On the validity of the formation of crystalline carbon nitrides, C3N4