I calculated the mass a third time, this time assuming nitrogen was molecular. This yields an atmospheric mass of 29.1 g/mol, of which CO₂ represents 604 ppm by mass, squaring with existing estimates. Then replacing 100 ppm of CO₂ with O₂ lowers the atmospheric mass by 1.2 mg/mol, or 4.1E-5, corresponding to a yearly added mass of 2.1 Gt.