Understanding Electrocatalytic Nitrate Reduction Catalyst Performance in Ideal and Practical Conditions

Abstract

Anthropogenic activities which modify atmospheric nitrogen gas to more reactive forms of nitrogen such as ammonium are essential for sustaining the growing global population and maintaining modern aspects of life. Electrocatalytic nitrate reduction is identified as a promising technology for nitrate remediation of waste streams and water sources but needs further development. Here we aim to understand the various levels of activity for transition metal nitrate reduction catalysts and evaluate catalyst activity under conditions which would be closer to remediation conditions. Namely, chloride is present in nearly all nitrate polluted waste streams and so we investigate the effect of chloride. Additionally, to remediate large volumes of waste streams, continuous flow electrochemical reactors are necessary, so we investigate the electrocatalysts in flow reactors to compare to batch reactor configurations. First, we use density functional theory (DFT) modeling and microkinetic modeling to identify key parameters (descriptors) influencing nitrate reduction activity on seven transition metals. The descriptors are N and O adsorption energies to the metal surface, because these adsorption energies control the strength of nitrate adsorption and correlate with hydrogen adsorption. The best performing catalysts adsorb nitrate strongly, but not so strongly as to poison the surface. The computed trends match the reported experimental activity measurements for these catalysts in the literature. We validated the adsorption energy trends by probing the adsorption of nitrate using cyclic voltammetry on Pt and Rh. We confirmed that Rh adsorbs nitrate more strongly than Pt, which explains the higher activity of Rh for nitrate reduction. With these descriptors we predict alloys that will be active for nitrate reduction based on their computed N and O adsorption energies, including PtRu. Second, we investigate the effects of chloride during nitrate reduction. We show that for active catalysts like Rh the nitrate reduction activity is more severely decreased compared to a less active catalyst like Pt when chloride is present. The decrease in activity is due to competitive adsorption between nitrate and chloride. The more severe chloride poisoning for Rh was unexpected because nitrate binds more strongly on Rh than on Pt. However, we show that chloride also binds so much more strongly on Rh than on Pt that chloride outcompetes nitrate for active sites. Using DFT we show a relation between nitrate and chloride adsorption energies on transition metals such that metals that adsorb nitrate strongly will adsorb chloride even more strongly, and thus be poisoned by chloride. To address this, we explore RhxSy as a chloride-resistant electrocatalyst. We show that RhxSy is active for nitrate reduction, but is poisoned by chloride to a similar extent as Pt. We use DFT and microkinetic modeling to show that possible active sites for nitrate reduction on RhxSy are sulfur vacancy sites that adsorb nitrate strongly, explaining the high activity, but also adsorb chloride strongly, explaining the moderate chloride poison resistance. Third, we compare the activity of Pt/C, PtRu/C, and Rh/C when used in a batch cell on a rotating disk electrode to a flow cell on a carbon felt. We show that the activity measured in the batch cell does not match that in the flow cell, but the trends are the same in both systems (Pt < PtRu < Rh). The activity trends match those expected from the O and N adsorption energies from our earlier work.