The question was inspired by a comment on a site I visit a lot, one to which I've posted a lot in nigh on three years. More about that later. Suffice it to say that I shall reply to here, rather than there, and attempt to place a link on the other site. There will be no more posting to that site until such time as its moderation policies are given a thorough overhaul! Nuff said for now...
The question was essentially this: given that we all know that petrol fumes sink to the ground at a filling station, why doesn't CO2 - which we also know is denser than air - also settle at ground level? Why are we not suffocated by the stuff - or does it only come up to ankle or knee level?
Imagine one were to trap gases inside balloons - one for hydrogen, one for oxygen, one for nitrogen, one for carbon dioxide - and then release them. The four balloons would behave exactly as the questioner supposes. The hydrogen balloon would quickly ascend, the CO2 balloon would rapidly descend, and the nitrogen and oxygen balloons would probably hover or sink slowly - due mainly to the weight of the balloon rubber - not the contents. The relative densities of hydrogen :nitrogen: oxygen: air::carbon dioxide are approximately 1 : 7 : 8 : 7.2 : 22. Gases lighter than air rise, those heavier than air sink. No surprises there.
When petrol fumes are released, they too sink quickly, at least to start with. A typical molecule in petrol is one of the isomeric octanes, general formula C8H18, with a relative vapour density of 57 - some 4 times greater than air.
But the petrol fumes would not stay for very long at ground level. Convection current carry them upwards, and gaseous diffusion would cause mixing with air even without convection. That's because gas molecules are in a state of constant motion, colliding with other molecules, millions of times a second, causing them gradually to diffuse ("spread") in all directions. The fumes gradually spread into all the space available - which could be a jar, a garage, a hangar, the entire atmosphere. Once the space is evenly occupied, the molecules then show no tendency to unmix. Why not? Answer: because the 1g force that acts on all molecules in air at sea level is insufficient to overcome the kinetic forces due to collision between molecules. Put more simply - a molecule that gets a strong bump from below will be knocked upwards, against the weaker force of gravity.
This is true for g=1, but is not true for progressively higher g forces.
Here's an example - always a a controversial one. Enrichment of the fissile uranium isotope U-235 needed for atomic power stations OR Hiroshima-type A bombs, requires separation from the more abundant U-238. This can be accomplished in gaseous diffusion plants, or in centrifuges that generate an intense g force. Either process requires that solid metallic uranium first be converted to the gaseous uranium hexafluoride (UF6).
There is a well known experiment that is done in schools, at any rate, those that still have a fume cupboard, to demonstrate that dense gases and/or vapours gradually diffuse to fill the space available, and then do not subsequently unmix from air.
One places of few drops of elemental bromine, Br2, a fuming red liquid in the lower jar, which is separated from the upper jar by a glass divider. One waits for the lower jar to fill completely with red-brown fumes. One then removes the separator. The fumes gradually fill both jars evenly, despite bromine vapour being 5 times denser than air.
There's a variant on the experiment that I devised while teaching to demonstrate the petrol vapour effect. One places a jar of bromine on top, and then removes the divider. Most of the bromine fumes sink immediately into the lower jar, behaving as if they were enclosed in a balloon. But the fumes then gradually diffuse back upwards to produce the same end-result as before.
The short term behaviour of the petrol fumes is called a bulk phase effect. It's the temporary behaviour of heavy molecules in close proximity, which behave briefly as if enclosed in a balloon. But once diffusion has caused mixing of heavy molecules with the lighter molecules of nitrogen and oxygen, unmixing does not occur at normal values of g.