Identification of carbon-encapsulated iron nanoparticles as active species in non-precious metal oxygen reduction catalysts

The widespread use of fuel cells is currently limited by the lack of efficient and cost-effective catalysts for the oxygen reduction reaction. Iron-based non-precious metal catalysts exhibit promising activity and stability, as an alternative to state-of-the-art platinum catalysts. However, the identity of the active species in non-precious metal catalysts remains elusive, impeding the development of new catalysts. Here we demonstrate the reversible deactivation and reactivation of an iron-based non-precious metal oxygen reduction catalyst achieved using high-temperature gas-phase chlorine and hydrogen treatments. In addition, we observe a decrease in catalyst heterogeneity following treatment with chlorine and hydrogen, using Mössbauer and X-ray absorption spectroscopy. Our study reveals that protected sites adjacent to iron nanoparticles are responsible for the observed activity and stability of the catalyst. These findings may allow for the design and synthesis of enhanced non-precious metal oxygen reduction catalysts with a higher density of active sites.

treated with Cl2 at 600 °C for 30 minutes with peak locations for reduced Fe species: Fe3C, α-Fe (mag) and α-Fe (spm). Treatment with Cl2 at 600 °C does not remove all reduced Fe species present in the as-prepared catalyst. The smaller absorption area signifies that only the exposed Fe is removed. We note that similar results treating with Cl2 at 650 °C have been previously observed. 1 treated with H2 at 600 °C for 30 minutes following Cl2 treatment at 900 °C with peak locations for reduced Fe species: Fe3C, α-Fe (mag) and α-Fe (spm). Additionally, a new species identified as Fe3S4 is present with fitting parameters given in Supplementary Table 4 which has been reported in previous literture. 2 It is evident that the presence of this Fe3S4 has no effect on catalyst activity. Treatment with H2 at 600 °C effectively reduces the Fe present in the Cl2treated catalyst to reform the reduced Fe species present in the as-prepared catalyst and catalyst treated with H2 at 900 °C. observed which is attributed to the destruction of the active species due to the temperature above that used during catalyst synthesis. In order to prevent temperature effects leading to deactivation, gas treatments were carried out at 900 °C and below. using ICP-OES, Cl content using ion selective electrode, and N content from CHN analysis.

Supplementary
Note that the wt% of Fe in the H2-treated sample is greater than in the other samples in part due to a decrease in total mass cause by the removal of Cl and etching of C during the treatment.  Fig. 4a). 16  material were obtained to be similar, but the coordination numbers were found to be different ( Table 1).

Sample Fe (wt%) Cl (wt%) N (wt%)
For the H2-treated catalyst multiple-scattering analysis was performed using FEFFIT and

Supplementary Note 2. VSM Fitting
VSM data was fit using the Langevin function for superparamagnetic particles given by: where M is equal to the magnetization (emu g -1 ), Msat is equal to the saturation magntization (emu g -1 ), Ms is equal to the spontaneous magnetization determined for Fe using the magnetic moment of an Fe atom in metallic Fe (2.2 µB) and the volume of a BCC Fe unit cell (emu cm -3 ), V is equal to the volume of a particle (cm 3 ), H is equal to the applied magnetic field, kB is the Boltzmann constant (cm 2 g s -2 K -1 ), and T is the temperature (K). L(x) = coth(x) -1/x is the Langevin function. Average particle diameter was calculated by determining the value of V for each sample and assuming spherical Fe particles.

Supplementary Note 3. Characterization of catalysts before and after ORR operation
In order to investigate the possibility of the formation of new active species during ORR the as-prepared and H2-treated catalysts were investigated during and after operation. The electrochemical activity was unchanged during operation as observed by CV ( Supplementary   Fig. 25). XPS obtained after ORR operation exhibits a decrease in pyridinic N and an increase in pyrrolic N while the oxydic and graphitic N remain (Supplementary Fig. 26). This result again suggests that the pyridinic N species are not required for ORR. Vibrating sample magnetometry (VSM) performed before and after ORR operation shows a decrease in the magnetization due to the dissolution of unprotected surface species while the signature of small superparamagnetic particles is maintained (Supplementary Fig. 27). In order to mitigate any effects from putative dissolved Fe, both catalysts were run in electrolyte solution containing up to 100 mM of ZnClO4 and MnClO4. The excess of Zn and Mn, which are ORR inactive metals, should fill the vacant N sites and prevent any dissolved Fe from coordinating. Using CV, no effect on the ORR activity was observed with either Zn or Mn in the electrolyte (Supplementary Fig. 28). Together, the electrochemical tests along with XPS and VSM show that there were no new Fe or Fe-N species formed during operation and suggest that the Fe particles encapsulated by C and graphitic N are responsible for the observed activity and stability of NPM catalysts.