Electrochemical characterization of evolving ionomer/electrocatalyst interactions throughout accelerated stress tests

Leiming Hu, Tim Van Cleve, Haoran Yu, Jae Hyung Park, Nancy Kariuki, A. Jeremy Kropf, Rangachary Mukundan, David A. Cullen, Deborah J. Myers, K. C. Neyerlin

Research output: Contribution to journalArticlepeer-review

13 Scopus citations

Abstract

The degradation of polymer electrolyte membrane fuel cells (PEMFCs) catalyst layers used for heavy-duty vehicles was examined using a catalyst-specific accelerated stress test (AST). High surface area carbon supported dispersed Pt (Pt/ HSC), annealed Pt (a-Pt/ HSC) and PtCo (PtCo/ HSC) alloy catalysts were examined over the course of 90,000 cycles by measuring changes in mass activity, O2 transport resistance, electrochemical active surface area ( ECSA), catalyst accessibility and ionomer-electrocatalyst interactions. Compared to a-Pt/HSC and Pt/HSC catalyst, the PtCo/HSC showed better initial mass activity, a larger initial mass transport loss, and faster degradation after the first 30k AST cycles, as a large portion of Co leached out during potential cycling. Pt/HSC showed higher initial performance relative to a-Pt/HSC but had faster degradation. STEM characterizations show that the ECSA losses are largely related to Pt dissolution resulting in either catalyst particle growth via the Ostwald ripening mechanism or redeposition in the membrane. Catalyst accessibility measurements showed decreased RH sensitivity for all three samples, while CO impedance measurements revealed a decrease in both Pt-water and carbon-water interactions. This implies that, Pt is either preferentially redepositing on the exterior of the carbon support, or that the ionomer is undergoing morphological changes enabling the enhanced intrusion of mesopores.

Original languageEnglish
Article number232490
JournalJournal of Power Sources
Volume556
DOIs
StatePublished - Feb 1 2023

Funding

This work was authored in part by Alliance for Sustainable Energy, LLC, the manager and operator of the National Renewable Energy Laboratory for the U.S. Department of Energy (DOE) under Contract No. DE- AC36-08GO28308 . This material is based on work performed by the Million Mile Fuel Cell Truck (M 2 FCT) Consortium, technology manager Greg Kleen. Funding was provided by the U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy , Hydrogen and Fuel Cell Technologies Office (HFTO). The X-ray scattering experiments were performed at beamline 9-ID-C at the Advanced Photon Source (APS) at Argonne National Laboratory (ANL). The X-ray spectroscopy experiments were performed at beamline 10-ID at the APS, which is operated by the Materials Research Collaborative Access Team (MRCAT). Use of the APS, an Office of Science user facility operated by ANL, is supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences , under Contract No. DE-AS02-06CH11357 . MRCAT is supported by DOE and by the MRCAT member institutions. The Talos F200X S/TEM tool was provided by US DOE, Office of Nuclear Energy, Fuel Cycle R&D Program, and the Nuclear Science User Facilities. RM acknowledges the support of Los Alamos National Laboratory under Contract No, 89233218CNA000001 operated by Triad National Security, LLC. The views expressed in the article do not necessarily represent the views of the DOE or the U.S. Government. The U.S. Government retains and the publisher, by accepting the article for publication, acknowledges that the U.S. Government retains a nonexclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this work, or allow others to do so, for U.S. Government purposes. The degradation of polymer electrolyte membrane fuel cells (PEMFCs) catalyst layers used for heavy-duty vehicles was examined using a catalyst-specific accelerated stress test (AST). High surface area carbon supported dispersed Pt (Pt/HSC), annealed Pt (a-Pt/HSC) and PtCo (PtCo/HSC) alloy catalysts were examined over the course of 90,000 cycles by measuring changes in mass activity, O2 transport resistance, electrochemical active surface area (ECSA), catalyst accessibility and ionomer-electrocatalyst interactions. Compared to a-Pt/HSC and Pt/HSC catalyst, the PtCo/HSC showed better initial mass activity, a larger initial mass transport loss, and faster degradation after the first 30k AST cycles, as a large portion of Co leached out during potential cycling. Pt/HSC showed higher initial performance relative to a-Pt/HSC but had faster degradation. STEM characterizations show that the ECSA losses are largely related to Pt dissolution resulting in either catalyst particle growth via the Ostwald ripening mechanism or redeposition in the membrane. Catalyst accessibility measurements showed decreased RH sensitivity for all three samples, while CO impedance measurements revealed a decrease in both Pt-water and carbon-water interactions. This implies that, Pt is either preferentially redepositing on the exterior of the carbon support, or that the ionomer is undergoing morphological changes enabling the enhanced intrusion of mesopores.The state-of-the-art PEMFC catalyst layer (CL) contains carbon supported Pt or PtCo alloy catalyst bound together by a PFSA (perfluorosulfonic acid) ionomer, which facilitates proton transport [4]. The morphology and distribution of the catalyst nanoparticles, the catalyst/ionomer interactions and the structure of the carbon support are critical to the CL's performance and durability [5,6]. For Pt nanoparticles, the degradation mainly comes in two ways, Pt mass loss and particle coarsening [7,8]. It has been shown that Pt solubility will increase with electrode potential until around 1.1 V, and then decreases at higher potentials due to oxide layer formation [9,10]. The dissolved Pt can migrate to the cathode/membrane interface, where it then gets reduced by hydrogen to form Pt crystals (often seen as a “Pt band”, in the membrane) [11,12]. Alternatively, Pt ions can redeposit to nearby nanoparticles, leading to particle growth (i.e., Ostwald ripening). It has also been reported that nearby Pt nanoparticles can migrate and coalesce into larger particles. In terms of PtCo alloy catalyst, a higher oxygen reduction reaction (ORR) activity can be achieved compared with Pt catalyst due to sub-surface Co element strain effect [13]. During long-term operations, however, Co can also dissolve into acidic polymer electrolyte at electrode potentials higher than 0.1 V [14,15]. Unlike dissolved Pt2+ ion, which can redeposit at low electrode potentials, dissolved Co2+ can hardly be reduced, because the Co/Co2+ has a lower redox potential than H2/H+ in acidic environments [13,16]. During voltage cycling, transition metals will slowly leach out leading to decrease of intrinsic activity of PtCo alloy catalyst [17]. The increased concentration of Co2+ in the ionomer phase can subsequently to increased oxygen transport resistance and decreased high-current-density performance [18]. Papadias et al. [19] investigated PtCo catalyst with different Co content and initial morphology, and found that the voltage losses induced by AST are related to increased kinetic and mass transfer losses, with a combined effect of Co leaching and ECSA loss. Carbon support is another factor that can affect catalyst durability. Compared with low surface area carbon support, high surface area carbon (HSC), and corresponding Pt deposition methods, can provide better initial Pt dispersion. Furthermore, catalyst nanoparticles located inside the carbon support tends to have better initial activity as it is free from purported sulfonate poisoning and more generically ionomer interactions [20,21]. It has been reported that catalyst nanoparticles on HSC have reduced coalescence [7,21,22]. However, HSC supported catalyst can still undergo severe electrochemical active surface area (ECSA) loss due to Pt dissolution and Ostwald ripening [21,22]. O'Brien et al. [23] conducted durability study of PtCo catalyst on different porous and solid carbon supports, and found that porous carbon supports can slow down the catalyst particle dissolution/precipitation, but also exhibit greater transport resistance due to enhanced Co leaching.The aim of this study is to examine the durability of several HSC supported Pt (dispersed - Pt/HSC, annealed - a-Pt/HSC, and alloyed PtCo - PtCo/HSC) catalysts and elucidate the changes in catalyst properties and catalyst-ionomer interactions induced from long-term operation. Membrane electrode assemblies (MEAs) containing these catalysts were exposed to 90k cycle catalyst-specific accelerated stress tests (ASTs) with periodic electrochemical diagnostics applied to examine the changes in catalyst activity, accessibility, ionomer coverage, sulfonate interaction, and electrochemical surface area. Improving the understanding the state of the electrode structure at the end of test informs both material design directions (e.g. catalyst design and support morphology), as well as engineering opportunities (e.g. higher loadings and novel structures) that can subsequently improve efficiency and reduce total cost of ownership for heavy duty vehicles.Membrane Electrode Assembly. – Three types of cathode catalyst with 50 wt% metal loading were used for this study, i.e., HSC supported dispersed Pt catalyst (Pt/HSC, TEC10E50E, TANAKA precious metals, Japan), annealed Pt catalyst (a-Pt/HSC, ElystPt500550, Umicore, Germany), and PtCo catalyst (PtCo/HSC, ElystPt500690, Umicore, Germany). Catalyst coated membranes (CCMs) were fabricated with an anode loading of 0.05 mgPt/cm2 (Pt/HSC, ElystPt200390, Umicore, Germany) and a cathode loading of 0.25 mgPt/cm2 with an ionomer (Nafion D2020 dispersion, Fuel Cell Store) to carbon ratio (I:C ratio) of 0.83. For Umicore a-Pt/HSC and PtCo/HSC catalyst, Nafion HP membrane was used. For Pt/HSC catalyst, Nafion 211 membrane was used. The membrane electrode assembly was assembled by sandwiching the CCM between two 50 cm2 Freudenberg H23C8 (Fuel Cell Store, USA) gas diffusion layers (GDLs). The gasket is polytetrafluoroethylene (PTFE) film with a chosen thickness to achieve a GDL compression ratio of ca. 18%.H2/O2 polarization curves for Pt/HSC, a-Pt/HSC and PtCo/HSC MEAs are shown in Fig. 1a. The beginning of test (BOT) performance of the PtCo/HSC sample has better oxygen reduction reaction (ORR) kinetics compared with the a-Pt/HSC sample, with about 30 mV higher cell voltage at 100 mA/cm2. The mass activity based on iR-free current density at 0.9 V is shown in Fig. 1b. The PtCo/HSC has about two times higher BOT mass activity compared with a-Pt/HSC sample. After the first 30k AST cycles, the PtCo/HSC sample shows a significant cell voltage loss, and the mass activity dropped by 71.3% from 877 mA/mgPt to 252 mA/mgPt. It is known that the high mass activity of the PtCo alloy catalyst largely comes from the geometric strain created by Co element [36]. The rapid decrease during the first 30k AST cycles suggests a large portion of Co leached out from the alloy, which is supported by the WAXS and STEM-EDS characterizations and will be discussed in a later section. From 30k to 90k AST cycles, PtCo/HSC and a-Pt/HSC samples show similar degradation rates, as the mass activity values continued to decrease by 12.3% and 19.6%, respectively. The Pt/HSC catalyst also shows higher initial mass activity compared with a-Pt/HSC but with a faster degradation rate. After 90k AST cycles, the mass activity of Pt/HSC dropped by 83% with an end of test (EOT) mass activity lower than a-Pt/HSC.CV curves were measured for Pt/HSC, a-Pt/HSC and PtCo/HSC MEAs at different stages of the durability test. As shown in Fig. 3, the PtCo/HSC sample BOT CV curve shows a broad featureless H-adsorption/desorption peak, which can be related to the Co alloying effect [17], while in contrast, both Pt/HSC and a-Pt/HSC samples show distinct characteristic peaks correspond to different facets of Pt nanoparticles at 0.05–0.4 V potential range. After 30k AST cycles, similar characteristic peaks start to show on PtCo/HSC sample, suggesting a leaching out of Co element from the alloy catalyst and less alloying effect. This loss of Co alloying effect is also evident from WAXS measurement results as shown in Fig. S1. For a-Pt/HSC sample, scattering peaks corresponding to different Pt facets get sharper after AST cycles. However, the peak location remains the same. In contrast, the peaks for PtCo/HSC sample have slightly different locations compared to a-Pt/HSC sample and keep shifting towards Pt-related locations. The Fourier transform of the X-ray absorption spectroscopy data acquired at the Pt L3 edge also supports the WAXS data showing loss of Co from the Pt lattice after AST cycling. As shown in Fig. S4, the Pt–Pt scattering in the 2.4 to 3 Å region shifts towards larger radial distances with AST cycling, approaching the Pt–Pt scattering path length of the Pt foil reference (Fig. S4). Fig. 3b presents the ECSA at different stages of AST cycles measured using CO stripping. During the first 30k AST cycles, all three samples exhibit a more pronounced ECSA loss. The a-Pt/HSC sample shows a 41% ECSA loss (40–23 m2pt/gpt), and the PtCo/HSC sample shows 31% ECSA loss (34–23 m2pt/gpt) after the first 30K AST cycles. The Pt/HSC sample has the largest extent of ECSA loss, down by 56% (62–27 m2pt/gpt) after the first 30K AST cycles. From 30k-90K cycles, the a-Pt/HSC and PtCo/HSC samples show similar ECSA loss, while the Pt/HSC shows a faster ECSA loss, likely due to its substantially smaller initial particle size of 2.2 nm. It should be noted that the ECSA here is normalized based on the BOT Pt loading. As shown in Fig. S5, after 90K AST cycles, Pt gets redeposited to the membrane and form Pt-band. For a-Pt/HSC, around 33.0% of Pt from the cathode catalyst layer gets redeposited to the membrane after 90K AST cycles. The ratio of Pt loss to the membrane is 36.3% for PtCo/HSC sample and 27.4% for Pt/HSC sample. Through the entire 90K cycles, the a-Pt/HSC and PtCo/HSC show similar ECSA retention, with a ECSA loss of 61% (or 42% after correction by taking into consideration of Pt loss to the membrane) and 54% (28% after correction), respectively. In contrast, the Pt/HSC has a total ECSA loss of 77% (68% after correction).The catalyst accessibility was measured via CO stripping voltammetry at different RH conditions. At low RH conditions, only the Pt surface that is in direct contact with Nafion ionomer can be accessed through CO stripping, while at high RH, Pt inside the carbon pores can also be active during CO stripping measurement [37]. The change of ECSA of Pt at different RH conditions provides information about the relative amount of Pt located inside carbon pores and at the exterior of the carbon support, or Pt surface in catalyst aggregates/agglomerates that has not been covered by ionomer. Fig. 4 shows plots of normalized ECSA as a function of RH and cycle number to aid in the visualization of catalyst accessibility. Both the a-Pt/HSC and PtCo/HSC sample had a BOT catalyst accessibility of ∼70%, while the Pt/HSC sample had a BOT catalyst accessibility of 45% at 30% RH. As the AST cycling continued, even though the overall ECSA values for all three samples decreased, the Pt accessibility continually increased. After 90k AST cycles, the catalyst accessibility for a-Pt/HSC and PtCo/HSC increased to 95% and 93% respectively, when 30% RH was used as the lower bound. This change in catalyst accessibility caused by catalyst durability AST cycles corresponds well with previous studies by Ramaswamy et al. [17] There are several possible factors that may cause this catalyst accessibility change. During the AST cycles, a large portion of Pt or PtCo catalyst located inside the carbon support could coalesce to larger particles, leading to loss of ECSA for catalyst located inside carbon support. Alternatively, Pt embedded in carbon support mesopores may dissolve and preferentially redeposit to the outside of the carbon support. A third plausible scenario, and one that has been discussed to a lesser extent is ionomer mobilization and intrusion into carbon pore necks, enabling enhanced Pt/ionomer interactions at the EOT.While accessibility measurements can relay a global depiction of Pt accessibility, to yield a more intimate examination of each electrodes evolving electrochemical interfaces, the relative change of Nafion ionomer coverages on both the carbon support and Pt electrocatalyst were examined using H2/N2 impedance. Here, the double layer capacitance was used to separate ionomer and water coverage and Pt and carbon surfaces by measuring capacitance at both low (20%) and (100%) RH conditions, with and without CO introduced to the cathode [28]. Shown in Fig. 7, as the AST proceeds, the double layer capacitance representing water coverage on Pt surface for all three types of MEAs keeps decreasing, suggesting a reduction in accessibility of Pt located inside the carbon support. As the Pt/water (Ptwater) interactions decreased, the signal from carbon/water (Cwater) interfaces also decreased. This can indicate either the collapse of internal carbon structure resulting from localized carbon corrosion, which has a reversible potential of 0.207 V [39] and is enhanced by Pt decorating the support, or the presence of more hydrophobic carbon pores due to the movement of Pt from the pore to the carbon support surface. Though it has been shown that the relative Pt surface in direct contact with ionomer increased with AST cycles (see Fig. 4), utilizing the capacitive approach we can also glean that the Pt/ionomer interactions overall either increased, as is the case for the a-Pt/HSC, or stayed relatively constant, as was the case for both PtCo/HSC and Pt/HSC.This work investigated the degradation of fuel cell catalyst layer for heavy-duty vehicle applications. MEAs with three different types of cathode catalyst, Pt/HSC, a-Pt/HSC and PtCo/HSC, were studied under 90K catalyst specific AST cycles. The PtCo/HSC sample showed much higher BOT mass activity compared to the a-Pt/HSC sample. The mass activity enhancement becomes less significant beyond 30K cycles due to leaching of Co from the catalyst particles, decreasing the intra-particle strain induced by alloying, as supported by the WAXS and STEM-EDS analysis. The Pt/HSC also showed higher initial mass activity but also suffers from faster degradation due to catalyst dissolution and redeposition. From the ECSA measurement, the a-Pt/HSC and PtCo/HSC sample show similar trend of ECSA loss from 30K to 90K AST cycles, while the well dispersed Pt/HSC suffered from more significant ECSA loss. From the H2/air polarization curves, the PtCo/HSC sample has a higher BOT mass transport resistance compared with a- Pt/HSC and Pt/HSC sample. The PtCo/HSC sample's mass transport loss becomes more significant after AST cycles. The catalyst accessibility measurement at different RH conditions using CO stripping method revealed that all three catalysts show less RH sensitivity after AST cycles, which may be a result of Pt preferentially redepositing on the exterior of the carbon support, or preferentially ECSA loss from Pt within carbons pores, along with the continued intrusion of Nafion into carbon pores/pore necks. Results from sulfonate coverage measurements indicate that Pt preferentially redeposits along hydrophobic domains at the ionomer/catalyst interface.This work was authored in part by Alliance for Sustainable Energy, LLC, the manager and operator of the National Renewable Energy Laboratory for the U.S. Department of Energy (DOE) under Contract No. DE-AC36-08GO28308. This material is based on work performed by the Million Mile Fuel Cell Truck (M2FCT) Consortium, technology manager Greg Kleen. Funding was provided by the U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy, Hydrogen and Fuel Cell Technologies Office (HFTO). The X-ray scattering experiments were performed at beamline 9-ID-C at the Advanced Photon Source (APS) at Argonne National Laboratory (ANL). The X-ray spectroscopy experiments were performed at beamline 10-ID at the APS, which is operated by the Materials Research Collaborative Access Team (MRCAT). Use of the APS, an Office of Science user facility operated by ANL, is supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under Contract No. DE-AS02-06CH11357. MRCAT is supported by DOE and by the MRCAT member institutions. The Talos F200X S/TEM tool was provided by US DOE, Office of Nuclear Energy, Fuel Cycle R&D Program, and the Nuclear Science User Facilities. RM acknowledges the support of Los Alamos National Laboratory under Contract No, 89233218CNA000001 operated by Triad National Security, LLC. The views expressed in the article do not necessarily represent the views of the DOE or the U.S. Government. The U.S. Government retains and the publisher, by accepting the article for publication, acknowledges that the U.S. Government retains a nonexclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this work, or allow others to do so, for U.S. Government purposes.

Fingerprint

Dive into the research topics of 'Electrochemical characterization of evolving ionomer/electrocatalyst interactions throughout accelerated stress tests'. Together they form a unique fingerprint.

Cite this