High Vacuum With Mercury And Glassware

If you want to build your own vacuum tubes, whether amplifying, Nixie or cathode-ray, you’re going to need a vacuum. It’s in the name, after all. For a few thousand bucks, you can probably pick up a used turbo-molecular pump. But how did they make high vacuums back in the day? How did Edison evacuate his light bulbs?

Strangely enough, you could do worse than turn to YouTube for the answer: [Cody] demonstrates building a Sprengel vacuum pump (video embedded below). As tipster [BrightBlueJim] wrote us, this project has everything: high vacuum, home-made torch glassware, and large quantities of toxic heavy metals. (Somehow [Jim] missed out on the high-voltage from the static electricity generated by sliding mercury down glass tubes for days on end.)

The pump itself is very simple. Drops of mercury catch bubbles of air from the vessel to be evacuated and eventually all that’s left is vacuum, at least until the mercury starts boiling. [Cody] even experiments with cooling the mercury in liquid nitrogen to lower its vapor pressure and get an even better vacuum. It looks like it succeeded, but with his rudimentary measurement technique we can’t be sure.

Anyway, if you’ve got a few dollar’s worth of glass tubing, a few kilograms of mercury, more than a few hours to wait, and the muscles to lift the mercury up to the top of the tube, you can build yourself a vacuum pump that’s perfectly adequate for making light bulbs or Nixie tubes. [BrightBlueJim] suggests updating this project from mercury to something like gallium that’s less toxic but still really dense. We like that idea.

So if you’ve looked on in envy as others have made their own tubes, even going so far as to build their own complete amplifiers, you now have one fewer excuse.

32 thoughts on “High Vacuum With Mercury And Glassware

  1. For my PhD thesis we used mercury diffusion pumps. Nasty things, but diffusion pumps require high molecular weights and relatively low vapor pressure, and the only other real alternative at the time was oil. Nowadays there are pretty cheap magnetically-levitated turbopumps which would do the trick.

    Yeah, we couldn’t have anything organic in our vacuum system. Because tritium.

      1. I would have thought that water saturated with salt might work? Mercury is too dangerous, even the fumes. And spill management would be a complete nightmare.

        There is an old term “mad as a hatter” because hatters dipped hats into mercury with their hands for form the shape of the hat. Mercury doesn’t come back out of your system. Instead it lodges in your brain causing “madness” or similar conditions.

        1. Hatters used a mercury compound soluble in fats and water. There are plenty of things that are poison when soluble in fats and water. Before you panic, and give out biochemical advice, think hard on the section in this project about the vapor pressure of liquid mercury. Think about the fact that liquid mercury can be used to plug a hole in a vacuum. Think about how elemental mercury will ever make it into your tissues and what compounds it can become, and how.

          Standard cleanup is to use sulfur. Don’t do it! If you are worried, ventilate a little. Same goes for the scary lead-in-solder BS.

      2. Perhaps another alternative would be “liquid metal” such as gallium/indium(/tin&other) alloys, such as are now used in place of cpu thermal paste. (I found this post several years after it was published but felt that this info would be relevant)

        1. For any intrepid hackers with an inclination to try this, I’ll share a couple problems you’ll run into.

          First, Cody makes constructing the Sprengel itself look like a piece of cake. It is definitely not. I recommend using the thickest boro your budget can buy. Tee joints in particular are incredibly difficult. Access to oxypropane is almost a necessity imo. I have no idea how he did this, even with MAPP. LP on it’s own will cause you no end of trouble as the glass will never actually flow. After you actually create this fragile glass specimen, now you have to mount it somehow to keep it upright as it’s obviously gravity driven. Good luck. Be incredible careful with any sections of tube you may have thinned as they will snap readily.

          The biggest problem after you managed the prior incredible feats is figuring out how to combat the galinstan wetting. EuGaInSn (and presumably any Ga/In alloy (though I’d love to be wrong here)) wets practically everything. Apparently it doesn’t wet gallium oxide, but haven’t purchased any to test this and coating the inside of the pump will prove difficult without some kind of CVD process.

          1. As a scientific glassblower of over 40 years, the thing that most helped him was luck. His technique shows more bravado than skill. Also, he used soda-lime glass, also known as soft glass, and his “torch” was just hot enough to get things soft (but never juicy so no real “seal” was achieved. There were a lot of “sharpies” in his seals. Ironically, soda-lime glass can be more forgiving as long as there are no post-seal thermal stresses going on. Otherwise, “snap!!!!”

            My issue was his constant use of “perfect vacuum.” That term was OK in the 1670s when Guericke was working on his air pump. But even by the time that Sprengel was working on his pump, it was known and accepted that “perfect” was a bit strong a term.

            Good catch on the gallinstan wetting issue.

            Oh, thin glass, by itself, is not too much of an issue. It is when there are any flaws in the glass, such as if the glass is dragged across other glass, that can cause flaws which will significantly increase the chance of failure.

    1. There are inorganic diff pump fluids available. There are silicone oils and PFE oils.

      Not sure what you consider pretty cheap for a mag-lev. Even used they are usually several thousand for a little one.

  2. Tesla used mercury based vacuum pumps… Curious that he attributed the “blue glow” of his tubes as a natural characteristic of electricity… not realizing all his tubes has mercury vapour in them. (Which actually accounts for the blue glow.)

    1. Ok sitting in front of an actual keyboard. The dreaded blue glow. It is true ionized mercury vapor will emit a pale blue light, but what is not widely understood is that under electrical discharge of ionized virtually unlimited current can flow. There is a saturation current but for something the size of a vacuum tube it is well into the amps range. This is how thyratrons work. This discharge will emirate from an anode and reach towards the cathode. In a total discharge it will span the whole of the gap. This kind of discharge, of which neon lamps, voltage regulator tubes, and gas discharge tubes are of the same family were common place up until IGBT’s and such. There are also mercury arc rectifier type tubes that contain a mercury pool and are evacuated nearly to the vapor point of the mercury at room temperature. inert gasses used to be used to flush the tube before evacuation. An arc was struck to the pool which vaporized additional mercury and the current would flow from the hot spot on the pool to an anode. There was also a secret ingredient along with the mercury which is a secret I will take to my grave as promised to an old timer. These rectifiers would emit so much blue light and up into ultra violet that anything around them would get sunburned!

      All of the above mentioned glow discharges have one thing in common. Lots and lots of current and very low voltages across the discharges. Its why mercury and hydrogen thyratron rectifiers were so popular. if mercury vapor were present in sufficient quantity to discharge glow then no serious experiment could be performed other than verifying the breakdown voltage and current of the glow itself.

      The more common blue glow that people talk of is of a different source. It is usually found clinging to the glass envelope or apparatus. This is an electron bombardment discharge with high electron voltages rather than a direct ionization discharge of low voltage(Sometimes as low as a few volts) and high current to sustain the discharge. This kind of glow is generally caused by electron bombardment. The remaining gas molecules are excited and emit a glow. It is not a discharge and typically represents little to no ion current flow. Thats why tube people say don’t mind the glow. Sometimes its benign gas, sometimes it is not. This can also be caused by high voltage surface breakdown. This i believe was the most common cause of this kind of glow in old systems. Residual gas evolving from glass surfaces would be bombarded by electrons set en motion my the extreme high voltage and glow. This threshold of glow will carry depending on the voltage and residual gas pressure. Air at the lower extremes (Below 10 -4 torr, above that pressure it will turn more purple, then eventually up to pink) will discharge nearly the same blue as mercury at higher pressures.

      As far as mercury in lab vacuum systems its a story of good and bad. With proper cryo trapping very low orders of vacuum pressure can be obtained in an extremely clean way in the early days. What really did them in was their relatively low pumping speed compared to oil diffusion pumps. If care is not taken though the mercury vapor can back up into the vacuum chamber or tubes and the system will basically destroy itself. The irony is mercury will be poisonous too anything you are trying to put in a vacuum.

      Ok, I officially forgot why I started writing this reply.

      Some observations on the pump you made.

      From old models I have seen they make sure they made sure the droplet sealed to the walls before encountering the vacuum chamber inlet. The dropper was a bigger diameter orifice and the vacuum inlet had a smaller orifice to insure the droplet did not loose a seal therefore allowing bypassing of gas around the drop. I also recall it was important for the vacuum inlet to angle down and in like a medical IV. Also I believe your capillary tube was a little small. The surface tension of mercury is pretty big and will allow a larger tube. These pumps were actually reasonably fast when perfected. One common practice was to use one to rough evacuate a chamber and a charcoal pump under heat. This also prevented any mercury from condensing in the chamber. then a valve was closed to isolate the chamber and charcoal pumps. The charcoal was then brought down to liquid air temperatures for a few hours and the chamber heated. The resultant vacuum was extremely good. Most of the early radio transmitter tubes were made in this manner. Langmuir had some patents with GE that prevented people from using diffusion pumps in tube production in the late teens and early 20’s so other schemes like this were used. When oil diffusion pumps came along there was no reason to stay with this type other than for research. The rotary vane pump was so cheap and reliable that it was a no contest scenario. Still to this day you cannot find a more reliable device than a Welch duo seal pump.

      1. I was referring to the 502.5nM (blue) emission line, typical of a mercury vapor rectifier tube, which I have seen many times. (Beginning back in 1967.) As I recall their was also a gas discharge lamp similar to neon, but with Argon in it that had a very distinctive blue emission. (Had one once, but have no clue at all what happened to it… Too bad. Bet the steam punk fans would pay a fortune for it nowadays.)
        My comment about Tesla was motivated from my thoughts when I was reading about his mercury vacuum pump. Having seen a mercury vapor rectifier, I reached the stated conclusion. Using such a system WILL leave mercury atoms in any tube thus evacuated.

      2. I know you promised an old timer, but I’d hate for the knowledge to die. Might be it’s already known, and your old mate didn’t realise that. But just in case, how about you take someone into your confidence? Or at least write it down somewhere. It might turn out handy in a future warp drive, or something.

  3. If you don’t mind a little bit of mercury vapour in your tubes (and lungs) this is a relatively cheap way to do it … But I do highly recommend pumping down the chamber with a HVAC pump first so your not waiting for days with the big stuff

    Also keep in mind you can get eBay turbo molecular pumps fairly cheaply (compared to the price of pounds of mercury) but they are fragile beasts and take some metal work to take the fairly large opening in to something more usable … And I believe these need to be pumped down with a more conventional vacuum pump first to function properly or at all

    1. Yes, turbomolecular pumps need a backing pump to run at all…
      Yes, you can get them cheap on ebay, !BUT! these pump require fancy ceramic bearings (there’s usually a reason they’re selling them cheap, new ones are not cheap by any means), not all will let you get away with using cheap ceramic skateboard bearings. Vacuum grade oil is also a must, but not is such quantities as for diffusion pumps.
      Also, for best performance, you need a clean supply of purge gas, usually nitrogen.

    1. I keep thinking that when the reservoir runs out of mercury, and he refills it, there’s a slug of air in the line which gets blown out of the nozzle and probably contaminates the vacuum-to-be. A needle float valve to shut off the feed when the reservoir gets low, or an intern to refill it before it runs out, might significantly shorten the pump-down time by eliminating those setbacks.

  4. Full details to make all this in The Scientific American book of projects for the amateur scientist (1960). Very likely there is a copy in your library. Also a particle accelerator, cloud chamber, vacuum forepump from a refrigerator pump, (check the star twinkle suppressor), vortex tube, x-ray setup. In fact, an updated version of this book would be fabtastic, and please, not with public school type fun-killing safety. I studied this book endlessly in Jr. High and must have used ideas from it hundreds of times in my career.

    There are some places with a PDF http://www.sciencemadness.org/library/books/projects_for_the_amateur_scientist.pdf

  5. These are called diffusion pumps. Very comon in Neon processing setups.
    You need a regular two stage mechanical pump to get a starting vacuum – then the diffusion pump can start working. No good at atmospheric pressure – will nto work.
    They use silicon oil now – dense and a lot safer than mercury – which is a nightmare in this scenario – all the old (well mostly dead now) Neon bombarder operators became bipolar – mad as hatters (look it up) most of em.
    For silicon oil use 704 or 702. 702 better at recovering from exposure to atmos pressure. 704 gets a better vacuum but tends to solidy and become useless when exposed to air while boiling. Check out Neon plants for gear.

    1. Not a diffusion pump. It is a mechanical pump. Diffusion pumps use mercury or oil vapor moving at high speed away from the vacuum chamber. The vapor bashed air molecule towards the fore pump, much like a turbo-molecular pump.

  6. Its also possible to use oil or mercury in a true diffusion pump where the medium (oil, mercury or ?) is heated and rises up to the top of a column or large diameter vessel that is also cooled by water or other refrigerant at the top portion the vapor of the medium then gets more dense and begins to drop down the cooled sides of the vessel pulling air molecules down with it, there are also baffles that are similar looking to the innards of a old coffee percolator to increase the effectiveness of the pump, more surface area .. etc. All diffusion pumps require a backing pump (also can be used for the roughing pump, which is a mechanical pump that removes the bulk of the air and gets the vacuum closer to “high vac” the diffusion pump or even turbo pumps that let us get into the really interesting area of ultra high vacuum are comparable to a super charger in a car except in reverse, the backing or roughing pump get real inefficient at higher vacuum states, they don’t have the ability to reach High vacuum, however turbo pumps and diffusion pumps cant operate at any pressure above a few 10’s of millibar and really are best when pressure is less than 3 or 4 millibar. Oil in a oil diffusion pump would burn if allowed to have any real air pressure at its inlet the blades of a turbo pump would deform and tear destroying it if operated in anything but a very good rough vacuum. so the diffusion pump or turbo “concentrates” the air molecules and feeds this to the roughing pump. I don’t know but the tiny gain from mercury over an oil diffusion pump isn’t worth the risk, hot mercury and its vapor in glass with a gate valve between it and me oil if it burns or escapes stinks but wont poison everyone in the building !! plus a proper sized diffusion pump with a decent backing/roughing pump will hit as high vacuum as any project may require. 1000 % more safe , plus EPA will cease demolish and send your lab home garage or other mercury spill site to a toxic dump in a split second. NO,,? Google Nicor gas company Chicago meter replacement mercury spill gas meters contain mercury, company spilled it during a meter upgrade , inside of several homes not cool, mercury spilled at any one home less than a teaspoon! check it out un real! leave mercury pumps alone stick to oil diffusion pumps, no other way to say it build a mercury pump and it will cause problems, some minor most very dangerous, possibly criminal, all for no reason! Leave mercury pumps as historical note!!! PLEASE!!

    1. I think you are talking about EPA over-step and job security more than mercury. There are plenty of ways to clean up mercury, and plenty of ways to just leave it alone and ventilate.

    2. Mercury hysteria. A PhD repeating a load of sh*t: https://www.thoughtco.com/can-you-touch-mercury-608434 Says mercury has “high vapor pressure” and moves instantly through your skin and effects your mental state so get medical aid immediately if you touch it, etc. etc. It is super serial!

      Nice rational and detailed discussion: http://www.heracliteanriver.com/?p=246

      Note, I would play with mercury in my hands and wash afterwards without worry. If a fluorescent lamp breaks I ventilate and make sure the room air is not circulated throughout the building. It is the exposed surface area of the mercury that matters.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.