A glass pump with mercury would sure look cool though.

Applied Science: Oh yeah, and I made a CRT out of some things I had lying around, NBD.

When I was in college one of my profs was telling stories about his vacuum experiences. One; he had carefully set the system up, using all the lessons learned earlier (e.g. new copper gaskets) but it would not pump down. Tore the rig apart and found a fly in one of the tubes. Turns out flies out gas rather well. :^)

I've worked with the large diffusion pump system for years and understood the theory, but THIS is what I've always wanted to see. Thank you this is awesome.

"Here's a cool video about my glass vaccuum pump" "oh don't mind this homemade crt, I was just using it because this didn't work for my mass spectrometer"

Glass diffusion pumps are still commonly used for neon sign work and were also in quite common use for producing one off vacuum tubes and the like. The advantage is in the ability to connect to an all glass manifold for gas back filling or to the pumped item directly with glassworking techniques rather than metal working techniques. As a side note DC-704 and other silicone based diff pump oils cause great havoc when exposed to atmosphere while hot resulting in an horrid black tar which cannot be removed except by aggressive mechanical means. The much safer (read: expensive) option is a Polyphenyl Ether based oil like Santovac-5 which contains no silicon.

As you explained in the video, the silicon oil absorbs unwanted substances, so the mechanical pump is always needed to bring the vacuum down, draining out such substances, as you did, and then getting out the residual gas. It also helps to drain down the oil vapour. If you turned the mechanical pump off, you'd contaminate your vacuum chamber, even if you cool the top of the diffusion pump with liquid nitrogen. Therefore you must have an automatic valve to shut out the connection with the chamber if electricity goes out. Actually they are a pain in the neck to use, so people prefer turbo-pumps which are far cleaner, if they can afford them, that is. By the way, great video! I really appreciate your resourcefulness.

Used Diff pumps a lot, no how they work, seen a hundred diagrams and only your channel showed me one running....cheers!

I applaud the glass maker that did this

1:55 You can not unsee those eyes on the vakuum pump, once you have seen them.

Hi, I worked for a company, before I retired, that made photomultipliers and they used mercury/glass diffusion pumps for many years. They also made some of the first television camera tubes and I think these were pumped down using mercury/glass diffusion pumps too. They had a glass blowing department where new glassware was made when required. Very interesting video, thanks very much.

While it might not be the most practical diffusion pump ever, that pump is still a thing of beauty.

Hello. I am a SEM operator. I was just studying up on SEM theory, particularly the pumping systems, turbomolecular pumps and oil diffusion pumps. I couldnt find a really good demo or video to see exactly the flow paths. Your video that is brand new, perfect timing, it really answered a lot of question of the actual flow paths for me. Thanks a million. I love your approach, discussed in the beginning of your behind the scenes video. I also learned most of my engineering skills on my own in life, garage, fixing and building things, not in the university. I am very grateful for you wonderful sharing. Raanan

New videos from this channel work better than SSRIs at alleviating depression.

Great video as always! Few notes though, as you have shown it, forvacuum mechanical pump could still reach absolute vacuum with infinite time, which is not true for real pumps. Real mechanical pumps have a hard limit on the pressure they can achieve (even given INFINITE time!), largest limiting factor being "dead space" of the pump(turbopumps excluded). Dead space is a volume of gas at exhaust pressure(1 atm for forvacuum pumps exhausting directly outside) which is left when exhaust valve is closed. If you think about plunger/piston type pump, it is the volume which is left when cylinder is compressed completely. Air in dead space at atmospheric pressure then is expanded to full cylinder volume, decreasing the pressure until intake valve opens, which limits the minimal pressure one can achieve. You don't seem to have a return path for electrons in your crt, you are charging the screen which screws with focus and makes it fuzzy. Also you could try to deposit ZnS as a classic green luminophore. Edit: fixed grammar to make it feel less like I had a stroke. Merged another comment.

I’ve been trying for a while to wrap my mind around how and why diffusion pumps work. Not quite there yet, but thank you for increasing the clarity a lot!

I like how I can sit down after working all day with work stuff running through my head and instantly be engrossed by one of your videos, and having just enough knowledge to understand what you are very eloquently explaining. I think it has more to do with the fact that you make it very easy to follow along even if someone has no clue what a diffusion pump is.

"Well, the reason I built the Hadron Collider in the basement was..." Answer in next weeks Applied Science video.

I'd think nichrome should work well as an electron emitter at dull red heat since chromium oxide is an excellent source of electrons, I'd heat it in air one time to get some oxide.

At work we use liquid ring pumps(very clever pump IMO), dry screw pumps, roots blowers, and steam jets(venturi) to pull high vacuum for distillation processes. We use this vacuum post reaction and during distillation of various chemicals.

LaB6 emitters work well when exposure to air will occur. In fact, my picture is the LaB6 cathode in my plasma chamber!

CRT's are super cool. We need to get you into a glass blowing class so you can make a suitable neck, and play around more with your glass experiments. Awesome video as always, thanks!

I first encountered oil diffusion pumps in graduate school. Our Beckman ultracentrifuge had a standard reciprocating piston pump to get most of the air out of the chamber. Then you switched on the oil diffusion pump to get a very high vacuum. I'd never heard of such a thing before and was fascinated by how they function with zero moving parts. This was in the early 80s, no internet or Google back then so I had a bit of a job finding the information.

Some serious work went into making that apparatus. Great video!

Love your channel! You are heating the diff pump oil too much, you do not want it to boil. The correct temp is just where there is slight movement on the surface of the oil and the temp varies with the type of oil used. 702 = 180-190C 704 = 220-230C Yes never ever use plastic tubing in vacuum systems - thick wall latex is acceptable. Me? I used to be a high vacuum service tech.

I worked where they mirrored glass on a machine that had a mercury diffusion pump, That thing could pull a vacuum in a 20 cubic foot chamber in like 5 seconds, then the aluminum filaments would start up and deposit onto the glass in about 5 seconds. It was very cool.

I haven't used one of these since the 60s!

One of my many "One day i want to..." projects is an open source CRT display.

Wow! Great video! I worked for a vacuum furnace company. They would have two guys clean the inside of the entire furnace with alcohol prior to the first run up. This minimized the outgassing but took the better part of a day. It would take quite a while to pump down(rough out) the first time. One time the valve controlling the diffusion pump vacuum flow was installed incorrectly. The roughing pump pulled the oil out of the diffusion pump coating the inside of the furnace. If I remember correctly, it took several days to clean it all out.

this is my preferred KZfaq channel hands down.

Dang that's a nice... diffusion pump.

Thank you. Worked as a mass spectroscopist using GCMS, He leak detectors, and isotopic ratio mass spectrometers. System pressures were 10 -8 microns or lower with a good vacuum. Use dry nitrogen or argon to let the system up to atmospheric pressure. This reduces pumping time. Water from water vapor in the air is an enemy as far a getting a good vacuum. You might get some vacuum gauges. A 1946 tube with a meter or a Pirani gauge is good for rough vacuums down to about 1 micron. For higher vaccums, use a hot cathode or a cold cathode vacuum gauge and controller.

Great explanation. Whoever invented this... wow. "We could push those last few molecules out with vapor molecules we boil and move around right inside the chamber!" I bet the rest of the people standing around looked at them like they were nuts.

Amazing. The amount of effort that goes into your videos is truly inspiring.

Very nice video and beautiful pump! Gives me flashback to my PhD there we used a 50 liter diffusion pump to vacuum a huge high voltage discharge system. The stove was firmly attached to its bottom and the running tap water was used for cooling. The engineers told me that the hot oil would burn if no pre-vacuum created prior to heating it up.

Man, that vacume pump demonstration with the beads is awesome. I never understood why it was so hard to "get it all" I do know. Thanks.

Try using the filament assembly from a miniature T5 fluorescent tube - it’s pre-coated with emitter!

I worked with lots of cryos and turbos and lots of mechanical pumps even with Roots blowers boosting them but I never worked with diffusion pumps. Very interesting.

Very nice explanation. One of the hard limits of a diffusion pump is the back flow from desorption of the pump fluid. I've only worked with mercury pumps preferring a Balzer any day

Nice of us to show us your massive bong, Ben!

Interestingly enough I was reading into DIY CRTs just this week and got interested in it again. Now this video. Can't be a coincidence!

It’s amazing what percolation type stuff can do

I am in need of a vacuum wood kiln. I was just reviewing Sprengel pumps and other less conventional pumps. Like you said the volume that needs to be evacuated makes these less practical. Especially when time is money. Awesome video as always. You are a fascinating dude.

Wish I could have seen this made :) love functional glass

"Obviously, boiling mercury is not so great."

The diffusion pump by itself is amazing. But the CRT demo is awesome. That has to be the most simple vacuum tube i have ever seen. I had never thought of making one using more standard laboratory glassware.

Imagine how awesome it would be if your neighbor's garage had this going on! Never change Applied Science, never change!

Waking up to this with a cup of green tea is delightful. Thank you for the explanation. Lab science has always been my favorite.

Great video as always! I didn't know that diffusion pumps came in glass! A lot of systems in the semi industry have moved away from diffusion pumps in favor of cryo or turbo pumps for achieving high vacuum in sensitive systems where any backstreaming(what you talk about around 5:37 ) could be catastrophic. If anyone wants to learn some more about vacuum systems "A User's Guide to Vacuum Technology" by John F. O'Hanlon is, from my understanding, the industry standard for vacuum systems. Certainly at a minimum quite a good resource. It was one of the required textbooks for my thin films class.

Worked with diff pumps for a long time. Cool to see one like this. Btw, you should have insulation near the heat source to prevent rapid cooling at the bottom. Also, with the plastic.. helium can penetrate some with no issue and it's even used to outgas helium (used as a carrier gas) from thicker material by pushing through plastic or PTFE loops.

Fantastic! Thanks for showing. I used Diff pumps for almost ten years in EM but of course even servicing them you can't see how they work. Great idea to show everyone!

That's some high-test screwing around right there.

Wow! This is really cool! As you can see from my KZfaq channel, I've been using diffusion pumps for many years, but never had the opportunity to "see" them in action! Great video! Thanks!

Everything about this video is insightful, amazing and fascinating at the same time. Very well explained too. Thanks!

These later episodes of Dr. Stone are absolutely lit!

As a note; Modem electron microscopes use lanthanum hexaboride emitters, and they have isolation valves such that the source never gets exposed to air. A separate vacuum backing for the source keeps it at very low vacuum (10^8 or better) and there are interlocks for vacuum level that must be closed before the sample chamber is opened to the source

Very cool video! I work at a high vacuum company, we use diffusion pumps large enough to fit inside. This is my first time actually seeing one operate. Its a concept quite difficult to understand with out a visual, I will be showing this to future recruits. I also liked your explanation of molecular flow in everyday terms. I'd like to know what type of oil you used. And yes, too bad that particular pump was made with such poor conductance. I'm not sure how difficult it would be to home build and Ion Pump? That would be very cool to see, and help you achieve UHV. Keep experimenting!

Diffusion pumps are really interesting pieces of equipment. The downside with them is that no matter what one does, they will always out gas oil into the chamber, due to the vapor pressure of the oil. Though, for a lot of stuff, that isn't a major issue.

This was like "Applied Science meets Glasslinger" ..... I approve!

Great video again Ben. Given the complexity of pump I think one has to be a master glass blower to make one. Very impressive!

Perhaps thE very best channel "out there" IMHO !

thank you for educating people on this channel. I look for next video!

Really appreciate the output & knowledge...👍

Thank you for this amazing video. I've always wanted to actually see what's going on in my metal diffusion pump. Now I know! The CRT is an astonishing achievement too, well done there. Barium carbonate is converted to barium oxide upon heating, and barium oxide would combine with with moisture from the atmosphere to form barium hydroxide so as you say quite rightly it cannot be subjected to the air once applied and heated.

Absolutly fascinating. Never thought about the gas molecules behaving differently at that pressure. The whole opaque thing makes sense now.

The most beautiful kind of vacuum pumps.

For some reason I found the tungsten trace quite lovely indeed! Bravo!

Being former electron microscope serviceman, you did make my day with this video.

I might have to build one of these so thanks for all the useful information!

Electron microscopes tend to use either pure tungsten or a LaB6 (lanthium hexaboride) tip if they are purely thermally driven. This means they are easily user replaced in the lab and we dont have to handle sealed units when replacing the tip. Using your material for extra brightness would caus problems with that. Extra brightness is always nice but honestly, we burn through tips every few years, so replaceable is key. For extra brightness we tend to use a FEG (field emission gun) which draws electrons out using electric fields (with or without additional heating). Really interesting video btw, having used oil diffusion pumps for years its nice to watch it working.

Today on applied science Im going to cover something that would take you a lifetime to figure out in a 15 minute video

Really cool! I think this is the first video that illustrates how diff pumps work that well on youtube. I actually seen one of those pumps in a store in Shanghai but didn't really think it can work (that well), given how small that inlet is.

Ben: if MacGuyver were real, he would have a statue of you that he prayed to every night. You are amazing!

Wanted to see inside an operating diffusion pump since I learned about the theory many years ago. Thanks for sharing!

Awesome glory of finest 19th century tech and someone who knows how to use it!

Thanks for this video. It was really cool!

Thank you Ben!

Absolutely fascinating

Instant click when i see Applied Science. Definitely learned something too, as always

The oil & chemical industries uses steam ejectors - those work off of the momentum transfer method too. Like the bottom of your diff. pump ejectors can be staged and the final pressure (vacuum) is determined by the overall compression ratio. I worked with those in the 90's and the roar of the steam was deafening.

Nice bong dude

Thanks for showing the 2 failures at the end.

your videos are awesome! Thanks for making them!

Now THIS is exhilarating!

Hey Ben! Your videos are so amazing.

Maybe try heating the filament slowly and to a lower peek current. I suspect the barium oxide is acting as an insulator allowing the tungsten to reach too high a temperature. Also, the barium carbonate could act as an oxidizer with respect to tungsten at high temperatures so best to decompose it slowly at a lower temperature.

Exceptionally well explained. Thank You so very much for your time and effort !!!

Never knew you needed that deep of a vacuum to get a cathode tube to work very interesting indeed 👌

“Bong? No officer that’s uhh my giant glass diffusion pump”

this has to be one of the craziest youtube channels ever...

Amazed as always

That's a majorly impressive hunk of glass. 😯

the reason why they don't use diffusion pumps as a single stage is because the pressure difference that they create is pretty low. so this can (very roughly speaking) either get your pressure from 1 bar to 0.99 bar or from 0.01 bar to almost 0, that's why it makes sense to use a mechanical pump as the first stage to get you to that 0.01 in the first place.

About your claim that you don't technically need the roughing pump (at ~5:10). My experience is if the outlet of the diffusion pump is above a certain pressure, the oil in the diffusion pump increasess in volume (due to dissolved gas) and spills over into your system. The excess air pressure can also burn (oxidize) the oil making it char. This is a problem especially when the power goes out -- if it comes back on unattended, your roughing pump doesn't pump down fast enough, and the pressure is too high when the oil starts moving again. Just nitpicking -- your videos are so cool!

Anyone who can build a CRT using a flask, coils of wire, glow-in-the-dark paint and a broken lightbulb is a science hero...

5:10 It's not just an efficiency problem. High vacuum pumps as diffusion and turbo pumps can not work at athmospheric pressure. And mechanical pumps (rotary vane, membrane etc.) can't reach high vacuum. But you're still right. By changing the liquid in the diffusion pump and the speed of a turbo molecular pump, you could stage multiple pumps to achieve a high vacuum (for example there are simple vacuum pumps you just connect to a fast flow of water. The high water flow creates a low pressure zone. But not the same principle as diffusion pumps. Bernoullis law instead of air molecules diffusing into the pumping liquid which drags them down). But that wouldn't be effective. Mechanical pumps are way more efficient at those pressures. As you explained, they are like syringes. They connect to an area, expand to decrease the pressure by expanding the area, then disconnect and expell the trapped gases to the exhaust. Diffusion pumps use a liquid to diffuse small amounts of gas and move it downwards, where it gets reheated to expell the gases and evaporate again. Turbo molecular pumps use fan blades at extremely high rpm to hit them into a direction away from the evacuated side. So basically at high pressure it's more effective to move gas by trapoing and releasing in a chamber. At low pressure you'll need pumps that physically move individual gas molecules. And i absolutely love your diffusion pump! Absolutely amazing to see it work through glass.

That diy CRT is dang neat!

Always amazing

I celebrate fixing my TV. Ben just makes his own.....

Bro making the craziest bongs ever