Modifying the Electrocatalyst–Ionomer Interface via Sulfonated Poly(ionic liquid) Block Copolymers to Enable High-Performance Polymer Electrolyte Fuel Cells

Polymer electrolyte membrane fuel cell (PEMFC) electrodes with a 0.07 mgPt cm–2 Pt/Vulcan electrocatalyst loading, containing only a sulfonated poly(ionic liquid) block copolymer (SPILBCP) ionomer, were fabricated and achieved a ca. 2× enhancement of kinetic performance through the suppression of Pt surface oxidation. However, SPILBCP electrodes lost over 70% of their electrochemical active area at 30% RH because of poor ionomer network connectivity. To combat these effects, electrodes made with a mix of Nafion/SPILBCP ionomers were developed. Mixed Nafion/SPILBCP electrodes resulted in a substantial improvement in MEA performance across the kinetic and mass transport-limited regions. Notably, this is the first time that specific activity values determined from an MEA were observed to be on par with prior half-cell results for Nafion-free Pt/Vulcan systems. These findings present a prospective strategy to improve the overall performance of MEAs fabricated with surface accessible electrocatalysts, providing a pathway to tailor the local electrocatalyst/ionomer interface.


Electrode Fabrication and Assembly
All catalyst coated membranes (CCM) were prepared using a similar catalyst ink formulation.Firstly, the ionomer (such as SPILBCP and Nafion D2020 Ion Power) was predispersed in acetonitrile with desired ionomer-to-carbon ratio (I/C) prior to adding water.Next, 47.7 wt% Pt/Vu (TKK, TEC10V50E) was added into a mixture of ionomer, deionized water (DI) and acetonitrile.The catalyst suspensions were then dispersed with 20 s of horn sonication followed by 20 mins of ice-bath sonication. 1,2The catalyst layer was ultrasonically sprayed onto Nafion NR211 membranes (IonPower) using a Sono-Tek spray station with 25kHz Accumist nozzle at a target catalyst loadings of 0.07 mg Pt /cm 2 for cathodes.Pt loadings on each individual electrode were verified by X-ray Fluorescence Spectroscopy (XRF) (Fisher XDV-SDD).Anode electrodes were prepared with PtCo/HSC (Umicore, Pt30 0670) dispersed in DI:n-PA (70 wt% water) with a I/C of 0.9 (Nafion:C mass ratio) and a 0.03 mg Pt /cm 2 loading.
Once fabricated, the CCMs were assembled into either 50 cm 2 hardware for performance metrics measurements such as H 2 /O 2 polarization data, H 2 /Air polarization data, kinetic activities, and ECA, or 5 cm 2 differential cells for O 2 limiting current experiments. 3The CCMs were sandwiched between either two 50 cm 2 SGL 29 BC gas diffusion layers (GDLs) at 25% compression or 5 cm 2 Freudenberg GDLs at 18% compression.The CCMs, GDLs, and polytetrafluoroethylene (PTFE) gaskets were then placed between the flow fields and the bolts tightened to 40 inch-pounds.

In situ Electrochemical Diagnostics
A customized Hydrogenics test station was used to perform all fuel cell testing.All the MEAs tested in this work were subjected to an established conditioning protocol 4 consisting of an initial break-in procedure that activates the cell and a series of voltage recovery steps which were followed by polarization measurements.

Break-in.
The break-in procedures begin by heating the cell to 80°C, and holding the cell at an open circuit potential (load equivalent flow rates of 0.8/2.5 = H 2 /Air L std /min), followed by a series of 5/10/5 voltage cycles in the fuel cell regime of 0.60 V-0.90 V for 4 minutes. 4ltage Recovery (VR).The voltage recovery (VR) step exposed the cell at 0.1 V cell under H 2 /Air (950/500 sccm respectively) for 2 hours at 40 °C and 150% RH.
H 2 /O 2 Polarization Curves.The test protocol involved measuring the I-V curves from 0.4 V to OCV at 80 °C at 100 kPa O 2 partial pressure (150 kPa total pressure) and 100 % RH for 4 min per point (average of last 1 min used) in the anodic direction.The ORR mass activities were reported at 0.90 V after applying high frequency resistance (HFR) and hydrogen cross-over corrections.
H 2 /Air Polarization Curves.The test protocol involved measuring the I-V curves from 0.3 V to OCV at 80 °C and at 150 kPa total pressure with 75 % RH for 4 min per point (average of last 1 min used) in the anodic direction.

CO Stripping Voltammetry.
Pt electrochemical surface area (ECA) was determined by integrating the CO stripping charge (Q CO ) obtained from cyclic voltammetry (CV) after the introduction of CO to an equilibrated electrode held at 0.2 V cell .The cathode feed was purged with pure N 2 at 0.25 L std /min prior to the first anodic sweep.CVs were performed immediately at 80 °C and different RH under H 2 /N 2 sweeping from 0.05 to 0.9 V at 20 mV/s.420 μC/cm 2 was assumed as the unit charge for CO integrated areas in determining the ECA.
O 2 Limiting Current Experiments.Limiting current measurements were performed at 80 °C and 75/30% RH, with 0.02, 0.03 and 0.05 mole fraction of oxygen.The limiting current was obtained at total cell pressures of 100, 150, 200 and 300 kPa.Limiting current was measured at constant voltages of 0.30, 0.24, 0.18, 0.12, and 0.06 V and held for 3 minutes.Due to the impact of hydrogen evolution on current densities obtained below 0.1 V, the maximum of the resulting current densities above 0.12 V was reported as the limiting current.Additional details on the procedure were previously described by Baker et.al. 3 AC impedance measurement.Electrochemical impedance spectroscopy (EIS) experiments were performed on fully conditioned 50 cm 2 MEAs with 0.07 mg Pt cm -2 at 80 °C and 100% RH using a Gamry Reference 3000 Potentiostat connected to a Gamry 30k Booster.EIS experiments were measured 50kHz-50mHz at 0.45 V vs RHE with 1 atm H 2 and N 2 at anode and cathode gas lines respectively.

Ink Characterization
Dynamic Light Scattering (DLS).The DLS measurements were performed using Zetasizer Nano ZS (Malvern Instruments Ltd, Malvern, U.K.).These measurements were performed on relatively dilute inks compared to inks used for MEA fabrication at 0.1 wt% Pt/Vu as desired by the technique to avoid multiple scattering.And the DLS measurements of ionomer solutions were carried out at 0.024 wt% using a disposable cuvette cell.At least five measurements were taken to ensure repeatability of the results.

Oxide Coverage and Kinetic Model Calculations
A similar methodology has been described elsewhere. 5In this study, the anodic sweep of cyclic voltammetry experiments (Figure 1d) was used to measure surface oxidization of Pt/Vu electrodes, which occurred between 0.45 and 1.05V cell .All spectra were first corrected by a background current measured at 0.45V to eliminate contributions from H 2 crossover.Slight differences in metal loading and electrochemical surface area were corrected by scaling the background corrected currents accordingly.These corrected spectra were then integrated from 0.45V to each measured potential until the upper bound of 1.05V was reached.These charges were then normalized by the total oxidization charge observed on the 0.6Nafion:C Pt/Vu electrode between 0.45 and 1.05V cell , to yield a fractional oxide coverage as a function of cell potential (Figure 4a) using Equation S1.The coverage values were directly applied in a model previously developed by Subramanian et al (Equation S2) to predict coverage-dependent ORR kinetics displayed in Figure 4b. 5 Note, this simplistic treatment does not account for differences in local O 2 concentration or permeability, which might result from PILBCP integration.Overall, we observe that the mixed ionomer MEAs exhibit better performance than any individual component analogues.However, further improvement will depend on the optimal ratio between two ionomers, to achieve a more appropriate electrode microstructure, and the interaction between the catalyst and PILBCP, which all be the focus of future research.represents the oxide coverage, cell voltage, and normalized oxidization current    , ,  , measured on different MEAs, and corresponds to the total oxidization charge observed    , on the 0.6Nafion:C Pt/Vu electrode evaluated between 0.45 and 1.05V.Figure S5 graphically represents how oxide coverage was calculated.

ORR Kinetics Model Parameters and Calculation 5
ORR kinetics measured as current density (i) depends many experimental variables such as the operating conditions (O 2 partial pressure , temperature , cathodic overpotential ), the intrinsic  2   properties of catalytic sites (exchange current density , oxide/adsorbate coverage ), and  0  parameters related to the reaction mechanism (reaction order w.r.t O 2 , charge transfer coefficient  , coverage dependent adsorption energetics ).R and F correspond to the universal gas constant   and Faraday constant, respectively.

Figure S4
Figure S4 Nyquist plots for fully conditioned Pt/Vu MEAs with indicated ionomer(s) at 80 ℃ and 100% RH.

Figure S5 .
Figure S5.Particle size distribution of dilute (0.024 wt%) ionomer dispersions in the solvent of 60 wt% water and 40 wt% acetonitrile characterized by DLS.

Figure S6 .
Figure S6.Integration of normalized CV data to calculate oxide coverage on 0.6SPILBCP:C MEA at 0.8Vcell.The red and black regions correspond to the cumulative oxidation charge experience on 0.6SPILBCP:C to 0.80V and 0.6Nafion:C electrodes to 1.05V, respectively.

Table S2 .
Average electrochemical active surface area (ECA) and roughness factor (RF) for MEA with Pt/Vu with loading of 0.07 mg Pt cm -2 after 8 VRs.

Table S1 .
Z avg diameter of dilute ionomer (0.024 wt%) and catalyst (0.1 wt%)+ionomer (0.024 wt%) dispersions characterized by DLS.The ratios are mass based.w stands by water, and AcN stands by acetonitrile.