...
Remember: molecules are tiny, and there's a lot of them in the air around you—there are more molecules within a literal hair's width of a pinhead your nose than one you could possibly count in a lifetime. And what are they doing? Bouncing into each other. Knock enough molecules in one direction and those molecules will knock into the next like dominoes, and that's pretty much how sound happens:
...
And the domino chain that is the sound wave moves through the air so fast that it looks like all the flames went out at once, but the flames get extinguished one after the next, at the speed of sound, but and the speed of sound is simply too fast to see. So the sound has direction, but the air doesn't flow. This lack of airflow is unlike blowing out a flame, where the air itself moves in a particular direction.
...
Suction pumps create flow in just the opposite way: remove some molecules and the ones next to it fall into the empty space left behind and, just like the dominoes, that next-door molecule leaves behind it's own void that gets filled by other molecules, and so on and so forth.
Both blowing and sucking create the domino-chain-like behavior known as viscous flow, and it's what we're used to and expect because we live out our entire lives in this gas we call 'air', always at atmospheric pressure. When viscous flow is slow it can be smooth (laminar) like the smoke from the flame, and when the flow gets fast it gets rough (turbulent), like the blown smoke. But what happens when the pressure gets so low that molecules hardly run into each other? No more domino chains, no more viscous flow: Molecular Flow.
...
Imagine the world's largest air hockey table, you place a few pucks on the table, and hit them in random directions. They'll hit each other sometimes, but they're much more likely to hit the wallwalls. And it's just as likely to bounce back into your own goal as it is to bounce into the opponent's goal. Now replace the flat of the table with the space inside a vacuum chamber, and replace the pucks with molecules; this is molecular flow: when a molecule is more likely to hit a chamber wall than it is to hit another molecule, and there is no general directionality direction to the flow (see figure the flow. check out the molecular flow figure (above the air hockey table), some arrows point forwards and some backwards), unlike viscous flow where all the arrows point generally forwards.
Pumping in a molecular flow
For a pump to create suction, it needs enough molecules around for the domino-chain behavior of a viscous flow, so in molecular flow a pump cannot pull air. But if that pump Can't suck, what does it it do? It traps. Like the goal on an air hockey table traps the puck when it flies in. But there's nobody knocking molecules towards the pump like there would be in air hockey, It's alllll random.!
Well if a molecule getting trapped behind a pump is random, How likely is a molecule to randomly fly into a pump that traps? That depends on how big the tube is leading to the pump.
...
Sadly adding a bigger pump to suck faster though the tiny tube only works in viscous flow. In molecular flow, the pump will only ever remove molecules at the rate they naturally fly through the smallest/longest tube in leading to the pump. Putting a bigger pump on the same tube won't pump any faster. The lowest conductance point sets the pace. =(
...
Ok, so you know how I said molecules bounce off walls? Not actually true! 😬 I lied to keep it simple, but I think you're ready for the truth: It turns out that any time a gas molecule hits a surface it actually sticks to that surface, just like how your breath fogs the mirror; this called adsorption (not to be confused with absorption).
And also just like fog on a mirror vanishes over time, those stuck molecules will eventually jump back off of the surface (Ii.e. evaporate, outgas, desorb), flying of in a random direction just as fast as it was when it hit the surface.
...
How long a molecule stays stuck before desorbing from a surface that depends on the type of molecule, the temperature, and the what the surface is made of. Evaporation of a molecule can happen anytime any time between almost instantly and basically never. And we'll never be a able to predict the moment exactly, but we usually have a general idea for how long, based on things like temperature, pressure, and material.
When a molecule is flying around a vacuum chamber, it's contributing to the total pressure, but a molecule is trapped on the surface doesn't contribute to pressure because it's not a gas, it's trapped on the that surface! So if a type molecule tends to jump off surfaces rarely, the pressure can stay low. And also if a certain type molecule tends to jump off quickly (from a given surface at a given temperature) it will hop around enough to quickly find its way into a pump, again the pressure drops.
The trouble for lowering pressure comes when a given type molecule jumps around enough to raise the pressure, but not enough to easily find it's way into the pump. Check out the Application of Vacuum Theory page to see how we deal with troublesome molecules.