Poly(ionic liquid) Ionomers Help Prevent Active Site Aggregation, in Single-Site Oxygen Reduction Catalysts

Anion exchange membrane fuel cells (AEMFCs) can produce clean electricity without the need for platinum-group metals at the cathode. To improve their durability and performance, most research investigations so far have focused on optimizing the catalyst and anion exchange membrane, while few studies have been dedicated to the effect of the ionomer. Herein, we address this gap by developing a poly(ionic liquid)-based ionomer and studying its effect on oxygen transport and oxygen reduction kinetics, in comparison to the commercial proton exchange and anion exchange ionomers Nafion and Fumion. Our study shows that the choice of ionomer has a dramatic effect on the morphology of the catalyst layer, in particular on iron aggregation. We also observed that the quality of the catalyst layer and the degree of iron aggregation can be correlated to the rheological properties of the catalyst ink. Moreover, this work highlights the impact of the ionomer on the resistance to oxygen transport and reports improved oxygen diffusion compared to Nafion, for poly(ionic liquid)s with fluorinated anions. Finally, the performance of the catalyst–ionomer layer for oxygen reduction was tested with a rotating disc electrode (RDE) and a gas diffusion electrode (GDE). We observed dramatic differences between the two configurations, which we attribute to the different morphologies of the catalyst layer. In summary, our study highlights the dramatic and overlooked effect of the ionomer and the limitations of the RDE in predicting fuel cell performance.


Note on ink solvent
The effect of solvent was investigated for Fumion and Nafion, which were tested in both 50% IPA/water and pure ethanol.Normally, a combination of alcohol and water is used in the ink, but the hereby synthesised polymers were insoluble in water or isopropanol, so pure ethanol was used instead.The same solvent was studied for Fumion and Nafion.The choice of solvent was found to have a small effect on the activity of the catalyst layer, but for the case of Fumion and Nafion, ethanol-based inks were found to be particularly challenging to deposit and results were less reproducible.Therefore, results with IPA/water solvents are shown in the manuscript.Figure S 12a shows a comparison of the activity of FePC/G with Fumion, in the case of IPA/water and ethanol solvent

Note on Fumion activation
Anion exchange ionomer are commonly delivered with a halogenated anion and should undergo ion exchange before use.In the case of Fumion, the ionomer comes in the brominated form and ion exchange usually happen by immerging the deposited catalyst layer in 1M KOH for 1 to 24 hours.In this work, we found that the activity of the catalyst layer does not improve after this ion exchange step either for the case of catalyst drop casted on an RDE electrode or spray coated on carbon paper and tested in GDE.For example, Figure S 12b shows the activity of the same RDE electrode before and after immersing it for 1h in 1M KOH.After soaking the electrodes in 1M KOH overnight the activity dropped slightly, possibly indicating that the ionomer is not stable in alkaline conditions.Therefore all the results shown in the manuscript were obtained after 1h KOH activation. of 10 mV/s.The catalyst is FePC/G and the ionomer is Fumion, the I/C ratio is 1 for both samples and the solvent for the catalyst ink is 50% IPA in water.The samples differ for the ionomer activation: the darker curve records the activity of the catalyst layer as prepared, while the lighter curve represents the same RDE electrode after soaking it 1h in 1M KOH c) Square wave voltammograms obtained in nitrogen-saturated 0.1M KOH, in static conditions, with a frequency of 2Hz, with 4mV steps and a modulation amplitude of 20mV.The catalyst is FePC/G (darker curve) and PC/G (lighter curve).The ionomer is Fumion, the solvent of choice for the catalyst ink is 50% IPA in water and the I/C ratio is 1.

Note on Fumion CV peaks
As it can been seen in Figure 7d in the manuscript, the cyclic voltammogram of FePC containing Fumion shows a broader peak shifted to higher potential, compared to other ionomers.To further understand the change in shape of this peak, we recorded square wave voltammograms (SWV).
Compared to a common CV, in SWV the potential is stepped up and down in every step, allowing to reduce the capacitative contribution and to isolate peaks originating from faradaic processes.The results are shown in Figure S 12c.As it can be observed, the low potential peak is actually composed of two separate peaks.Since this extra peak was not observed for any other ionomer, we hypothesise that it originates from the ionomer.To confirm this hypothesis, we repeated the experiment with the same catalyst without iron, for which no CV peak is normally observed.The PC catalyst with Fumion presented a broad peak at around 0.45V vs RHE, which again was no observed in the absence of Fumion.This confirms that the peak is likely originate from an electron transfer at the ionomer.

Note on Impedance
There is still debate regarding the best fitting for impedance data collected with a rotating disc electrode.However, there is general consensus that the high frequency semicircle can be modelled using a Randel circuit, composed of a polarization resistance (Rp), double layer capacitance (Cdl) and uncompensated resistance (Ru). 2 The uncompensated resistance, or known as ohmic drop, represents the ohmic resistance of the electrolyte between the reference and the working electrode, while Cdl represents the charging and discharging of the electric double layer at the surface of the electrode.The main component of the polarization resistance is the charge transfer resistance, which is related to the kinetics of heterogeneous electrochemical processes.The semicircle present at lower frequency is generally associated with oxygen transport.This can either be represented by a simple resistance (Ro), 3,4 or by a Warburg element (Zo) for the case of a transmissive boundary 2 .

Figure S 8 :Figure S 9 :
Figure S 8: Dynamic light scattering of the catalyst inks, diluted 100 times.After dilution, the catalyst inks contained 0.04mg of catalyst and 0.04mg of ionomer per mL of ink.The solvent was either ethanol or 50% isopropanol in water.

Figure S 12 :
Figure S 12: Further electrochemical studies of Fumion ionomer a) Linear sweep voltammogram obtained in oxygen-saturated 0.1M KOH, at a rotational speed of 1600 rpm and scan rate of 10 mV/s.The catalyst is FePC/G and the ionomer is Fumion, the I/C ratio is 1 for both samples.The samples differ in the choice of solvent for the ink: 50% IPA in water (darker curve) and ethanol (lighter curve) b) Linear sweep voltammogram obtained in oxygen-saturated 0.1M KOH, at a rotational speed of 1600 rpm and scan rate

Finally,Figure S 13 :
Figure S 13: Equivalent circuit for fitting of Impedance spectroscopy measurements

Figure S 15 :
Figure S 15: Effect of catalyst loading on the oxygen reduction performance.a) shows the oxygen reduction performance of FePC/G with Nafion ionomer, at the loadings indicated in the legend.b) shows the same results, when using PIL2 as an ionomer.For all the measurements, an I:C ratio of 1 was used.c) shows the current density at 0.85V vs RHE as a function of the catalyst loading

Figure S 16 :
Figure S 16: Optical microscopy images of the RDE electrode with a drop-casted layer of 0.14 mg cm -2 FePC/G with the ionomers Nafion, PIL2 and PIL3, with an I:C ratio of 1.All the inks contained the following: 4mg/ml FePC/G, ionomer with I/C ratio=1.The solvent used was ethanol for PIL2, PIL3 and 50% isopropanol in water for Nafion.

Table S1 :
elemental composition at the locations shown in Figure S6, as determined by EDX Position C (%w) N (%w) O (%w) F (%w) S (%w) Fe (%w) Catalyst Layer Morphology -Rheology and Particle Size Figure S 7: Rheology of the catalyst inks.All the inks contain 4mg/ml FePC/G catalysts, with various ionomers in a I/C ratio of 1.The solvent is either ethanol of 50%w isopropanol in water a) Shear stress as a function of shear rate at low shear rates, for the inks in ethanol, b) Shear stress as a function of shear rate at low shear rates, for the inks in 50% isopropanol in water c) Viscosity as a function of shear rate in ethanol d) viscosity as a function of shear rate in 50% IPA/water

Table S2 :
Hydrodynamics radius and polydispersity of the aggregates in the catalyst ink, measured with dynamic light scattering and averaged over at least 3 repeats of 10 runs each.It should be noticed that since the polydispersity index is high, the average diameter can be overestimated, 1 therefore these results should only be taken as a qualitative indication of the catalyst ionomer interaction.