Sciencemadness Discussion Board

Making Liquid Nitrogen

M4D1NV3N70R - 19-8-2005 at 17:04

I'm wondering whether I could make my own liquid nitrogen using things available around the house, junkyard, or home improvement center. I'm on an ultra tight budget - $20 max, so scrounging everything I can is going to be a requirement for me.

Compressing air to 3000 psi seems awefully expensive, and potentially dangerous (I've heard of pipe fittings going through skulls). Not only that, It would seem to require fractional distillation after liquifying the air, which would really make it expensive.

Cascading refridgeration sounds more economical and less dangerous, but I can't find many details online (maybe I'm just looking in the wrong places?), or even how much LN2 it generates.

Has anyone out there made their own LN2? How did you do it?

saps - 19-8-2005 at 17:16

When you compress some air, you're putting energy in and it heats up. You can get it back to room temperature just by running it through some coils which aren't special, they just have a lot of surface area to let heat flow out to the room. Now if you let that compressed gas expand again, it does work, loses energy, and gets cold.
You can chain a few stages like that together. For example, the compressed air in the second stage can run through coild immersed in the cold air from the first stage. It then starts off cold and gets even colder when it expands. But I dont believe that 3000 is enought presure. and you will NEVER be able to produce LN2 with just 20 dollars.

M4D1NV3N70R - 19-8-2005 at 17:46

Originally posted by saps
When you compress some air, you're putting energy in and it heats up. You can get it back to room temperature just by running it through some coils which aren't special, they just have a lot of surface area to let heat flow out to the room. Now if you let that compressed gas expand again, it does work, loses energy, and gets cold.
You can chain a few stages like that together. For example, the compressed air in the second stage can run through coild immersed in the cold air from the first stage. It then starts off cold and gets even colder when it expands.
Yes, I'm familiar with the basic principles, they are described quite well in the Refridgeration Technology thread on this forum. What I'm asking about are dirt cheap inexpensive setups or even just ideas that others may have come up with.
Originally posted by saps
But I dont believe that 3000 is enought presure. and you will NEVER be able to produce LN2 with just 20 dollars.
Well, $20 is the max I can spend in addition to scrounging, elbow grease and ingenuity.

I was thinking maybe old air conditioners from the junk yard, jury rigged into a daisy chain using some piping from the home improvement center, and possibly replacing the coolants with something more effective, building up to using methyl chloride in the unit just before the liquification stage.

But that cascading setup still leaves me with a need for fractional distillation to get pure LN2.:(

Fleaker - 19-8-2005 at 18:16

What's wrong with purchasing some? It's not too expensive, and many places will loan you a dewar with a deposit on it. I'm lucky enough to get free liquified gases (there's a cascade cooling plant right next door to the store that sells it)

Should be readily obtainable if you have any large welding store or medical supply.

M4D1NV3N70R - 19-8-2005 at 18:37

Originally posted by Fleaker
What's wrong with purchasing some?

Because once it's gone it's gone. And I need an ongoing supply of reasonable quantities.

BromicAcid - 19-8-2005 at 20:33

Refrigeration Technology

A thread with a similar angle to this one, the main purpose of that thread instead to make liquid air. Explosivo posted a nice picture of a liquid air producing setup and there is some relevent discussion.

I invested some time into attempting some cascade cooling. Finding air conditioners that still work in peoples garbage was an easy task and I collected four of them in two weeks. Simple tests involving measuring the temperature in the cooling coils with a thermometer of two different air conditioners then putting the two together (cold side against hot side) did result in a lower temperature then either could acheive on their own. The problem with cascade cooling as I saw it though was that the coolant in the systems would not have worked for the low temperature employed, Marvin states that methyl chloride could be substitued into one of the systems and this used to cascade ethylene which could acheive temperatures to condense air.

But how to change the coolant in these two air conditioners? Both methyl chloride and ethylene were employed at some time for these things, as a matter of fact I managed to find a local shop that serived industrial machines that said they could put ethylene in an air conditioner (though they didn't say they would do it for me).

I really don't see a way to get just the liquid nitrogen without first taking out everything else from the air, Organikum made a post some time ago on a setup to produce nitrogen from the air free of oxygen, maybe you should search for that for ideas, I believe that it used the passage of air through hot copper turnings to acheive the removal of the oxygen.

evil_lurker - 19-8-2005 at 23:32

I'd say get the biggest damn thermos you can find for $20 and go to a welding shop.

A good thermos will keep Liquid N for 2-3 days.

Alternatively, a semen storage container would work too, but at a higher cost.

I can tell you right now that there practically is no way in hell your gonna scrounge up the equipment to freeze air down to its condensation temp.

Sadly this is the way it is with many chemical processes, its simply impossible/impractical to do it on a home scale.

Fleaker - 20-8-2005 at 14:45

A good quality dewar will hold it for weeks on end. I know that a 35L dewar takes about a month to go completely dry, but if you need a constant supply, then you'd best save up :-\

ChemicalBlackArts - 20-8-2005 at 23:20

Or you could just go to a welding shop and buy it for 3 dollars per gallon. Just make sure you have a solid nalgene container to store it in.

Fleaker - 21-8-2005 at 10:52

3 dollars per gallon is a pretty good price, I called a large national distributor and they wanted 30 USD for 4 liters and they wanted a 70 dollar deposit on the dewar. Needless to say, I found another supplier.

It's a shame that those Nalgene 'dewars' only hold LN2 for maybe 36 hours depending on their size. Good enough for a few distillations or low temperature cryo research though :-)

hdcwr0x2 - 9-10-2005 at 20:40

I've done quite a bit of research into this path myself. I was in the process of building a phase change cooler for my computer using propane for a refrigerant (boils at -40something C) and I have all the parts I need although I'm a little lazy at the moment. Anyway, if you need more information, I'd reccomend looking at these two sites.

Also, if you search google there is a diagram of an enormous 3 stage cascade used by phillips to produce LN2. It uses propane for the first stage, ethylene for the second, and methane for the third with a boiling in the -160s c. This is enough to liquify compressed air. All of this can be done with standard refrigeration compressors, piping etc, although you'll want to change the oil in the compressor to be compatible with the refrigerent that you will be using in each stage. Finally, if you're actually going to carry though with this, then this gas database from airliquide should prove indispensible.

Good luck and PM me if you need any further info! Maybe someday I can attempt something like this if I get off my ass.

Chris The Great - 19-10-2005 at 23:15

Just an idea...

Could one liquify a gas, such as ethylene, and then lower the pressure below atmospheric to further drop the boiling point until it could liquify air?

Just popped into my head, it seems like a good idea....

12AX7 - 20-10-2005 at 00:17

IIRC, refrigeration is a lot more efficient at elevated pressures. For sure, you'll be pumping a hell of a lot of evaporated gas volume for not much actual mass if you do it like that.

A shame that you need so much cooling to get cryo... for a given power input at the lowest stage, you need that plus its refrigerator's power use to be handled by the next stage, and so on. Nasty exponential growth there, a good reason to make the refrigerators as efficient as possible!


Marvin - 20-10-2005 at 07:58

In terms of working at all, the important values are the triple point of the refrigerent which determines the lowest temp you can achieve, and the critical point of the gas you are trying to liquify. When they overlap it can work.

Practically things become more difficult. You have to worry about compression ratios, and otherwise trivial things like how cold a gas can become before it freezes the pump trying to compress it.

There is of course the method using just air. Its been suggested by more compitant people than myself, that a fridge pump can do 30atm at least if pushed, though not I suspect with reasonable life. If you can get 30atm then liquid air is just a matter of lots of metal tubing and insulation and risk of explosion :) Hampson's machine only ran at 50atm. It wont be cheaper than buying LN2, but then neither would the cascade method I suspect.

garage chemist - 21-10-2005 at 11:43

Guess what I have been planning to do with my 40-Bar- producing fridge compressor even long before acquring it! :cool:

Cascade cooling with several cooling systems using different refrigerants will never work. Period. Think of the heat the compressor motors produce- 100% of it are subtracted from the cooling performance of EACH STAGE. Only compressors that have motors cooled by separate systems and with insulated pistons and cylinders can be used. Fridge comps rely on the residual coldness of the gas from the evaporator to cool the motors!

The Linde process using nothing else but "air and metal" is the only way that actually offers a CHANCE for success.

Marvin, you are absolutely right in stating that making liquid air is just a matter of lots of metal tubing and insulation when a suitable source of pressurized air of at least 30 bar is available (well, the issue of air purification, especially removing the oil mist produced by the compressor is also important but it can be done with lots of cotton wool stuffed into the metal pipe. Drying and removing CO2 from the air can be done simultaneously by a single chemical: potassium hydroxide, employed in the ambient pressure suction line of the compressor).
My idea for producing pressure in excess of 60 bar (for increased efficiency) is to pressurize the inlet of the fridge comp to 8 bar using an ordinary compressor. In theory, the comp multiplies this by 40 (in reality the motor will become blocked at maybe 80 bar due to the extreme overload which is put on it and the parts- think of the pressure difference it has to overcome).
If this fails, a huge 200 Bar pressurized air bottle (maybe a scuba tank?) will be used as the pressure source.

I have read through Ullmann's section on "Cryogenic Technology" with great interest, everyone considering building an air liquefaction apparatus must read this ( I really mean "must"!).
The importance of having a heat exchanger of extremely high efficiency (at least 99% for economic operation, every percent lacking from 100% efficiency means a huge overall efficiency loss) is emphasized several times. Most normal exchanger technologies reach maybe 70%.
Put every bit of your skill and knowledge into the heat exchanger, everything relies on it.

The best approach to the heat exchanger would be to use Linde's original design as the material for it is the easiest to acquire (no huge rolls of capillary copper tubing required like in other designs).
It consists of two coaxial copper tubes, the inner containing the high- pressure air and the outer the expanded air which precools the high- pressure air. The arrangement is coiled up in order to save space and make insulation easier (also very important. I'll use self- expanding PU foam to embed the whole exchanger. Linde used a wooden box filled with wool.).
I have already looked for the copper tubes and found that 8mm tubes fit nicely into 13mm tubes leaving maybe 2-5mm of air space between them, which is perfect.
Cost will be the limiting factor for the length of the exchanger. It will be at least 5m long (coiled up of course). Using a low flow rate in the exchanger also increases the efficiency a lot, but reduces the cooling capacity of the system and makes insulation even more important.
Overall, a low flow rate shifts the construction problems from the exchanger to the insulation.

The insulated vessel where the air will hopefully liquefy will be an all-steel thermos flask, tested for suitability by filling with bought liquid nitrogen and observing the evaporation rate. I already tested a 0,5L- thermos flask and the outside didn't become noticeably cold, also no boiling of the LN2 was observed inside it after the walls had become cold, which is a very good sign.

The connection of the exchanger to the flask will be made by a cork (only material suitable for getting it airtight without having to worry about glassification at cryogenic temps, also the vessel will be very easy to remove from the connection in order to measure temperature inside it or look for liquid air) with a 13mm hole in the middle.
The 13mm outer copper tube will be pushed through the hole in the cork, the inner high pressure line sticking out of it will end in the "Joule- Thompson Valve".

The "Joule- Thompson Valve" will consist in the high- pressure line ending inside the thermos being hammered flat as to allow only a small amount of air to escape and hoping that the 60 Bar won't push it open again. If it gets pushed open, soldering it and filing a tiny opening into the pipe will be tried.

When I have succeeded in filling my gas flask with ammonia, I will work on this project. I can already see the entire thing before my eyes, I have planned everything.

Bump for new kind of technique

watson.fawkes - 8-6-2010 at 08:16

Precis: About 1 L / day output rate. Key components are a pre-built cryocooler and a nitrogen separation membrane. Parts acquired on eBay; project cost under $500. Marginal cost of LN2 estimated at $1.15 / L.

not_important - 8-6-2010 at 08:42

Looks nice, certainly within hobbyist range, although obtaining the cryocooler at an acceptable cost appears to be a matter of luck. The membrane separation units are slowly becoming more common, plus there's the possibility of cascading several surplussed portable oxygen supply units to give oxygen enriched air and fairly pure nitrogen.

mr.crow - 8-6-2010 at 16:25

Thanks for the link, I never knew such a thing existed!

If you want liquid nitrogen dress like a student or professor and go to a university. They will give it to you for free because its not worth charging a few dollars to a credit card. They even had liquid helium but you need a better flask for that.

Edit: Video from Maker Faire

[Edited on 9-6-2010 by mr.crow]

densest - 8-6-2010 at 17:54

In the US, welding supply stores are usually gas supply stores as well. I bought a good quality dewar (expensive) and get LN2 for (IIRC) $0.50 a liter. The dewar keeps it for weeks.

JohnWW - 9-6-2010 at 13:18

What do you want to use the liquid N2 for, MadInventor? Surely not as a cooling drink on a hot summer day.

watson.fawkes - 10-6-2010 at 04:49

Quote: Originally posted by not_important  
Looks nice, certainly within hobbyist range, although obtaining the cryocooler at an acceptable cost appears to be a matter of luck. The membrane separation units are slowly becoming more common, plus there's the possibility of cascading several surplussed portable oxygen supply units to give oxygen enriched air and fairly pure nitrogen.
Luck: yes, the guy got really lucky in scoring his cryocooler. And now that his device has gotten so much web attention, the remaining once-a-year deals will be even harder to get. It's too bad that these are just not a DIY kind of part, what with all the machining tolerances required for helium as a working fluid. Perhaps the thermoacoustic devices currently in research will prove to be, given that the moving part is typically a piezoelectric transducer.

Nitrogen supply: The home medical units that have become so common recently are pressure-swing adsorption (PSA) units. The zeolite packing inside them preferentially adsorbs oxygen (and argon). There are other zeolites that preferentially adsorb nitrogen. Packing the columns with this other material yields quite pure nitrogen, > 99% as I recall, in a single stage. The raffinate (waste) gas is oxygen enriched (? 30-40%), but not pure. These devices should be considered extraction devices, not separation ones. Cascading them will work, but at rather much larger costs. The raffinate gas from an oxygen concentrator is still something like 10% O2, so you need quite a number of stages.

Incidentally: Last year I scored a batch of three-way solenoid valves (the key expensive part in a PSA system) on the cheap from a surplus vendor. I'll know more about sourcing these zeolites when I get to building a unit, as I'm planning to make the columns swappable for three outputs: O2, N2, dry air (with silica gel packing).

elementcollector1 - 25-3-2014 at 17:20

Does anyone know of a very small model of portable LN2 generator? This is important to me for a possible college research project.

Zyklon-A - 25-3-2014 at 17:47

I'm not sure if you already saw this, as it was posted earlier in this thread, anyway, there's this:

elementcollector1 - 2-4-2014 at 12:11

Saw it. It's small, but not quite as small as I was hoping. I guess I'll lay out my requirements, and then see if it's even mechanically possible:
-Portable! I don't really want to plug this in, unless that 'charges' it. The 27V DC to run the version below is doable (perhaps with 2 12V lead-acid batteries instead), but this would add a lot of size for just battery storage. Ideally, this would be powered by a series of 9V's - but that's just silly in terms of amp-hours.
-Small. As in, really small. Looking at the setup outlined here (, I could probably get rid of the nitrogen tank, and the dewar itself (as any liquid nitrogen produced will immediately be used, and then possibly sent back to the cooling unit as either gas or slightly-warmer liquid). I could also theoretically get rid of the nitrogen separation procedures, but then I'd have to deal with liquid oxygen. While colder than LN2, it's also more reactive, so I'm hesitant to introduce a safety risk. Also, I could probably stand to scale down whatever I can, as I won't need a high output.

This is for the purpose of cooling superconducting materials (to complement my recent interest), so once the pump starts flowing it could probably stand to recycle liquid nitrogen that has been warmed by the superconductor and pathway tubing.

Zyklon-A - 2-4-2014 at 16:08

By the way, How do you buy that LN2 generator? (the one I linked) I couldn't find anywhere on the site or any links posted on actually purchasing the thing. Am I missing something obvious?

neptunium - 2-4-2014 at 17:27

have yo looked at cryocooler yet? $20 seems low since everybody on ebay wants one (it seems) but i just got 2 of them and this one is working fine !

right now sitting at -91C ! and dropping!

0402142120.jpg - 132kB

neptunium - 2-4-2014 at 18:13


0402142200.jpg - 76kB

elementcollector1 - 2-4-2014 at 18:26

Quote: Originally posted by Zyklonb  
By the way, How do you buy that LN2 generator? (the one I linked) I couldn't find anywhere on the site or any links posted on actually purchasing the thing. Am I missing something obvious?

You probably need to contact them to find prices.
How big are these cryocoolers?

neptunium - 2-4-2014 at 18:41

about the size of a vacuum cleaner`s motor so not that big at all....but expansive ! somewhere arround 400 bucks!, however , these unit can run for years or decades continously without any problem....its worth it !

Chemosynthesis - 2-4-2014 at 20:56

You might also consider PMing ElizabethGreene here. She works at a cryonics institute.

elementcollector1 - 2-4-2014 at 21:00

I did not know that. I'm not sure what I would ask, though...
A vacuum cleaner motor is relatively small... Unfortunately, I think I'd need smaller (not knowing the actual dimensions of the motor - glancing at my vacuum cleaners, it's probably a 6" diameter by 12" height cylinder.

jock88 - 3-4-2014 at 05:52

Some light reading on Cryocoolers here:

Are they dooable in an amateur lab/garage?

Zyklon-A - 3-4-2014 at 06:08

As long as you have plenty of money and room to put stuff, almost anything is doable in an amateur lab. Room isn't the problem for me, it's the money.:mad:

neptunium - 3-4-2014 at 11:47

you can get lucky on ebay sometimes but you`d have to babysit the site because they go fast...
superconductor technology superfilters have one inside its quite a heavy box for shipping but you could still get your hands on one...they run on AC despite what the label says.... they come in all sizes i bet you could find a smaller one but its gonna be even harder to find...
Or you can always buid a joule thomson exchanger! if you save a bunch of money!
good luck

elementcollector1 - 7-4-2014 at 10:57

This is probably a workable size: Cryocooler on Ebay
Might have to shift things around a bit, but it looks doable at the very least. Can't say I like the price, though.

Ha! I couldn't build a miniature vacuum pump when I tried, I doubt I'd have much luck with a miniature JT cooler. Although I will be requesting this:
Of course, I can't really tell what 'miniature' means in this context...

neptunium - 22-4-2014 at 10:42

the smaller you go the longer it will take to get as low as LN2 temp because you`ll be lacking the power to condense nitrogen out of the air.
more power wil get you there quicker and you`ll end up with more product.
If you are going to spend the money why not make it worth your time ?

elementcollector1 - 22-4-2014 at 11:12

If I am going to build the product (keeping in mind that the product has size and power limitations) why not make it worth the build? I don't want to be carrying around a portable generator on my back...

neptunium - 9-5-2014 at 17:47

i have made some improvement (and spent more than i care to admit) but i think i am getting good at this!
Those cryocooler cost a little but are nowhere near the price on those liquid nitrogen generator!
Here is what i did...
with an older superfilter taken appart to reveal the cryocooler, i built a cooling jacket to go arround it
and fitted it with intake and out for the water .. (water is indeed a much better cooland than air)...

here is what the set up looks like

set up2.jpg - 41kB

Despite what is indicated on the superfilter , those things (the cryocooler ) runs on AC! but the rest of the unit is DC powered .. dont waste your time trying to run it on DC like i did...

Next build a Jacket outta PVC pipe and PVC cement outfitted with in and out for the water ..

cooling jacket2.jpg - 39kB

with that set up and a Pt probe thermometer in a dewar i was able to get to -195C ! (-320 F!!!)

coldest recorded2.jpg - 32kB

This is by far the lowest temperarute i have ever been able to obtain in my lab! If i remove the thermometer which is introducing a little bit of heat from outside i might be able to hit the -200!

elementcollector1 - 21-5-2014 at 10:27

Finally found the whole version of that document from earlier, but the build is very confusing...
What is the purpose of the wire insert?
Is it possible to reduce the ridiculous size of the cooling tube?
What is the power supply for this device, and where did they get a figure of "as low as 10 mW" for cooling power requirements?

Still, this looks at least *possible* to build at home... Although that being said, I still have to account for the pressurized input (which will likely require buying an adorably tiny air compressor).

Zyklon-A - 21-5-2014 at 10:36

Cool! I'm interested in hearing your results, I want a LN2 generator myself, but will wait for someone else to try it before I do!:D Or I might just buy one, I guess I should contact the people who sell those little one's.

The Volatile Chemist - 22-5-2014 at 05:12

Funny. I saw this post yesterday and today I saw an instructables about it. I haven't read it yet, I'll do so in a bit, but I thought it'd be applicable.

[Edited on 5-22-2014 by The Volatile Chemist]

Bezaleel - 22-5-2014 at 06:42

That's a great achievement, neptunium! ::thumbs up::

On the other hand, reaching that temperature is one thing, but generating LN2 may require so much power that you will only be able to have no more than just a few drops of it.... I'm very interested to hear your further progress.

neptunium - 23-5-2014 at 15:27

true, i have only been able to get a few drops from condensation but the temperature is there and the cryocooler is rated for a few hundred watts.
i have had leaks problems with the cooling jackett but i should be able to get as much as i want given enough time...

smaerd - 23-5-2014 at 19:31

Wow neptunium! -195*C that's impressive. My physics professor told our class that he hand built the cryogen system to do his doctoral thesis(at least I believe so). I'm really taken back by peoples home efforts at cryogenics. I didn't realize many people were willing to work with the Joule Thomson effect given the pretty extreme pressures involved here. Forget who it was, maybe it was magicjigpie, used to have in their signature "Friends don't let friends build autoclaves", or something like that. Love to see results, if you could post a basic schematic or sketch I'd love to see it. One day when I build a home I wouldn't mind having something set up to get liquid air.

The Volatile Chemist - his pressure swing absorber is pretty intense!

Bezaleel - 25-5-2014 at 14:12

Quote: Originally posted by neptunium

I'm impressed!!! (which I don't say often)

You shared some photo's above, but I'd certainly be interested in a more detailed write-up. E.g. what you based your choice of cryopump on. And how did you remove the oxygen, H2O, CO2 and noble gases from your air? Or did you use nitrogen from a pressure bottle as input? More technical details? (If time permits, of course.)

Do you have any special experiments in mind you created you LN2generator for?

The Volatile Chemist - 26-5-2014 at 06:53

Those low temperatures are insane (-195!)! If you put Ice around the thermometer probe, would that limit heat intake?

Zyklon-A - 26-5-2014 at 07:23

neptunium, is that your video? Or did you just post a link to a video you were copying.
If that is you video, then you just made liquid air, not nitrogen (of course you know this).
Seems like a risky way to handle liquid are.
Cardboard? The nitrogen will boil off more rapidly than the oxygen, and will take heat out if the air, thus possibly cooling more oxygen to below its boiling point. So the concentration of liquid oxygen will steadily rise, which could be dangerous near carbon containing compounds (eg. Cardboard).
I'm sure you've thought this through though.
Assuming that is your video.

neptunium - 31-5-2014 at 04:54

no i just copy the link in response to Bezaleel...
I have had some power issue with my variac lately , but i also noticed a change in the design of those
the one i have is an older model and takes about 2 to 3 hours to reach these temperature thus not powerful enough to generate a sufficiant amount of liquid air..
the newer unit can get low fast! very fast and within a few seconds generate a few ml of liquid air.

Yes Zylonb i`ve wrok with detectors cooled by LN2 and was trainned to handle it and air at cryo temperature, however there is no danger with the few drops i generated but you are correct this must be kept in mind and it fascinates me to see the clea liquid slowly turning light blue! and responding to a magnet!

imsmooth - 4-11-2014 at 19:03

I'm sure, as someone commented above, you've seen this video:

This is a summary of my efforts making LN2 without the aid of a pre-made cryocooler. I did it the old-fashioned way. As an added benefit, I used pure N2 from a homemade pressure-swing adsorber.

The tutorial for the LN2 generator is at

The pressure swing adsorbed is at

I also put together a highly accurate cryogenic digital thermometer. It is more accurate than the Auber, which I initially used, and measures to tenths of a degree, which helps me follow the cooling process more closely. You can read about this at

Hopefully, this will help those of you interested in this.

j_sum1 - 5-11-2014 at 01:32

Seriously cool. And not a bad video at all -- good skills there too.
Surprisingly, your cat took that quite well.

imsmooth - 5-11-2014 at 05:45

The cat is very curious.
Once the weather gets cooler where I live and the humidity goes down I'm going to run the generator again. In the summer there is so much humidity the scrubbers get saturated.

Unlike the cryocooler method, I can make a good amount of liquid in a short time, especially since my cool-down is 45 minutes. If anyone is interested in doing this he can use a larger compressor (like 12 scfm) to get higher production rates.

The PSA is also a great addition because I can get high flow rates of 99% pure nitrogen gas with just one stage.