Operando and ex situ comparison of Pt and Pt3Co catalyst degradation under ORR in PEMFC
Abstract
Platinum bimetallic alloys represent a promising class of catalysts for the oxygen reduction reaction in proton exchange membrane fuel cells. Among them, Pt3Co is characterised by higher performances than bare Pt, but also by different stability, as Co leaching is known to take place during fuel cell operation. To underline differences of catalyst behaviour in real operating conditions, the evolution of bare Pt and Pt3Co catalyst nanoparticles are here compared from pristine conditions, up to catalyst activation and aging via specific break-in and Accelerated Stress Tests (ASTs) procedures, respectively. Changes in catalyst chemistry were monitored via x-ray absorption and photoelectron spectroscopies, and via SEM-EDX. Results were combined with morphology analysis carried out via small-angle x-ray scattering. Results from both operando and ex situ measurements show as for bare Pt catalyst, both particle morphology and the ratio among metal-to-oxidised Pt do not change remarkably after the break-in, and that the Electrochemically Active Surface Area (ECSA) strongly reduces due to average particle size growth from 2.28 to 6.21 nm within the first 3000 AST cycles. Conversely, in Pt3Co catalyst, Co leaching strongly affects the break-in stage, by reducing particle size and decreasing the fraction of metallic Co. During ASTs, leaching continues also after the formation of the Pt-rich skin, which formation contributes to slow down ECSA reduction, with particle size growing from 2.59 to 6.14 nm in between 3000 and 6000 cycles only.
Full text
1 Operando and ex situ comparison of Pt and Pt3Co catalyst degradation under ORR in PEMFC Marco Bogar1, Yurii Yakovlev2, Simone Pollastri3,4, Tomáš Hrbek2, David Kalabis2, Giovanna Marussi5, Matteo Crosera5, Roberto Biagi3,6,7, Heinz Amenitsch8, Rodolfo Taccani1, Iva Matolínová2 1 Department of Engineering and Architecture, University of Trieste, Via Alfonso Valerio 6/1, 34127 Trieste, Italy 2 Department of Surface and Plasma Science, Faculty of Mathematics and Physics, Charles University, V Holešovičkách 2, 180 00 Prague 8, Czech Republic. 3 Department of Physics, Computer Science and Mathematics, University of Modena and Reggio Emilia, 41125, Modena, Italy 4 ELETTRA - Sincrotrone Trieste S.C.p.A., SS 14 - km 163,5, 34149, Basovizza, Trieste, Italy. 5 Department of Chemical and Pharmaceutical Sciences, University of Trieste, Via Giorgeri 1, Trieste 34127, Italy 6 Istituto Nanoscienze (NANO-S3), Consiglio Nazionale delle Ricerche (CNR), via G. Campi 213/a, Modena, 41125, Italy 7 Centro H2-MORE, University of Modena and Reggio Emilia, 41125, Modena, Italy 8 Graz University of Technology, Institute for Inorganic Chemistry, Stremayrgasse 9, 8010 Graz, Austria.
2 Supplementary Material Supplementary Tables Table S1. XANES analyses of the Pt3Co catalyst in pristine conditions (in situ, within EC cell). Linear combination fitting results of the XANES spectra presented in Figures 1b and 1c of the main manuscript, collected from the MEA loaded with the Pt3Co catalyst nanoparticles in pristine conditions. Spectra collected at the Pt L3-edge Spectra collected at the Co K-edge Pt0 96 ± 3 Co0 6 ± 2 PtO 4 ± 3 Pt3Co 53 ± 3 Co(H2O)62+ 22 ± 5 CoO 19 ± 3 Table S2. Complete results from the EXAFS refinements on spectra from samples in pristine conditions. Crystallographic data and structural parameters as obtained from the R-space fit by using the theoretical references. Catalyst Absorber N Atom % r (Å) S02 R-factor σ2 (Å2) ΔE0(eV) Pta Pt 12a,b Pt - 2.748 (3)a 0.57 (8)a 0.008a 0.005 (1)a 8 (1)a Pt3Co Pt 8c Pt - 2.706 (9) 0.73 (9) 0.025 0.005 (4) 7.0 (1) 4c Co - 0.009 (*) Co 12c Pt 57 2.685 (7) 0.8 (2) 0.009 0.008 (2) -5* 12d Co 10 2.632 (7) 0.007 (5) 6e O 33 2.068 (7) 0.005 (5) a Reported from previous work of the same authors [1]. b Shell calculated from the crystallographic data of metallic Pt of Wyckoff [2]. c Shell calculated from the crystallographic data of Pt3Co of Geiser and Martin [3]. d Shell calculated from the crystallographic data of metal Co of Häglund et al. [4]. e Shell calculated from the crystallographic data of CoO of Saito et al. [5]. *Fixed value Table S3. ICP-MS. Complete ICP-MS results. Obtained from waters produced during operando SAXS measurements. Catalyst Pt/C Pt3Co Analyte Pt Pt Co Concentration measured during: break-in / AST (μg/L) - / - - / - 31.0 / - Limit of detection (μg/L) 0.01 0.01 0.05 Table S4. XANES linear combination fitting results from spectra measured in operando during break-in with the Pt3Co catalyst. Linear combination fitting results within the XANES region of the spectra collected at the Co K-edge from the MEA leaded with the Pt3Co catalyst nanoparticles measured in operando conditions (Figures 2e and 3e). Pristine After break-in Co0 6 ± 1 9 ± 5 Pt3Co 53 ± 2 17 ± 6 Co(H2O)62+ 21 ± 5 56 ± 9 CoO 18 ± 3 18 ± 4 Table S5. Results of the XANES linear combination fitting analyses on data collected on ex situ MEAs loaded with the Pt3Co catalyst. Linear combination fitting results within the XANES region of the spectra collected from the MEA loaded with the Pt3Co catalyst nanoparticles measured ex situ at the most significant time points, selected based on operando electrochemical and SAXS analysis. Pristine After 1250 cycles After 2500 cycles After 5000 cycles Spectra collected at the Pt L3-edge Pt0 93 ± 7 93 ± 9 92 ± 9 88 ± 7 PtO 7 ± 2 7 ± 5 8 ± 6 12 ± 8 Spectra collected at the Co K-edge Pt3Co 61 ± 4 48 ± 3 68 ± 5 74 ± 5 Co(H2O)62+ 21 ± 5 37 ± 8 12 ± 8 4 ± 4 CoO 18 ± 3 15 ± 2 20 ± 3 22 ± 3
3 Supplementary Figures Figure S1. XAS characterization in pristine conditions. (a) Spectrum of the MEA loaded with the Pt3Co catalyst nanoparticles measured at the Pt L3-edge with the best linear combination fit and residual curve obtained using metallic Pt (Pt0) and PtO2·H2O. (b) Spectrum of the MEA loaded with Pt3Co measured at the Co K-edge with the best linear combination fit and residual curve obtained using metal CoO and the spectra of Pt3Co(*) and Co(H2O)62+(*). (c) EXAFS first shell fit on the spectrum collected on the MEA loaded with the Pt3Co nanoparticles and measured at the Pt L3-edge. (d) Fourier-Transformed EXAFS of the Pt L3-edge spectra of Figure 1b, collected from the MEAs hosting the Pt and Pt3Co catalyst nanoparticles referred to a foil of metallic Pt; the red dashed line is used to highlight the shift to lower “r” values for the Pt3Co (light blue line). Notes: (*) spectra from the work of Takao and co-workers [6].
4 Figure S2. In operando NAP-XPS. (a, b, c) Pt 4f and (d, e, f) Co 2p XPS spectra collected from the cathode electrode loaded with the Pt3Co catalyst. Measurements: (a,d) in pristine conditions, (b,e) by applying a constant 1.5 V potential after the break-in procedure, and (c,f) after 50 CV cycles with the potential sweeping between 1.0 and 1.5 V. Note: the PtX+ represents the fraction of highly-oxidised Pt specimen which distinction is beyond the scope of this work. PtZ+ represents the overall oxidised Pt fraction. Figure S3. Pt catalyst, results of SAXS data fitting during the break-in step. Time resolved evolution of the parameters composing the analytical model used for fitting SAXS patterns recoded in operando represented in Figure 2a. Form factor parameters representing the catalyst nanoparticles: (a) forwarded scattering probability, IP, (b) mean particle
5 diameter, DP, and (c) standard deviation, σP, within the Schultz distribution; (d) calculated particle volume fraction, φ. Structure factor parameters: (e) fractal number, Df, and (f) calculated radius of gyration of the fractal aggregate, Rg. Debye-Anderson-Brumberger form factor used to represent the Vulcan support: (g) forwarded scattering probability, A; the correlation length value, ξ was kept fixed at 36.07 nm, as previously done [1,7]. Voigt peak parameters used to model the ionomer peak from Nafion electrolyte: (h) peak intensity, IIP, (i) peak width, σIP, and (j) peak position, qIP; the shape factor (sIP) was kept fixed at 0.45, as previously done [1,7]. Power law parameters, used to model the GDLs: (k) forwarded scattering probability, C, and (l) power law exponent, p. Note: dashed lines are guides for the eye only. Figure S4. Pt3Co catalyst, results of SAXS data fitting during the break-in step. Time resolved evolution of the parameters composing the analytical model used for fitting SAXS patterns recoded in operando represented in Figure 2c. Form factor parameters representing the catalyst nanoparticles: (a) forwarded scattering probability, IP, (b) mean particle diameter, DP, and (c) standard deviation, σP, within the Schultz distribution; (d) calculated particle volume fraction, φ. Debye-Anderson-Brumberger form factor used to represent the Vulcan support: (e) forwarded scattering probability, A; the correlation length value, ξ, was kept fixed at 36.07 nm, as previously done [1,7]. Voigt peak parameters used to model the ionomer peak from Nafion electrolyte: (f) peak intensity, IIP, (g) peak width, σIP, and (h) peak position, qIP; the shape factor (sIP) was kept fixed at 0.45, as previously done [1,7]. Power law parameters, used to model the GDLs: (i) forwarded scattering probability, C, and (j) power law exponent, p. Note: dashed lines are guides for the eye only.
6 Figure S5. Pt catalyst, results of SAXS data fitting during ASTs. Time resolved evolution of the parameters composing the analytical model used for fitting SAXS patterns recoded in operando represented in Figure 4b. Form factor parameters representing the catalyst nanoparticles: (a) forwarded scattering probability, IP, (b) mean particle diameter, DP, and (c) standard deviation, σP, within the Schultz distribution; (d) calculated particle volume fraction, φ. Structure factor parameters: (e) fractal number, Df, and (f) calculated radius of gyration of the fractal aggregate, Rg. Debye-AndersonBrumberger form factor used to represent the Vulcan support: (g) forwarded scattering probability, A; the correlation length value (ξ) was kept fixed at 36.07 nm, as previously done [1,7]. Voigt peak parameters used to model the ionomer peak from Nafion electrolyte: (h) peak intensity, IIP, (i) peak width, σIP, and (j) peak position, qIP; the shape factor (sIP) was kept fixed at 0.45, as previously done [1,7]. Power law parameters, used to model the GDLs: (k) forwarded scattering probability, C, and (l) power law exponent, p. Fill dots (•) represent data retrieved from fitting SAXS patterns collected in operando conditions (figure 4b), while empty triangles (Δ) represent data retrieved from fitting SAXS patterns collected ex situ (Figure S7). Dashed lines are guides for the eye only. Figure S6 Ex situ and operando comparison of SAXS patterns. Operando SAXS patterns are compared with patterns collected from twin MEAs where aging was stopped at specific time points of the AST. All MEAs were loaded with the Pt nanoparticles as cathode catalysts.
7 Figure S7. Pt3Co catalyst, results of SAXS data fitting during ASTs. Time resolved evolution of the parameters composing the analytical model used for fitting SAXS patterns represented in Figure 4e. Form factor parameters representing the catalyst nanoparticles: (a) forwarded scattering probability, IP, (b) mean particle diameter, DP, and (c) standard deviation, σP, within the Schultz distribution; (d) calculated particle volume fraction, φ. Debye-AndersonBrumberger form factor used to represent the Vulcan support: (e) forwarded scattering probability, A; the correlation length value (ξ) was kept fixed at 36.07 nm, as previously done [1,7]. Voigt peak parameters used to model the ionomer peak from Nafion electrolyte: (f) peak intensity, IIP, (g) peak width, σIP, and (h) peak position, qIP; the shape factor (sIP) was kept fixed at 0.45, as previously done [1,7]. Power law parameters, used to model the GDLs: (i) forwarded scattering probability, C, and (j) power law exponent, p. Fill dots (•) represent data retrieved from fitting SAXS patterns collected in operando (figure 4e), while empty triangles (Δ) represent data retrieved from fitting SAXS patterns collected ex situ (Figure S9). Dashed lines are guides for the eye only. Figure S8. Ex situ and operando comparison of SAXS patterns. Operando SAXS patterns are compared with patterns collected from twin MEAs where aging was stopped at specific time points of the AST. All MEAs were loaded with the Pt3Co nanoparticles as cathode catalysts.
8 Figure S9. Number-weighted particle size distribution. Time resolved evolution of number weighted size distribution (within the Schultz distribution) of (a) the bare Pt nanoparticles and (c) the Pt3Co catalyst nanoparticles undergoing ASTs, calculated from the results from least square fitting parameters shown in Figures 5a and 5c, respectively. Skewness of the probability distribution function (ν) was further calculated for both (b) the Pt, and (d) the Pt3Co catalyst nanoparticles. Dashed lines are guides for the eye only. Figure S10. As collected XANES spectra at the Co K-edge on the MEA loaded with the Pt3Co catalyst in ex situ conditions in transmission mode in pristine conditions and after having applied ASTs.
9 References [1] M. Bogar, Y. Yakovlev, S. Pollastri, R. Biagi, H. Amenitsch, R. Taccani, I. Matolínová, Capabilities of a novel electrochemical cell for operando XAS and SAXS investigations for PEM fuel cells and water electrolysers, J. Power Sources 615 (2024) 235070. https://doi.org/10.1016/j.jpowsour.2024.235070. [2] R.W.G.Wtckoff, Crystal Structures Volume 1, American Mineralogist Crystal Structure Database, 1963. [3] A.H. Geisler, D.L. Martin, A new superlattice in Co-Pt alloys [8], J. Appl. Phys. 23 (1952) 375. https://doi.org/10.1063/1.1702216. [4] J. Häglund, A. Fernndez Guillermet, G. Grimvall, M. Körling, Theory of bonding in transition-metal carbides and nitrides, Phys. Rev. B 48 (1993) 11685–11691. https://doi.org/10.1103/PhysRevB.48.11685. [5] S. Saito, K. Nakahigashi, Y. Shimomura, X-ray diffraction study on CoO, J. Phys. Soc. Japan 21 (1966) 850–860. https://doi.org/10.1143/JPSJ.21.850. [6] S. Takao, O. Sekizawa, K. Higashi, G. Samjeské, T. Kaneko, T. Sakata, T. Yamamoto, T. Uruga, Y. Iwasawa, Visualization Analysis of Pt and Co Species in Degraded Pt3Co/C Electrocatalyst Layers of a Polymer Electrolyte Fuel Cell Using a Same-View Nano-XAFS/STEM-EDS Combination Technique, ACS Appl. Mater. Interfaces 12 (2020) 2299–2312. https://doi.org/10.1021/acsami.9b16393. [7] M. Bogar, Y. Yakovlev, J. Nováková, A.M. Darabut, M. Kriechbaum, H. Amenitsch, R. Taccani, I. Matolínová, A small angle X-ray scattering approach for investigating fuel cell catalyst degradation for both ex situ and in operando analyses, Int. J. Hydrogen Energy 58 (2024) 1673–1681. https://doi.org/10.1016/j.ijhydene.2024.01.261.