Atomic Fe in N-doped carbon (FeNC) electrocatalysts for oxygen (O 2 ) reduction at the cathode of proton exchange membrane fuel cells (PEMFCs) are the most promising alternative to platinum-group-metal catalysts. Despite recent progress on atomic FeNC O 2 reduction, their controlled synthesis and stability for practical applications remains challenging. A two-step synthesis approach has recently led to significant advances in terms of Fe-loading and mass activity; however, the Fe utilisation remains low owing to the difficulty of building scaffolds with sufficient porosity that electrochemically exposes the active sites. Herein, we addressed this issue by coordinating Fe in a highly porous nitrogen doped carbon support (~3295 m 2 g -1 ), prepared by pyrolysis of inexpensive 2,4,6triaminopyrimidine and a Mg 2+ salt active site template and porogen. Upon Fe coordination, a high electrochemical active site density of 2.54×10 19 sites g FeNC -1 and a record 52% FeN x electrochemical utilisation based on in situ nitrite stripping was achieved. The Fe single atoms are characterised pre-and post-electrochemical accelerated stress testing by aberration-corrected high-angle annular dark field scanning transmission electron microscopy, showing no Fe clustering. Moreover, ex situ X-ray absorption spectroscopy and low-temperature Mössbauer spectroscopy suggest the presence of penta-coordinated Fe sites, which were further studied by density functional theory calculations. catalysts [7,8] or introducing axial ligands. Different FeN x active site axial ligands have been recently proposed following in situ Mössbauer, x-ray absorption spectroscopy, nuclear inelastic scattering or electron paramagnetic resonance. [9,10] Some of them bearing close resemblance to biological systems, [11] such as N axially coordinated FeN 4 sites resembling heme. [12] However, spectroscopic discernibility is often challenging in these typically heterogeneous FeNC catalysts, [13] therefore experimental structure-activity correlations are hard to conclude. To overcome experimental limitations, the effect of O axial ligands on model FeNC systems has been calculated by density functional theory, [14,15] although the effect of other possible axial ligands on different Fe sites (pyridinic and pyrrolic) has not been fully considered. [16] Alternatively, to improve catalyst performance, the number of active sites can be increased, an approach which has shown significant progress in recent years. [17] To selectively form a high density of atomic Fe sites and avoid undesired Fe-induced carbothermal reduction, Fellinger and coworkers first identified that the high temperature pyrolytic step (800-1000ºC) should be decoupled from the Fe loading, by using a suitable N x site template. [17][18][19][20] The decoupled two-step synthetic approach to prepare FeNC O 2 reduction catalysts has led to remarkable progress; Mehmood et al. recently showed bulk FeN x site density (SD Mössbauer, , Eq. 1-2) up to 7.4×10 20 sites g FeNC -1 . The reported in situ nitrite strippin...