Although metal-free catalysts have been widely explored for ORR due to their low-cost and long-term stability, their ORR activity especially in acidic solution cannot meet the requirement of substituting for the commercial Pt/C. Metal-nitrogen-carbon (M-N-C) and metal carbides (M
3C) based carbon hybrids are the most attractive electrocatalysts for ORR especially in acidic electrolyte (
Table 2), where M represents for transition metal (Fe and Co) [
127-
148]. Since Jasinski firstly discovered that cobalt phthalocyanine was effective for the reduction of oxygen under alkaline condition in 1964, M-N
4 macrocycles (porphyrins, pathalocyanines, and tetraazaannulenes) were then developed as ORR electrocatalysts despite of their high cost, poor thermal stabilities, and poor conductivities [
149]. Series of M-N-C materials were subsequently prepared by pyrolyzing various transition metal, nitrogen, and carbon precursors in inert atmosphere. The O¼O bond is prone to be broken by the central metal, which greatly promotes the reduction of oxygen on the surface of M-N-C catalysts. In general, many researchers confirm that M-N
x sites in M-N-C catalysts are active centers for ORR [
150-
152]. M-N
4 units, as the basic structural blocks in metal macrocycles, are demon-strated as the active sites for ORR by Bouwkamp-Wijnoltz using in-situ Mossbauer spectroscopy [
49]. However, many researchers argue that transition central metal plays the role in the promotion of graphitization and formation of different type of nitrogen (pyridinic, pyrrolic, and graphitic N), instead of acting as the active sites for ORR. First-principles calculations prove that pyridinic and graphitic N act as the main active sites in N-doped carbon materials [
153]. Graphitic N determines the limited charge density and pyridinic N is favorable to enhance the onset potential of oxygen reduction [
154]. Despite many efforts have been made, the intrinsic nature of active sites for ORR in M-N-C species is still under debate.