To understand diffusion vs perfusion limited you have to know the equation for diffusion (Fick's). J = A*D(P1-P2)/x
where a = cross sectional area, D = diffusion coefficient (CO2 diffuses 20x more than O2, CO diffuses 200x more than O2; D is the diffusion constant and is proportional to solubility and inversely proportional to MW), P1-P2 is the difference in partial pressures (explained by KDBurton), and x is the thickness (can increase in ARDs, CHF, or pulmonary fibrosis, etc)
The easiest way to think about it is Nitrous Oxide versus Carbon Monoxide.
Nitrous oxide does not bind hemoglobin at all and is totally dependent on it's solubility in the blood. It reaches it's peak concentration in the blood very quickly (diffuses quickly until it reaches saturation) and thus, because it finishes it's diffusion quickly, it is NOT diffusion limited, it is PERFUSION limited. This is explained in Ficks by the P1-P2 term (P1 quickly equals P2, and diffusion quickly stops)
CO on the other hand binds Hb avidly and is removed from the blood, meaning the partial pressure of CO in the blood is almost always 0. This means that P1-P2 is almost never 0, meaning that diffusion will never reach 0. This means that it is almost entirely depends on diffusion and is DIFFUSION limited.
Oxygen is somewhere in between.
It both binds Hb (duh) and is soluble in the blood. There are 2 situations, low cardiac output, and high cardiac output. At low cardiac output, there is more time for blood to have contact with the alveoli. Since there is more time, more diffusion takes place and the blood will saturate with oxygen. This sounds like nitrous oxide right? P1-P2 eventually becomes 0, no more diffusion occurs, it's not diffusion limited...therefore it's PERFUSION limited at low cardiac output.
At strenous exercise, cardiac output increases, and now the blood is shooting through the capillary bed quickly. The oxygen does not have as much time to saturate the blood...thus it's dependent on diffusion (P1-P2 never reaches 0) and is DIFFUSION limited.