Current Status and Future Development of Catalyst Materials and

---

Current Status and Future Development of Catalyst Materials and Catalyst Layers for Proton Exchange Membrane Fuel Cells: An Industrial Perspective Dustin Banham and Siyu Ye* Ballard Power Systems, 9000 Glenlyon Parkway, Burnaby, BC V5J 5J8, Canada ABSTRACT: Proton exchange membrane fuel cells (PEMFCs) have already penetrated many commercial markets (e.g., portable power, backup power, materials handling, and buses) and are poised to greatly expand in the automotive market with both Toyota and Hyundai recently commercializing small fleets. As this occurs, catalysts for PEMFCs will experience ever greater demands on cost, activity, and durability. This Perspective outlines the technology timeline and characteristics of the most promising catalysts currently being developed and discusses the remaining challenges for both platinum group metal and nonprecious metal catalysts. Finally, the importance of combined catalyst and catalyst layer design strategies is highlighted, and a brief discussion on the future outlook of this field is provided.

I

increased sales allow for decreased cost through economies of scale. This is true for almost every component of a PEMFC, including the bipolar plates, gas diffusion layers, ionomer, and membrane. Unfortunately, there is still one component (platinum group metal catalysts) that could get more expensive as sales/volumes increase. In fact, it is anticipated that the contribution to total stack cost from PGMs (plus application) will increase from 21% at 1 000 fuel cell systems/year to 45% at 500 000 systems/year.3,4 At both the cathode and anode of a PEMFC, platinum group metals (PGMs) are currently required to catalyze the desired redox reactions (hydrogen oxidation at the anode, and oxygen reduction at the cathode). As these metals are commodities and are all quite scarce, increased demand for PEMFCs will only serve to increase the price of these catalysts if the loading is not reduced significantly from current levels. Because of the sluggish kinetics of the oxygen reduction reaction (ORR) (∼5 orders of magnitude slower than hydrogen oxidation kinetics5), the majority of the PGMs are required at the cathode. This challenge is widely recognized in the PEMFC community and has led to a strong focus on improving the catalysts used for the ORR at the cathode. This includes both improving the activity and utilization as well as the durability and stability of these catalysts. A major commercial market for PEMFCs will

n the face of ever-growing environmental concerns, the need for advanced clean energy technology has never been more apparent. The global commitment to reduce our dependency on combustion of fossil fuels was recently highlighted on December 12, 2015, in the Paris Agreement, which is the world’s first comprehensive climate agreement. As it is recognized that climate change is a global problem that will require a global solution, 195 countries agreed to sign this broadreaching agreement. To meet the coming energy challenges, a broad suite of alternative energy and renewable resources will be required, including biomass, solar, hydro, wind, and nuclear to name a few. As many of these alternative energy sources are intermittent, it is widely recognized that energy storage will become increasingly important to help stabilize demand on the grid. Hydrogen has been identified as one of the most promising means for storing energy and will certainly play a large role in helping to meet the coming energy challenges.1 Because of their ability to efficiently convert the stored chemical energy in hydrogen into useable electrical energy, proton exchange membrane fuel cells (PEMFCs) will be a central technology in any envisioned version of the anticipated “hydrogen economy”. In fact, PEMFCs are already being commercialized for a broad range of applications, including portable power, backup power, materials handling, and buses. Additionally, some automotive original equipment manufacturers (OEMs) have commenced sales of fuel cell electric vehicles (FCEVs).2 The penetration of PEMFCs into these commercial markets is expected to have exponential growth, as © 2017 American Chemical Society

Received: November 30, 2016 Accepted: February 3, 2017 Published: February 3, 2017 629

DOI: 10.1021/acsenergylett.6b00644 ACS Energy Lett. 2017, 2, 629−638

Perspective

http://pubs.acs.org/journal/aelccp

ACS Energy Letters

Perspective

for catalyst layer researchers, Pt utilization can mean access of reactants to the Pt surface at a given current density. Clearly, all three interpretations are important, but it is necessary to clarify which one is being discussed as they have very different meanings. In this Perspective, it will be explicitly stated which type of “Pt utilization” is meant whenever this term is used. Finally, the terms electrochemical surface area (ECSA) and electrode platinum surface area (EPSA)8 must be defined. ECSA is typically reported in units of square meters per gram and refers to the catalyst surface area normalized to the mass of the catalyst particle. This is the measurement of interest to most synthetic chemists and catalyst researchers as it relates to definition 1 for Pt utilization. EPSA is a measure of real platinum surface area to geometric surface area of an MEA, and has units of square centimeters of platinum per square centimeter of MEA (cm2 Pt/cm2 MEA). Because of the close analogy to traditional “roughness factor” in electrochemistry and materials science,9 EPSA is often also referred to as the MEA roughness factor.6,10,11 EPSA is a critical value for catalyst layer researchers and is related to Definition 2 for “Pt utilization”. Classif ication and Current Status of ORR Catalysts. During the past decade, a wide variety of highly promising ORR catalysts have been developed. Broadly, these catalysts can be categorized as (1) Pt/C, (2) Pt and Pt alloy/dealloy, (3) core− shell, (4) nonprecious metal catalysts, (5) shape-controlled nanocrystals, and (6) nanoframes. The approximate development timelines for each catalyst type is shown in Figure 1. The timeline in Figure 1 is highly subjective and is meant to give only a general ranking of technology readiness for each catalyst family. No specific dates are provided as this would be rather presumptuous and will depend largely on where current and future research efforts are focused. It should be noted that these catalysts are usually supported on carbon and other nanomaterials. As this Perspective is not focused on support material, readers are referred to other review papers published on this topic.12,13 Additionally, the nanostructured thin film (NSTF) catalyst developed by 3 M can be considered a hybrid of several of the catalysts listed above and is also highly promising.14,15 However, as this catalyst is a free-standing structure, and typically “ionomer-free” when used as a catalyst layer, it is fundamentally different from those listed in Figure 1, with many unique advantages and challenges. Thus, this particular catalyst will not be covered in this Perspective, and for further understanding of the NSTF, readers are referred to an overview published by 3M.15 As mentioned, the “current status” and development timeline shown in Figure 1 is subjective and, among other factors, depends largely on what applications are being considered. For example, from a purely automotive perspective, the nonprecious metal catalysts (NPMCs) would likely be considered the furthest from commercialization, as they currently do not meet performance, durability (performance loss during voltage cycling20), stability (performance loss during potentiostatic/ galvanostatic experiments20), or power density requirements. However, NPMCs are now close to meeting the requirements for portable power applications. In addition to differences in the development timeline of each family of catalyst, each one offers unique advantages and disadvantages from an industrial or commercialization perspective. At a high level, these differences are summarized in Figure 2 and will be discussed in more depth in the following section.

be automotive; thus, it is this market that drives much of the long-term cost analysis and catalyst requirements. An excellent overview on catalyst requirements for the automotive industry was recently published by General Motors (GM).6 However, in the short- to midterm, other markets which are currently more developed such as portable power, backup power, materials handling, and buses will represent a larger total market share of PEMFC products. Each of these markets has different product (and catalyst) requirements. Thus, while it is prudent to strive toward commercial automotive targets, it is also wise for catalyst developers to keep the requirements of these other markets in mind, as they may offer more immediate opportunities to commercialize some of the next-generation catalysts currently being developed.

While it is prudent to strive toward commercial automotive targets, it is also wise for catalyst developers to keep the requirements of other markets in mind, as they may offer more immediate opportunities to commercialize some of the next-generation catalysts currently being developed. Clarif ication of Terminology. When considering the activity of a PGM catalyst, there are two primary metrics: (1) specific activity (μA/cm2) and (2) mass activity (A/mg). The specific activity of a catalyst provides information on the inherent turnover frequency and is thus an important parameter for researchers working to tune catalyst structure and electronic properties of their catalyst to maximize the exchange current density toward the oxygen reduction reaction. As specific activity is concerned only with the processes occurring at the surface of the catalyst, it is an excellent metric for single-crystal studies where the mass of the electrode is not a factor of interest. In fact, it is through analyzing and comparing specific activities that many of the most significant catalyst advances have been made, including the seminal work on Pt3Ni(111),7 which demonstrated a specific activity 90-fold higher than commercial Pt/C and had a profound impact on global catalysis research efforts. However, for industrial applications, cost is the primary consideration; thus, the mass activity is ultimately what matters as this metric is directly transferable to cost. Importantly, this metric considers both inherent activity (specific activity) and utilization (ratio of surface atoms to bulk). Thus, for this Perspective, when discussing next-generation ORR catalysts, the primary activity metric will be mass activity. Another important concept prior to delving into deeper analysis is “Pt utilization”, which can have very different meanings depending on whether a catalyst or catalyst layer perspective is taken. Definition 1: For catalyst researchers, Pt utilization is based on the concept of “dispersion” and is a measure of the ratio of surface atoms to bulk. In this sense, Pt utilization has no dependence on how well the cathode catalyst layer (CCL) is designed. Definition 2: For catalyst layer researchers, Pt utilization can mean the ratio of electrochemically accessible Pt area in the membrane electrode assembly (MEA) to the expected Pt area [based on transmission electron micrsoscopy (TEM), X-ray diffraction (XRD), rotating disk electrode (RDE), or other ex-situ measurements]. Definition 3: Also 630

DOI: 10.1021/acsenergylett.6b00644 ACS Energy Lett. 2017, 2, 629−638

ACS Energy Letters

Perspective

Figure 1. Development timelines for Pt, Pt alloy/dealloy, core−shell,16 nonprecious metal,17 shape-controlled,18 and nanoframe19 ORR electrocatalysts.

Figure 2. Benefits and remaining challenges for each of the primary categories of electrocatalysts.

Pt/C. For the past decade, commercial PEMFC products have relied heavily on Pt/C catalysts. When first introduced, these catalysts offered significant advantages over unsupported Pt black because of the much smaller nanoparticles that are achievable with supported catalysts. The simplicity of these catalysts is both a benefit and a drawback. From a synthetic perspective, there is little room to tailor activity and durability when limiting the design to a single element (Pt). In fact, further improvements in activity and durability with

conventional Pt/C now rely on advances in catalyst supports resulting in “catalyst−support” interactions12,21 which have been reported to enhance both activity and durability of PGMbased ORR catalysts. While promising, these approaches are unlikely to meet long-term mass activity requirements using conventional Pt nanoparticles. From a manufacturing perspective, simpler systems are advantageous. However, while having only one component in the synthesis (Pt) may be considered an advantage, the reality is that at large scales, catalyst 631

DOI: 10.1021/acsenergylett.6b00644 ACS Energy Lett. 2017, 2, 629−638

ACS Energy Letters

Perspective

promising core−shell catalysts is often Pd, which when included, decreases the ECSA by ∼60%.37 Thus, for practical use, it is imperative that less expensive cores (PGM alloy or non-PGM) are explored. Fortunately, this work is already underway, with highly promising results at the RDE level.38,39 Shape-Controlled Nanocrystals. While at an earlier stage of development than core−shell catalysts, shape-controlled catalysts (Figure 2) appear to be a highly promising class of ORR catalyst because of their extremely high mass activities.40 As described above, mass activity is dependent on specific activity and Pt utilization (Pt dispersion). Core−shell catalysts most strongly exemplify the “Pt utilization” strategy to generate high mass activities. Conversely, shape-controlled catalysts rely primarily on achieving extraordinarily high specific activities to generate high mass activities. In this way, these catalysts have the closest ties to the fundamental single-crystal studies mentioned earlier in this Perspective. In principle, these catalysts attempt to recreate the ideal crystal structure identified by single-crystal studies,7 but at the nanometer scale.40 An excellent example of this is the 9 nm Pt2.5Ni cctahedra developed at the Georgia Institute of Technology. Despite the relatively poor Pt utilization (Pt dispersion) afforded by the large 9 nm particles in this study, a mass activity of 3.3 A/mg was achieved (more than 7-fold higher than the 2020 DOE MEA target, albeit at the RDE level).41 This was accomplished through maintaining the ideal Pt2.5Ni(111) crystal structure (which has >50× higher specific activity versus commercial Pt/C) at the nanometer scale. Despite the great promise shown by these catalysts, it is still too early in the development timeline to make firm conclusions on their commercial viability (the majority of testing on this family of catalyst has been at the RDE level only). In particular, there are two issues which may prove highly challenging for these catalysts when used in an operating fuel cell: (1) ECSA and (2) durability and stability. ECSA: To to achieve high performance at high current densities with a PGM loading of 0.1 mg/cm2, an ECSA of ∼50 m2/g will be required assuming no substantial advances in ionomer (discussed later in this Perspective). Unfortunately, it has proven difficult to maintain the desired crystal face as the particle size is decreased to ≪10 nm, leaving ECSA as a key challenge for this family of catalysts for automotive applications. For nonautomotive applications (e.g., portable power, backup power, materials handling), high current densities are not as critical. For these applications, efficiency is a key metric, and they thus operate at lower current densities. In this case, achieving an ECSA of >50 m2/g is less of a concern, and the much higher mass activity of these advanced catalysts versus commercial Pt (or Pt alloy) catalysts could provide a valuable increase in efficiency for products used in these other important markets. Durability and Stability. It has been reported that shape-controlled catalysts have lower stability than commercial Pt/C when subjected to voltage cycling.40,42 In fact, it has been suggested that this type of catalyst exists only in a “metastable” state and will inevitably change to the thermodynamically preferred “round” shape following voltage cycling.42 The use of core− shell type strategies may help to overcome this limitation, with one report showing almost no loss in activity following an aggressive voltage cycling protocol for a PtPd−Ni core−shell octahedral catalyst.43 Ultimately, MEA testing will be required to properly understand the merits and limitations of this catalyst type. In this regard, some MEA testing performed at GM has already indicated that some of these shape-controlled

manufacturing costs are minimal in comparison to PGM costs.22 Pt-Alloy. For these reasons, Pt-alloy (e.g., PtCo, PtNi) catalysts are becoming the new baseline catalyst at the commercial level as they are able to achieve high mass activities while also demonstrating similar or better durability compared to Pt/C. The technological maturity of these catalysts is highlighted by Toyota’s recent announcement that a PtCo-alloy is currently used in the Mirai.2 The improved electrocatalytic activity of Pt-alloys (such as PtCo, PtNi, PtFe, PtCr, PtV, PtTi, PtW, PtAl, and PtAg) has been attributed to (i) the smaller Pt− Pt bond distances resulting in more favorable sites that enhance the dissociative adsorption of oxygen and (ii) the structuresensitive inhibiting effect of OHads. Considerable work has been carried out over the past decades on carbon-supported binary alloys or ternary alloys that demonstrate 2−3 times higher mass activity vs Pt/C.23,24 To improve the stability and durability of these catalysts, further work on preleaching of Pt-alloy catalysts has been performed to remove the base metal (deposited on the carbon surface or poorly alloyed to the Pt).25 Post-treatments by either acid leaching (starting with Co/Ni rich alloys) and/or heat treatment of Pt alloys have also been performed, resulting in increased stability and activity, due to the formation of a Pt-rich skin.26−28 These approaches have proven quite successful in improving the durability of alloy catalysts, with similar or improved durability versus Pt/C reported at the MEA/stack level.25,29,30 Core−Shell. Significant progress has been made in recent years on the highly promising “core−shell” family of ORR catalysts. As shown in Figure 2, the core−shell concept relies on having the active ORR catalyst (Pt) located only on the surface of the nanoparticle, with another metal (most typically Pd) making up the bulk. This unique design and concept theoretically can allow for the highest possible Pt utilization (surface Pt to bulk) and thus from a cost perspective is highly attractive (provided less expensive cores can be developed). Additionally, the inherent rate of the ORR can be tuned through changing the core, which has both structural and electronic impacts on the Pt shell.31−34 However, perhaps the largest potential benefit of this class of catalyst is the extraordinarily high ECSA afforded by the high Pt dispersion. This benefit goes beyond the obvious cost advantage, as will be discussed in the Remaining Challenges section, could be critical to achieving PGM loadings 50 m2/g19,46) due to their relatively thin frames (