytic polymer, forming larger microporous carbon agglomerates (Figs. 1(c) and 1(d)) [3].

The electrolytic polymer, typically referred to as ionomer, is believed to form a nanometer-thin film, surrounding the carbon spheres [4]. The carbon spheres are decorated with catalytically active Pt particles of 2–5 nm in size [5]. Such electrodes provide a high specific interfacial area of electrode and electrolyte, using relatively low amounts of Pt, while providing a reasonable electrical conductivity and diffusivity for fuel and oxidant gases.

Gases and electrons are supplied to the electrode by additional, noncatalytic, gas diffusion media, whose purpose is as well to efficiently remove the reaction products such as heat and water (Fig. 1(b)) [6]. These functional layers typically consist of woven or nonwoven carbon fibers with approximately 10 μm in diameter and up to several millimeters in length. This fibrous layer, often referred to as GDL, provides electrical conductivity, as well as transport of gas and water through its pore space. Usually, the fibrous GDL is connected to the catalytically active catalyst layer by a microporous interlayer. This MPL as well consists of 30–60 nm sized carbon spheres, but is bound by polytetrafluoroethylene (PTFE) polymer, and contains no catalytically active materials.

In conclusion, the morphology of the different materials and features within a fuel cell electrode cover a wide range of sizes from few nanometers up to several millimeters. The composition, alignment, shape, and distribution of these features are crucial and ultimately determine the performance of the device. Therefore, investigation and control of the morphology of fuel cell electrodes and gas diffusion media by imaging is one of the major fields of research in energy materials. Three-dimensional morphological imaging of fuel cells enables deeper understanding and successively enables an improvement of performance-limiting mechanisms in fuel cell operation. For instance, the tortuosity of the combined pore space of GDL, MPL, and catalyst layer in a fuel cell influences the mass-transport resistance of oxygen reaching the reactive sites, ultimately limiting cell power at higher current densities [7]. Another prominent example is the composition of the carbon spheres and Pt particles within the catalyst layer. During cell operation, carbon corrosion and Pt accumulation can lead to loss of specific electrode surface area [8]. Improvements to such morphology driven failure mechanisms are best possible via direct observation by tomographic methods.

In the past couple of years, various tomography methods have been optimized for imaging different aspects of energy materials. Three-dimensional transmission electron microscopy (TEM) tomography has been utilized to image Pt catalyst particles in fuel cell catalyst layers (Fig. 1(d)) [5,9]. TEM was used due to its high resolution in the range of 0.3 nm, therefore being the only feasible method for imaging such small particles. From this, Pt particles size distributions, as well as a particle dispersion in 3D, could be obtained. Further, the composition of carbon spheres in single agglomerates as imaged in Fig. 1(d) was investigated [8,10]. This is especially interesting for imaging degradation phenomena, for which identical location TEM is oftenly used to reveal structural changes ex situ at the same location before and after aging [11,12]. In recent work, 3D TEM was further used to image an (Cs+ stained) ionomer layer surrounding single carbon spheres within such an agglomerate [4].

From such TEM measurements, we gain insight to the particle-size distribution, to the ionomer film thickness and coverage, as well as to the morphology of carbon sphere agglomerates. However, along with the high resolution of 3D TEM comes the restriction in the field of view of about approximately 500 nm [5]. This forbids imaging larger sample volumes and, especially with respect to ionomer distribution, probably does not account for a representative reconstruction. Moreover, same as any other electron-based tomography method, the obtained image contrast is dependent on the atomic number (Z) of the imaged materials. Therefore, materials of similar Z can hardly be distinguished.

Larger sample volumes of electrode material can be imaged by focused ion beam/scanning electron microscope (FIB–SEM) tomography [13]. FIB–SEM tomography works by successive FIB milling and SEM imaging. The field of view thus is restricted by the field of view of the SEM and by the amount of milling steps and the milling distance. This typically leads to imaged sample volumes of few cubic micrometers. The resolution of FIB–SEM tomography is given by the resolution of the SEM images and by the FIB milling distance. This typically leads to an anisotropic resolution of approximately 3 × 3 × 10 nm. Here, small Pt particles cannot reliably be resolved anymore (Fig. 1(c)). However, the resolution is high enough to image the single carbon spheres, and the field of view is large enough to cover several carbon agglomerates (TEM typically covers only few carbon spheres). The enhanced representativeness enables the estimation of gas diffusivity through the porous electrode, the determination of porosity, as well as pore- and grain-size distributions within the fuel cell catalyst layer [5,14].

So-called nano X-ray tomography with a resolution of 50 nm and a field of view of roughly 16 μm3 has been shown to cover a similar purpose as FIB–SEM tomography [15]. However, the slightly worse resolution forbids imaging of single carbon spheres but the method is still able to distinguish carbon agglomerates from each other, allowing estimations of porosity, pore- and grain-size distribution of the primary pore space, and the carbon agglomerates. In contrast to FIB–SEM tomography, nano X-ray tomography has the general advantage of not being destructive, mitigating any adulterations from the FIB milling process. However, particularly for resolutions of about 50 nm, samples must be specially prepared by using lift out techniques. Generally, X-ray tomography is able to cover rather large sample volumes, always while sacrificing resolution when increasing the imaged volume. For the purpose of imaging larger sample volumes, so-called micro X-ray tomography with a resolution in the range of 500 nm can be applied to image the GDL of fuel cells [16]. For imaging even larger sample volumes, which can cover the whole cross section of a fuel cell, synchrotron X-ray tomography is especially suitable (Figs. 1(a) and 1(b)). The high beam energy of the synchrotron X-ray source (up to approximately 80 keV versus 8 keV in nano X-ray tomography) yields in a high signal-to-noise ratio. This exemplarily enables a field of view of 4 mm at a resolution of 1.1 μm, which is good enough for estimating gas transport and pore morphology of the fibrous GDL in a representative sample volume.

Those often applied tomography methods are currently being complemented by novel, less established techniques, which enable a surplus of information. Serial block-face electron microscopy, for instance, is very similar to FIB–SEM tomography, but substitutes the FIB milling step by sectioning with an ultramicrotome [17]. By doing so, polymer within the sample is not altered by the FIB milling process, and polymeric samples can be analyzed. Of course, artifacts resulting from microtome sectioning need to be taken into account as well when evaluating its suitability. Already well-established in life sciences, the field of array tomography could also be of use in energy material science. This technique relies on the serial sectioning and keeping of microtome slices of a specimen, which enables subsequent imaging and analysis of the single slices by all kind of techniques [1820]. A derivate of this technique is especially promising for fuel cell imaging: Differently functionalized atomic force microscopy tips would enable novel insight to many fuel cell related open questions in 3D [21]. For instance, a hydrophobicity-tailored tip could distinguish hydrophobic and hydrophilic domains in 3D within proton exchange membranes, or electrochemically active tips could directly measure areas of different membrane conductivity (as already shown in 2D) [22]. Further, an electronically conductive tip could map regions of varying conductivity within catalyst layers, or a phase-contrast tip could distinguish ionomer from carbon particles, therefore measuring the ionomer distribution throughout a larger volume of a catalyst layer.

Regarding transport phenomena, neutron spectroscopy has been applied to image water transport in proton exchange membranes [23] and GDLs [24]. However, due to the dynamic behavior of water generation and removal, 3D imaging is difficult, as each state of humidification would have to freeze to allow imaging from different angles. Further, to date neutron spectrometers have a resolution not high enough for imaging water in catalyst layer pores. Generally, it must be noted that the performance of the above-mentioned tomography methods is a challenge of its own, and that the achievement of high-quality results can be time consuming. Especially in methods such as FIB–SEM tomography, larger sample volumes obligate more FIB milling steps, which expand production time and enhance risks for systematic errors.

The above-discussed tomography methods are able to image every relevant morphological feature of a fuel cell. From very small features, such as nanometer-sized Pt particles (TEM) up to whole GDLs in the range of several millimeters (X-ray), there are no open questions of how the different features, or layers, look like. However, the strongly varying field of view of the different methods forbids the resolution of all these features within only one tomographic dataset. This means that it is not directly possible to reveal any information about the Pt particles in a fuel cell electrode, if the imaged sample volume should include information about the gas transport through the fibrous GDL at the same time. This issue opened up the quite new research field of multiscale tomography. Research in this field aims at overcoming the technical limitations of the different tomographic methods when it comes to integration of the information of the single methods within one dataset. This is typically attempted by a combination of results from imaging with virtual design and statistics.

The combination of tomographic data from different imaging methods into single, multiscale datasets aims at a rather far aim: The ideal dataset should present a field of view, covering the whole fuel cell, at a resolution high enough to resolve even small Pt particles. This would allow reliable analysis of gas transport throughout the whole fuel cell, starting from large pores in the GDL and ending in tiny nanometer-sized pores within carbon agglomerates, also taking into account any interface-effects between the single layers. Proton and electron conductivity could be modeled analogously. Pt distribution would be known for the whole catalyst layer, and carbon, ionomer, and pore space would be distinguishable, allowing fuel cell performance predictions by pure modeling of reactant transport to and from the reactive sites at a certain reaction rate. Naturally, this is aiming at the impossible. However, by smart virtual design approaches, certain open questions in fuel cell research can be answered better by multiscale approaches as it could be when only taking into account single tomography methods. In this quest, few studies have already been presented in literature. Already in 2010, Ostadi et al. performed X-ray tomography and FIB–SEM tomography on the same gas diffusion medium, consisting of a fibrous GDL (X-ray) and MPL (FIB–SEM) [16]. Although they analyzed gas diffusion through each layer, they did not combine information of the two into an integral dataset. Combining information from X-ray and FIB–SEM tomography of fuel cell gas diffusion media was later performed by Wargo et al. [7] and Göbel et al. [25]. This is beneficial as the experimentalist is ultimately interested in the gas diffusion properties of the whole composite, including any interface-effects, and not just a single isolated layer. Further, the integral pore- and particle-size distribution would be of high interest, but cannot be assessed when treating the two layers separately. At this point, a virtual design approach could solve the issue. Morphological parameters and effective mass transport values obtained from the microporous morphology (as obtained from FIB–SEM) can be assigned to regions of the X-ray dataset where the MPLs are visible, but not resolved on the nanoscale. To give a perspective of the feasibility of such an approach, we show it exemplarily on hand of a gas diffusion electrode consisting of a fibrous GDL, a microporous layer, and a catalyst layer (for demonstrative purposes artificially created with geodict software).

Figure 2(b) shows an artificially created X-ray dataset of a gas diffusion electrode, in which the micrometer-sized fibers, the pores within the GDL, and the polymer binder (hydrophobic agent) are well resolved. However, as the resolution of X-ray tomography at such large field of views is too low, the microstructure of the MPL and catalyst layer is not resolved (red and green phases). These functional layers appear as solid phases within the X-ray dataset. The microstructure of these layers can be assessed by FIB–SEM tomography, at a much smaller field of view (Figs. 2(a) and 2(c)). Within these smaller volumes, morphology and transport parameters can be calculated and fed back to the global X-ray dataset as effective values within the respective phases. From that, multiscale analysis can be performed for the whole gas diffusion electrode.

Figure 3 shows the combined pore- and grain-size distribution of the gas diffusion electrode. This was obtained by back-feeding the pore- and grain-size distribution from FIB–SEM analysis into the global X-ray dataset while normalizing to the respective volume fractions. This allows a multiscale and spatially resolved morphological analysis. For instance, it is interesting to see that the mean particle size abruptly changes to smaller values along the depth of the gas diffusion electrode (at the interface of gas diffusion layer microporous layer), while the mean pore size reveals a much smoother transition. The latter is beneficial for gas transport, as sudden tapering of the transport paths is mitigated. In contrast, electron conductivity is insensitive for sudden changes in morphology.

Besides morphological parameters, the effective transport of gases through the whole gas diffusion electrode can be evaluated more precisely when combining information from the different imaging techniques. In the example of the gas diffusion electrode from above, the diffusivities of the MPL and the catalyst layer can be assigned as effective values to the unresolved (solid appearing) phases of MPL and catalyst layer within the X-ray reconstruction. From this, the overall diffusivity of the gas diffusion electrode can be calculated. For instance, the effective Laplace diffusivities of the MPL and the catalyst layer in through-plane direction are calculated to be 67.3% and 45.8%, respectively. The corresponding overall diffusivity of the integrated dataset is calculated to be 60.7% (units represent the percentage of diffusivity compared to the bulk value). At this point, experimental validation is necessary to evaluate the validity of such an analysis, for instance, as reported by Terao et al., where FIB–SEM tomography-based porosity, pore-size distribution, and diffusivity are validated by different experimental methods [26].

In addition to the purely tomographic analysis of the electrode morphology, also experimental techniques can be helpful. For instance, Cetinbas et al. used the differences of the pore-size distribution obtained by tomographic methods to experimental values from gas adsorption and mercury intrusion porosimetry to conclude the distribution of ionomer throughout the catalyst layer [27].

Further, besides the assignment of effective values (transport and morphology) to microscopically unresolved phases, as described previously, a direct morphological “upscaling” can be made. This was first presented in fuel cell application by Thiele et al., who integrated morphological data of Pt particles from TEM tomography into a FIB–SEM reconstruction of its carbon support matrix [5]. Information about the Pt particle-size distribution was obtained by high-resolution 3D TEM tomography and, with this information as boundary condition, Pt particles were stochastically inscribed into a FIB–SEM reconstruction of the carbonaceous phase of a fuel cell electrode.

All the previously-discussed approaches mark efforts toward direct fuel cell performance prediction, based on calculated reactant transport to 3D distributed catalytically active sites through the tortuous morphological networks of a fuel cell electrode. In this perspective, a fully resolved fuel cell reconstruction would enable the direct calculation of reactant pathways to the catalytically active Pt particles. Fully resolved in this sense means that the field of view of the reconstruction is large enough to cover a representative fuel cell volume (for instance, 2 × 2 mm in lateral size and the complete height of the fuel cell), while at the same time revealing a resolution high enough to cover even nanometer-sized features. Further, the different materials contained in the fuel cell must be distinguished. This means that the reconstruction discriminates between carbon materials (fibrous GDL, MPL, and carbon support of catalyst layer), pore space, Pt, PTFE binder, and polymer electrolyte. Ideally, the latter is even resolved in hydrophilic and hydrophobic domains, mapping areas of different proton conductivity. Further, the carbonaceous matrix should include information about the additional electric resistances arising at the interfaces of different carbon spheres or fibers.

From this, in principle, reactant transport within the respective material phase to the Pt particles could be calculated. By analyzing the reactant supply at every single Pt particle, and given a certain reaction rate, the fuel cells overall oxygen reduction reactions per time could be calculated, allowing conclusion about the total power output. This could in principle be even performed dynamically when taking into account the produced water, clogging the pore space and thus hindering gas supply. This would, for the first time, allow predictions toward fuel cell performance purely based on tomographic analysis.

However, as mentioned before, such a completely resolved fuel cell reconstruction is impossible to obtain with methods available today. Therefore, possible workarounds by statistical approaches and virtual design will be discussed in the following. The presented ideas are conclusions, combinations, and outlook to the multiscale tomography approaches presented in the Approaches for Multiscale Analysis of Fuel Cell Electrodes section. As before, we will restrict to the reconstruction of a gas diffusion electrode, consisting of fibrous GDL, MPL, and catalyst layer.

As a starting point, an X-ray tomography of a representative fuel cell volume can be obtained. This directly can image the carbon fibers (or carbon fibers plus PTFE coating, if the GDL is treated to be hydrophobic), the pores space of the gas diffusion layer, as well as potential cracks of larger size within the MPL and the catalyst layer. The MPL and catalyst layer, however, appear as solid phases. Because of the Pt content within the catalyst layer and the differing contents of PTFE/ionomer, the two solid appearing layers can be distinguished.

As a next step, the carbonaceous microporosity of the catalyst layer and MPL must be added to the global dataset. By FIB–SEM tomography, information about the formation of carbon agglomerates, and carbon spheres within these agglomerates, can be obtained. This information can be morphologically integrated into the global dataset by two different ways. First, the much smaller FIB–SEM reconstruction could be inscribed directly by stringing together the reconstructed volume until the respective layer in the global X-ray dataset is “filled-up” (Fig. 4(a)).

Special care needs to be taken considering the interface of the single FIB–SEM reconstructions. Here, morphological interpolation could avoid abrupt changes in morphology, possibly adulterating calculated transport parameters across this interface. Further, the FIB–SEM reconstruction should be large enough to cover a representative volume element. The second way to inscribe microporosity to the catalyst layer and the microporous layer is by a statistical approach (Fig. 4(b)). Carbon spheres could be inscribed stochastically, so that the morphological parameters obtained by the FIB–SEM reconstruction are fulfilled. Hence, carbon spheres could be placed into the respective layer of the global dataset in a way that this layer matches the porosity, the tortuosity, as well as pore- and grain-size distribution of the FIB–SEM reconstruction. This stochastic approach has already been shown for catalyst layers in many studies [2831]. Here, it should be noted, that for both of the described approaches, the resolution of the global dataset would be in such a high range, that calculations of morphological and transport-related parameters would require huge computational resources: If we consider the datasets used above, we find that the pore space in the catalyst layer is discretized into elements with overall roughly 350 mio nodes. The MPL has over 500 mio nodes, and the pore space of the not as well resolved GDL is formed by ca. 30 mio nodes. If we take resolutions and volume fractions into account, we would end up with roughly 900 × 1012 nodes in a combined dataset that does not make use of homogenization. If one would like to determine the diffusivity of such a combined dataset, one would have to solve Fick's law (we assume one field variable, the concentration, and one degree-of-freedom per node) and thus a system with overall 900 × 1012 degrees-of-freedom.

However, at this stage, one would have a large-volume fuel cell reconstruction with the following features included and discriminated from each other: Fibers and the PTFE coating of the gas diffusion layer, global shape of the MPL and the catalyst layer (including any cracks), and the microporous carbonaceous matrix within these layers (including primary and secondary pores). As a next step, the ionomer and PTFE binder within the catalyst layer MPL (respectively) must be inscribed. This could be done stochastically upon information obtained from 3D TEM, as exemplarily shown by Lopez-Haro et al. in case of ionomer distribution [4]. The ionomer and PTFE phase could be inscribed to match the film thickness, volume fraction, and coverage determined in 3D TEM tomography. In a last step, Pt particles must be inscribed into the catalyst layer. This can as well be performed stochastically to match the Pt particle volume fraction and size distribution obtained from TEM tomography, as already shown by Thiele et al. [5].

The result of this procedure would be a fully resolved, hybrid fuel cell reconstruction, based on virtual design and actual tomography data. All components of the fuel cell electrode would be morphologically resolved, enabling the calculation of gas, water, protons, and electrons to and from the catalytically active sites. This, together with a given reaction rate at each Pt particle, would for the first time allow dynamic fuel cell performance modeling, purely based on morphological reconstruction. In addition, morphological changes within the electrode could be simulated. For instance, the effect of ionomer swelling within the catalyst layer, which is a current challenge for hydrocarbon ionomers [32], could be simulated by varying the inscribed film thickness. Also, the effect of ionomers with different ion exchange capacities on fuel cell performance could directly be modeled by assigning a different proton conductivity to the ionomer phase. It must be noted that generating high-quality reconstructions itself, regardless if it is TEM, FIB–SEM, or X-ray is a challenge, and that the calculation of transport parameters within fuel cell reconstructions is not trivial. Reconstructions might contain artifacts, especially in invasive methods such as FIB–SEM tomography. Further, the segmentation of a tomography of porous media is challenging and needs to be done properly [33]. Last, as reconstructions are always a discretization of real morphology, the calculated transport parameters are dependent on the resolution of the reconstruction, which needs to be compensated for in material analysis [14,34]. In this regard, it should also be noted that innovative strategies for implementing contact resistances, especially those of carbon fibers and spheres, should be developed. Here, perfect fibers and spheres could be virtually inscribed into the dataset, with the boundary condition of filling as much of the carbonaceous solid volume fraction as possible. From the contact points of the inscribed fibers and spheres, areas of increased contact resistance could be derived. However, despite probable shortcomings, such approaches toward a morphologically fully resolved reconstruction could open a new era of tomography-based analysis in fuel cells, especially when combined with experimental validation.

Tomography-based analysis of fuel cells, especially when combined with experimental validation, is a powerful tool to understand and control fuel cell morphology, and the transport processes within. Elaborate imaging techniques, such as 3D TEM, FIB–SEM tomography, or X-ray tomography, are able to cover various morphological features of the fuel cell, ranging from the Pt particle distribution on the nanoscale up to reconstructions of the fibrous GDL. However, it remains challenging to properly combine the information obtained from the different imaging techniques into one single, large-scale, comprehensive, and fully resolved fuel cell representation. Approaches by virtual design and parametrically driven stochastics could overcome this hurdle. Soon such approaches could provide for tomography-based but virtually generated hybrid reconstructions, which allow purely morphology-dependent prediction of overall fuel cell performance.

  • Bundesministerium fur Bildung und Forschung (Grant No. 03SF0544D).

Gröger, O. , Gasteiger, H. A. , and Suchsland, J.-P. , 2015, “ Review—Electromobility: Batteries or Fuel Cells?,” J. Electrochem. Soc., 162(14), pp. A2605–A2622. [CrossRef]
Litster, S. , and McLean, G. , 2004, “ PEM Fuel Cell Electrodes,” J. Power Sources, 130(1–2), pp. 61–76. [CrossRef]
Wilson, M. S. , Valerio, J. A. , and Gottesfeld, S. , 1995, “ Low Platinum Loading Electrodes for Polymer Electrolyte Fuel Cells Fabricated Using Thermoplastic Ionomers,” Electrochim. Acta, 40(3), pp. 355–363. [CrossRef]
Lopez-Haro, M. , Guétaz, L. , Printemps, T. , Morin, A. , Escribano, S. , Jouneau, P.-H. , Bayle-Guillemaud, P. , Chandezon, F. , and Gebel, G. , 2014, “ Three-Dimensional Analysis of Nafion Layers in Fuel Cell Electrodes,” Nat. Commun., 5, p. 5229. [CrossRef] [PubMed]
Thiele, S. , Fürstenhaupt, T. , Banham, D. , Hutzenlaub, T. , Birss, V. , Ziegler, C. , and Zengerle, R. , 2013, “ Multiscale Tomography of Nanoporous Carbon-Supported Noble Metal Catalyst Layers,” J. Power Sources, 228, pp. 185–192. [CrossRef]
Cindrella, L. , Kannan, A. M. , Lin, J. F. , Saminathan, K. , Ho, Y. , Lin, C. W. , and Wertz, J. , 2009, “ Gas Diffusion Layer for Proton Exchange Membrane Fuel Cells—A Review,” J. Power Sources, 194(1), pp. 146–160. [CrossRef]
Wargo, E. A. , Schulz, V. P. , Çeçen, A. , Kalidindi, S. R. , and Kumbur, E. C. , 2013, “ Resolving Macro- and Micro-Porous Layer Interaction in Polymer Electrolyte Fuel Cells Using Focused Ion Beam and X-Ray Computed Tomography,” Electrochim. Acta, 87, pp. 201–212. [CrossRef]
Mayrhofer, K. J. , Meier, J. C. , Ashton, S. J. , Wiberg, G. K. , Kraus, F. , Hanzlik, M. , and Arenz, M. , 2008, “ Fuel Cell Catalyst Degradation on the Nanoscale,” Electrochem. Commun., 10(8), pp. 1144–1147. [CrossRef]
Gontard, L. C. , Dunin-Borkowski, R. E. , and Ozkaya, D. , 2008, “ Three-Dimensional Shapes and Spatial Distributions of Pt and PtCr Catalyst Nanoparticles on Carbon Black,” J. Microsc., 232(2), pp. 248–259. [CrossRef] [PubMed]
Schulenburg, H. , Schwanitz, B. , Linse, N. , Scherer, G. G. , Wokaun, A. , Krbanjevic, J. , Grothausmann, R. , and Manke, I. , 2011, “ 3D Imaging of Catalyst Support Corrosion in Polymer Electrolyte Fuel Cells,” J. Phys. Chem. C, 115(29), pp. 14236–14243. [CrossRef]
Dubau, L. , Castanheira, L. , Berthom, G. , and Maillard, F. , 2013, “ An Identical-Location Transmission Electron Microscopy Study on the Degradation of Pt/C Nanoparticles Under Oxidizing, Reducing and Neutral Atmosphere,” Electrochim. Acta, 110, pp. 273–281. [CrossRef]
Perez-Alonso, F. J. , Elkjær, C. F. , Shim, S. S. , Abrams, B. L. , Stephens, I. E. , and Chorkendorff, I. , 2011, “ Identical Locations Transmission Electron Microscopy Study of Pt/C Electrocatalyst Degradation During Oxygen Reduction Reaction,” J. Power Sources, 196(15), pp. 6085–6091. [CrossRef]
Ziegler, C. , Thiele, S. , and Zengerle, R. , 2011, “ Direct Three-Dimensional Reconstruction of a Nanoporous Catalyst Layer for a Polymer Electrolyte Fuel Cell,” J. Power Sources, 196(4), pp. 2094–2097. [CrossRef]
Hutzenlaub, T. , Becker, J. , Zengerle, R. , and Thiele, S. , 2013, “ How Coarsening the 3D Reconstruction of a Porous Material Influences Diffusivity and Conductivity Values,” ECS Electrochem. Lett., 2(2), pp. F14–F17. [CrossRef]
Wargo, E. A. , Kotaka, T. , Tabuchi, Y. , and Kumbur, E. C. , 2013, “ Comparison of Focused Ion Beam Versus Nano-Scale X-Ray Computed Tomography for Resolving 3-D Microstructures of Porous Fuel Cell Materials,” J. Power Sources, 241, pp. 608–618. [CrossRef]
Ostadi, H. , Rama, P. , Liu, Y. , Chen, R. , Zhang, X. X. , and Jiang, K. , 2010, “ 3D Reconstruction of a Gas Diffusion Layer and a Microporous Layer,” J. Membr. Sci., 351(1–2), pp. 69–74. [CrossRef]
Müllner, T. , Zankel, A. , Mayrhofer, C. , Reingruber, H. , Höltzel, A. , Lv, Y. , Svec, F. , and Tallarek, U. , 2012, “ Reconstruction and Characterization of a Polymer-Based Monolithic Stationary Phase Using Serial Block-Face Scanning Electron Microscopy,” Langmuir, 28(49), pp. 16733–16737. [CrossRef] [PubMed]
Micheva, K. D. , and Smith, S. J. , 2007, “ Array Tomography: A New Tool for Imaging the Molecular Architecture and Ultrastructure of Neural Circuits,” Neuron, 55(1), pp. 25–36. [CrossRef] [PubMed]
Caplan, J. , Niethammer, M. , Taylor, R. M. , and Czymmek, K. J. , 2011, “ The Power of Correlative Microscopy: Multi-Modal, Multi-Scale, Multi-Dimensional,” Curr. Opin. Struct. Biol., 21(5), pp. 686–693. [CrossRef] [PubMed]
Wacker, I. , and Schroeder, R. R. , 2013, “ Array Tomography,” J. Microsc., 252(2), pp. 93–99. [CrossRef] [PubMed]
Alekseev, A. , Efimov, A. , Loos, J. , Matsko, N. , and Syurik, J. , 2014, “ Three-Dimensional Imaging of Polymer Materials by Scanning Probe Tomography,” Eur. Polym. J., 52, pp. 154–165. [CrossRef]
Aleksandrova, E. , Hiesgen, R. , Friedrich, K. A. , and Roduner, E. , 2007, “ Electrochemical Atomic Force Microscopy Study of Proton Conductivity in a Nafion Membrane,” Phys. Chem. Chem. Phys., 9(21), pp. 2735–2743. [CrossRef] [PubMed]
Bunn, J. R. , Penumadu, D. , Woracek, R. , Kardjilov, N. , Hilger, A. , Manke, I. , and Williams, S. , 2013, “ Detection of Water With High Sensitivity to Study Polymer Electrolyte Fuel Cell Membranes Using Cold Neutrons at High Spatial Resolution,” Appl. Phys. Lett., 102(23), p. 234102. [CrossRef]
Owejan, J. P. , Trabold, T. A. , and Mench, M. M. , 2014, “ Oxygen Transport Resistance Correlated to Liquid Water Saturation in the Gas Diffusion Layer of PEM Fuel Cells,” Int. J. Heat Mass Transfer, 71, pp. 585–592. [CrossRef]
Göbel, M. , Godehardt, M. , and Schladitz, K. , 2017, “ Multi-Scale Structural Analysis of Gas Diffusion Layers,” J. Power Sources, 355, pp. 8–17. [CrossRef]
Terao, T. , Inoue, G. , Kawase, M. , Kubo, N. , Yamaguchi, M. , Yokoyama, K. , Tokunaga, T. , Shinohara, K. , Hara, Y. , and Hara, T. , 2017, “ Development of Novel Three-Dimensional Reconstruction Method for Porous Media for Polymer Electrolyte Fuel Cells Using Focused Ion Beam-Scanning Electron Microscope Tomography,” J. Power Sources, 347, pp. 108–113. [CrossRef]
Cetinbas, F. C. , Ahluwalia, R. K. , Kariuki, N. , de Andrade, V. , Fongalland, D. , Smith, L. , Sharman, J. , Ferreira, P. , Rasouli, S. , and Myers, D. J. , 2017, “ Hybrid Approach Combining Multiple Characterization Techniques and Simulations for Microstructural Analysis of Proton Exchange Membrane Fuel Cell Electrodes,” J. Power Sources, 344, pp. 62–73. [CrossRef]
Lange, K. J. , Sui, P.-C. , and Djilali, N. , 2011, “ Pore Scale Modeling of a Proton Exchange Membrane Fuel Cell Catalyst Layer: Effects of Water Vapor and Temperature,” J. Power Sources, 196(6), pp. 3195–3203. [CrossRef]
Lange, K. J. , Sui, P.-C. , and Djilali, N. , 2010, “ Pore Scale Simulation of Transport and Electrochemical Reactions in Reconstructed PEMFC Catalyst Layers,” J. Electrochem. Soc., 157(10), pp. B1434–B1442. [CrossRef]
Siddique, N. A. , and Liu, F. , 2010, “ Process Based Reconstruction and Simulation of a Three-Dimensional Fuel Cell Catalyst Layer,” Electrochim. Acta, 55(19), pp. 5357–5366. [CrossRef]
Mukherjee, P. P. , and Wang, C. Y. , 2006, “ Stochastic Microstructure Reconstruction and Direct Numerical Simulation of the PEFC Catalyst Layer,” J. Electrochem. Soc., 153(5), pp. A840–A849. [CrossRef]
Holdcroft, S. , 2014, “ Fuel Cell Catalyst Layers: A Polymer Science Perspective,” Chem. Mater., 26(1), pp. 381–393. [CrossRef]
Vierrath, S. , Güder, F. , Menzel, A. , Hagner, M. , Zengerle, R. , Zacharias, M. , and Thiele, S. , 2015, “ Enhancing the Quality of the Tomography of Nanoporous Materials for Better Understanding of Polymer Electrolyte Fuel Cell Materials,” J. Power Sources, 285, pp. 413–417. [CrossRef]
Klingele, M. , Zengerle, R. , and Thiele, S. , 2015, “ Quantification of Artifacts in Scanning Electron Microscopy Tomography: Improving the Reliability of Calculated Transport Parameters in Energy Applications Such as Fuel Cell and Battery Electrodes,” J. Power Sources, 275, pp. 852–859. [CrossRef]
Copyright © 2018 by ASME
View article in PDF format.

References

Gröger, O. , Gasteiger, H. A. , and Suchsland, J.-P. , 2015, “ Review—Electromobility: Batteries or Fuel Cells?,” J. Electrochem. Soc., 162(14), pp. A2605–A2622. [CrossRef]
Litster, S. , and McLean, G. , 2004, “ PEM Fuel Cell Electrodes,” J. Power Sources, 130(1–2), pp. 61–76. [CrossRef]
Wilson, M. S. , Valerio, J. A. , and Gottesfeld, S. , 1995, “ Low Platinum Loading Electrodes for Polymer Electrolyte Fuel Cells Fabricated Using Thermoplastic Ionomers,” Electrochim. Acta, 40(3), pp. 355–363. [CrossRef]
Lopez-Haro, M. , Guétaz, L. , Printemps, T. , Morin, A. , Escribano, S. , Jouneau, P.-H. , Bayle-Guillemaud, P. , Chandezon, F. , and Gebel, G. , 2014, “ Three-Dimensional Analysis of Nafion Layers in Fuel Cell Electrodes,” Nat. Commun., 5, p. 5229. [CrossRef] [PubMed]
Thiele, S. , Fürstenhaupt, T. , Banham, D. , Hutzenlaub, T. , Birss, V. , Ziegler, C. , and Zengerle, R. , 2013, “ Multiscale Tomography of Nanoporous Carbon-Supported Noble Metal Catalyst Layers,” J. Power Sources, 228, pp. 185–192. [CrossRef]
Cindrella, L. , Kannan, A. M. , Lin, J. F. , Saminathan, K. , Ho, Y. , Lin, C. W. , and Wertz, J. , 2009, “ Gas Diffusion Layer for Proton Exchange Membrane Fuel Cells—A Review,” J. Power Sources, 194(1), pp. 146–160. [CrossRef]
Wargo, E. A. , Schulz, V. P. , Çeçen, A. , Kalidindi, S. R. , and Kumbur, E. C. , 2013, “ Resolving Macro- and Micro-Porous Layer Interaction in Polymer Electrolyte Fuel Cells Using Focused Ion Beam and X-Ray Computed Tomography,” Electrochim. Acta, 87, pp. 201–212. [CrossRef]
Mayrhofer, K. J. , Meier, J. C. , Ashton, S. J. , Wiberg, G. K. , Kraus, F. , Hanzlik, M. , and Arenz, M. , 2008, “ Fuel Cell Catalyst Degradation on the Nanoscale,” Electrochem. Commun., 10(8), pp. 1144–1147. [CrossRef]
Gontard, L. C. , Dunin-Borkowski, R. E. , and Ozkaya, D. , 2008, “ Three-Dimensional Shapes and Spatial Distributions of Pt and PtCr Catalyst Nanoparticles on Carbon Black,” J. Microsc., 232(2), pp. 248–259. [CrossRef] [PubMed]
Schulenburg, H. , Schwanitz, B. , Linse, N. , Scherer, G. G. , Wokaun, A. , Krbanjevic, J. , Grothausmann, R. , and Manke, I. , 2011, “ 3D Imaging of Catalyst Support Corrosion in Polymer Electrolyte Fuel Cells,” J. Phys. Chem. C, 115(29), pp. 14236–14243. [CrossRef]
Dubau, L. , Castanheira, L. , Berthom, G. , and Maillard, F. , 2013, “ An Identical-Location Transmission Electron Microscopy Study on the Degradation of Pt/C Nanoparticles Under Oxidizing, Reducing and Neutral Atmosphere,” Electrochim. Acta, 110, pp. 273–281. [CrossRef]
Perez-Alonso, F. J. , Elkjær, C. F. , Shim, S. S. , Abrams, B. L. , Stephens, I. E. , and Chorkendorff, I. , 2011, “ Identical Locations Transmission Electron Microscopy Study of Pt/C Electrocatalyst Degradation During Oxygen Reduction Reaction,” J. Power Sources, 196(15), pp. 6085–6091. [CrossRef]
Ziegler, C. , Thiele, S. , and Zengerle, R. , 2011, “ Direct Three-Dimensional Reconstruction of a Nanoporous Catalyst Layer for a Polymer Electrolyte Fuel Cell,” J. Power Sources, 196(4), pp. 2094–2097. [CrossRef]
Hutzenlaub, T. , Becker, J. , Zengerle, R. , and Thiele, S. , 2013, “ How Coarsening the 3D Reconstruction of a Porous Material Influences Diffusivity and Conductivity Values,” ECS Electrochem. Lett., 2(2), pp. F14–F17. [CrossRef]
Wargo, E. A. , Kotaka, T. , Tabuchi, Y. , and Kumbur, E. C. , 2013, “ Comparison of Focused Ion Beam Versus Nano-Scale X-Ray Computed Tomography for Resolving 3-D Microstructures of Porous Fuel Cell Materials,” J. Power Sources, 241, pp. 608–618. [CrossRef]
Ostadi, H. , Rama, P. , Liu, Y. , Chen, R. , Zhang, X. X. , and Jiang, K. , 2010, “ 3D Reconstruction of a Gas Diffusion Layer and a Microporous Layer,” J. Membr. Sci., 351(1–2), pp. 69–74. [CrossRef]
Müllner, T. , Zankel, A. , Mayrhofer, C. , Reingruber, H. , Höltzel, A. , Lv, Y. , Svec, F. , and Tallarek, U. , 2012, “ Reconstruction and Characterization of a Polymer-Based Monolithic Stationary Phase Using Serial Block-Face Scanning Electron Microscopy,” Langmuir, 28(49), pp. 16733–16737. [CrossRef] [PubMed]
Micheva, K. D. , and Smith, S. J. , 2007, “ Array Tomography: A New Tool for Imaging the Molecular Architecture and Ultrastructure of Neural Circuits,” Neuron, 55(1), pp. 25–36. [CrossRef] [PubMed]
Caplan, J. , Niethammer, M. , Taylor, R. M. , and Czymmek, K. J. , 2011, “ The Power of Correlative Microscopy: Multi-Modal, Multi-Scale, Multi-Dimensional,” Curr. Opin. Struct. Biol., 21(5), pp. 686–693. [CrossRef] [PubMed]
Wacker, I. , and Schroeder, R. R. , 2013, “ Array Tomography,” J. Microsc., 252(2), pp. 93–99. [CrossRef] [PubMed]
Alekseev, A. , Efimov, A. , Loos, J. , Matsko, N. , and Syurik, J. , 2014, “ Three-Dimensional Imaging of Polymer Materials by Scanning Probe Tomography,” Eur. Polym. J., 52, pp. 154–165. [CrossRef]
Aleksandrova, E. , Hiesgen, R. , Friedrich, K. A. , and Roduner, E. , 2007, “ Electrochemical Atomic Force Microscopy Study of Proton Conductivity in a Nafion Membrane,” Phys. Chem. Chem. Phys., 9(21), pp. 2735–2743. [CrossRef] [PubMed]
Bunn, J. R. , Penumadu, D. , Woracek, R. , Kardjilov, N. , Hilger, A. , Manke, I. , and Williams, S. , 2013, “ Detection of Water With High Sensitivity to Study Polymer Electrolyte Fuel Cell Membranes Using Cold Neutrons at High Spatial Resolution,” Appl. Phys. Lett., 102(23), p. 234102. [CrossRef]
Owejan, J. P. , Trabold, T. A. , and Mench, M. M. , 2014, “ Oxygen Transport Resistance Correlated to Liquid Water Saturation in the Gas Diffusion Layer of PEM Fuel Cells,” Int. J. Heat Mass Transfer, 71, pp. 585–592. [CrossRef]
Göbel, M. , Godehardt, M. , and Schladitz, K. , 2017, “ Multi-Scale Structural Analysis of Gas Diffusion Layers,” J. Power Sources, 355, pp. 8–17. [CrossRef]
Terao, T. , Inoue, G. , Kawase, M. , Kubo, N. , Yamaguchi, M. , Yokoyama, K. , Tokunaga, T. , Shinohara, K. , Hara, Y. , and Hara, T. , 2017, “ Development of Novel Three-Dimensional Reconstruction Method for Porous Media for Polymer Electrolyte Fuel Cells Using Focused Ion Beam-Scanning Electron Microscope Tomography,” J. Power Sources, 347, pp. 108–113. [CrossRef]
Cetinbas, F. C. , Ahluwalia, R. K. , Kariuki, N. , de Andrade, V. , Fongalland, D. , Smith, L. , Sharman, J. , Ferreira, P. , Rasouli, S. , and Myers, D. J. , 2017, “ Hybrid Approach Combining Multiple Characterization Techniques and Simulations for Microstructural Analysis of Proton Exchange Membrane Fuel Cell Electrodes,” J. Power Sources, 344, pp. 62–73. [CrossRef]
Lange, K. J. , Sui, P.-C. , and Djilali, N. , 2011, “ Pore Scale Modeling of a Proton Exchange Membrane Fuel Cell Catalyst Layer: Effects of Water Vapor and Temperature,” J. Power Sources, 196(6), pp. 3195–3203. [CrossRef]
Lange, K. J. , Sui, P.-C. , and Djilali, N. , 2010, “ Pore Scale Simulation of Transport and Electrochemical Reactions in Reconstructed PEMFC Catalyst Layers,” J. Electrochem. Soc., 157(10), pp. B1434–B1442. [CrossRef]
Siddique, N. A. , and Liu, F. , 2010, “ Process Based Reconstruction and Simulation of a Three-Dimensional Fuel Cell Catalyst Layer,” Electrochim. Acta, 55(19), pp. 5357–5366. [CrossRef]
Mukherjee, P. P. , and Wang, C. Y. , 2006, “ Stochastic Microstructure Reconstruction and Direct Numerical Simulation of the PEFC Catalyst Layer,” J. Electrochem. Soc., 153(5), pp. A840–A849. [CrossRef]
Holdcroft, S. , 2014, “ Fuel Cell Catalyst Layers: A Polymer Science Perspective,” Chem. Mater., 26(1), pp. 381–393. [CrossRef]
Vierrath, S. , Güder, F. , Menzel, A. , Hagner, M. , Zengerle, R. , Zacharias, M. , and Thiele, S. , 2015, “ Enhancing the Quality of the Tomography of Nanoporous Materials for Better Understanding of Polymer Electrolyte Fuel Cell Materials,” J. Power Sources, 285, pp. 413–417. [CrossRef]
Klingele, M. , Zengerle, R. , and Thiele, S. , 2015, “ Quantification of Artifacts in Scanning Electron Microscopy Tomography: Improving the Reliability of Calculated Transport Parameters in Energy Applications Such as Fuel Cell and Battery Electrodes,” J. Power Sources, 275, pp. 852–859. [CrossRef]

Figures

Grahic Jump Location
Fig. 4

Two methods for adding the microporous carbon matrix to the microporous layer identified by X-ray. In (a), a microporous FIB–SEM reconstruction of the microporous layer is directly inscribed by stringing together the reconstructed volume. In (b), the microporosity is added to the microporous layer by stochastically placing spheres, so that the whole layer matches the porosity and tortuosity as well as the pore- and grain-size distribution of the FIB–SEM reconstruction.

Grahic Jump Location
Fig. 3

Integral and spatially resolved multiscale particle- and pore-size analysis of the global gas diffusion electrode reconstruction. The region starting at approximately 150 μm in depth can be assigned to the microporous layer and catalyst layer. Information within this region is obtained by FIB–SEM tomography, whereas all other information is obtained from X-ray tomography. By this, morphological features ranging from several nanometers up to micrometers can be covered within one dataset (MGS: mean grain size, GVF: grain (or particle) volume fraction, MPS: mean pore size, and PVF: pore volume fraction).

Grahic Jump Location
Fig. 2

(b) Global X-ray dataset of a gas diffusion electrode. Fibers, pore space within the fibers, microporous layer (MPL), and catalyst layer (CL) can be discriminated. The microporosity of the microporous layer and the catalyst layer cannot be resolved by X-ray, but by FIB–SEM tomography, as depicted in (c) and (a), respectively. Calculated parameters from the FIB–SEM reconstruction can be inscribed into the respective areas of the X-ray dataset. From this, multiscale tomography-based analysis can be performed for the whole gas diffusion electrode.

Grahic Jump Location
Fig. 1

Cartoon (top) and real imaged (bottom) morphology of a fuel cell. It is shown that morphological features range from dimensions of several micrometers down to few nanometers. Different methods are thus needed to image these features in varying resolutions and field of views. In this figure, this is X-ray tomography (a) and (b), FIB–SEM tomography (c), and TEM tomography (d).

Tables

Errata

Discussions

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In