Picture of Homemade liquid nitrogen generator
ln2 jet.jpg

Did you ever think you could make liquid nitrogen in your own garage? This is an industrial process so how can an individual do this? Still doubt me? Intrigued? Read on.

As a lover of science I tried to think of a challenging project that was out of the ordinary. After going through the internet web and Youtube I realized that no one had made liquid nitrogen in his home. Yes, I did see some videos where some would use a Stirling Cooler from a cryorefrigerator and use this to condense nitrogen on the exterior of the cold-head. While one is making liquified gas, this is done using a prefabricated machine. I wanted to make the machine that liquefies the gas. Furthermore, a cryocooler has a very low production rate. You will only get about 500 - 1000ml per day. On the following pages I will walk you through the basics of how to build your own liquid nitrogen generator. Using easily obtained materials you can liquefy nitrogen or air. The unit cools to -320F in under 50 minutes. The production rate is about 350 cc/hr.

A full tutorial and plans are at This page goes over theory, thermodynamics and more detail on where to get components and how to build this. This Instructable serves as a general introduction to how this baby is put together.

I have just added a new web tutorial on how to make your own N2 gas from the air. I will add this as a new Instructable in the next few days. You can get a link for it at the end of this one.

I have also built a high-precision cryogenic digital LCD thermometer for this project, which you can buy for yourself. You can see how it compares with an Omega digital thermometer here.

Ok. The video above gives you a quick 3 minute overview of the project. At the end of this tutorial I briefly mention the PSA I made for making the pure N2 from the air for the generator. If you're ready for 320 degrees below zero we can begin...

Remove these adsRemove these ads by Signing Up

Step 1: Overview

Picture of Overview
early version2.jpg

The liquefication generator has a few basic components. Starting in order:

1. Scrubber - This removes CO2 and H20 from the gas stream. Without this the water and CO2 would freeze and clog the tubing and valves

2. Filter - We need to remove any micro-particles that can clog our compressor valves

3. Compressor - This compresses the gas to high pressure. Two important factors are the pressure and flow rate. This project uses an oil-free scuba compressor delivers a pressure of 3500 psi at a flow rate of 3 SCFM (I jacked it up to 4 SCFM). It is possible to use a regular refrigerator compressor, but the production rate will be significantly reduced

4. Pre-cooler - This cools the hot, compressed gas before entering the cooling tower.

5. Regenerative cooling tower - Hot compressed gas flows through a counter-current system to cool the gas to cryogenic temperatures. Expanded, non-liquid gas returns to get recompressed.

6. Throttle - This is a needle-valve that enables a controlled expansion of the gas without losing the pressure behind it.

7. Baffle - This reduces the velocity of the expanded gas so it does not dissipate the cooled liquid into the gas-stream. It also provides a larger surface area for condensation.

7. Reservoir- This is the collection system that collects the gas. Heat exchange with the environment is minimal.

Above is a picture of an early version of the generator using a recycle can. Later you will see an improved version.

Step 2: Compressor Filter

Picture of Compressor Filter

You need to protect the compressor's valves and cylinders from debris and water. You can purchase an inexpensive vacuum filter from a distributer. I bought a Vaccon 10 um filter. The part number is VF500F.

Step 3: Compressor

Picture of Compressor

This is the most expensive component. I used a RIX oil-free SA-3E. This delivers 3 SCFM @ 3500psi (230 ATM). I modified the motor and pulleys to deliver 4 SCFM. A higher flow gives you a faster cool-down and production rate. The high pressure allows you to have a larger temperature drop when you throttle the gas to a lower pressure. You can use a regular refrigerator compressor, but you will be waiting a long time to drop 400 degrees Fahrenheit from the ambient temperature to -320F if you only have 40 ATM of pressure.

My compressor allows me to get to -320F in 45 minutes. I am guessing that a standard compressor at 40 ATM will take 6 times longer, or 4 1/2 hours. Of course, if the flow rate is less then you have to add more time.

Step 4: The Scrubber

Picture of The Scrubber
scrubber top.jpg
molecular seive.jpg
scrubber new.jpg

You can make this components out of PVC pipe. I used 4" PCV. You need to get the correct fittings to get to a pipe size that you can then attach to your compressor. The scrubber uses a material called zeolite, or molecular sieve. I specifically used 13x molecular sieve and added some 4A sieve material that had a color indicator added so I know when it has reached saturation. You determine the size of the scrubber based on your flow rate. If you want 3-4 SCFM then you need about 10-15lbs of the material and size the scrubber based on this.

Zeolite is a naturally occurring, finely porous material. By fine I mean the holes are only angstroms in size. Molecular sieve is man-made, but the structure is the same. CO2 and water enter and bind within the pores, letting the other gases, like O2 and N2, to pass through. The color-indicator will change from light blue to gray when the material can not adsorb any more water. One regenerates the material by heating to 350F, driving off the CO2 and water.

Now, back to the device. I cut a disc of 200 micron screening. This holds the 10 lbs of material in place so it does not get sucked into the compressor, but still allows for air-flow. I glued this material between the 4-to-2 inch reducer and the coupler fitting. We need to filter smaller material, so I took 11 MERV air-conditional filter material and cut it into a large circular disc. This should filter 1-3um particles. If you can get a 12 MERV this is better. I fashioned a circular ring of stainless steel that I got at a hardware store and fixed the filter material on top of the 200um screening. Then, poured in 10 lbs of 13x and 4A sieve. I then made another filter disc out of the air-conditioning material and fixed it on the top with a stainless steel ring. This keeps the material clean. You don't want particles clogging the sieve's pores.

The top of the scrubber needs three inputs. One is to allow for fresh gas input. The other connects to the regenerated gas. This is gas from the cooling tower that did not liquefy. The gas is already cool, free of CO2 and water, and in my case is 98.5% nitrogen. I don't want to waste this gas to the atmosphere so I reuse it, reducing the amount of fresh N2 that I need to make. This is why it is called regenerated gas because it comes from the original gas I fed into the tower. The third input allows any excess gas to vent out.

The only other connection is at the bottom where you connect the scrubber to the compressor. I placed a 10um vacuum water trap between the scrubber and the compressor. This will remove particles that can fowl up the compressor, and also serve as a check that no water gets into the cylinders. I also used this spot as a place to insert a gas-analysis meter that I made to monitor the purity of the N2.

You will need to buy a pipe/tube bender to make the coils and rings. I've shown a picture of one above. Mine is meant to bend 1/4" tubing/rod.

Step 5: Pre-cooler

Picture of Pre-cooler

The compressed gas leaving the compressor is hot. You want to remove this heat using an ice bath. The pre-cooler is sized to fit into a large bucket.

Basically, the precooler is a long coil of tubing. I used 20 feet of 1/4" 304 seamless stainless steel with a 0.035" wall thickness. Again, one could go to 0.027" wall, but this got too thin for me. I did not want a wall rupture at 3500 PSI. I go through the engineering calculations for verifying the integrity of the tubing at the tutorial site.

I added 0.016" thick aluminum fins which increased dT/dt even further by allowing better heat transfer. I carefully drilled a hole 1/64" smaller than the tubing. I then cut the square and snapped the fin in place. I can then submerse the cooler in an ice bath or expose it to sub-freezing outdoor temperatures when available.

Step 6: Regenerative Cooling Tower

Picture of Regenerative Cooling Tower
original coil packing.JPG
original coil in can.JPG
coil with teflon1.JPG
coil with teflon2.JPG

This is the part that will take the most work. My final tower uses a large concrete cardboard form container that is 50 inches high and 24 inches in diameter. However, I had good success with an earlier version that used a 36 gallon plastic recycle can. The compressed gas goes into the tower through a long helical coil and re-expands through a needle valve at the end into the reservoir. The cold gas that did not liquefy returns through Teflon tubing around this coil. This cold gas cools the steel tubing further. This process repeats until the steel tubing is cold enough to liquefy the gas.

I used 305 stainless steel 1/4" tubing with a 0.035" wall thickness. There are many places that sell this near you. If you are going to do this with a standard refrigerator compressor than you can use copper tubing. Just make sure it is rated for the pressure from the compressor. Now the PTFE (Teflon) tubing is critical. It is one of the few materials that is flexible and can withstand cryogenic temperatures. I used plastic tubing in the beginning, but it would eventually crack.

My tower uses 80 feet of tubing. You can do this with 40-60 feet, but this increases the cool-down time. Again, everything is about tradeoffs. You need to insert the steel coil into the Teflon tubing. You want the Teflon tubing diameter to be a little bigger than the coil. Mine was 1" corrugated tubing which allows it to be flexible. This wraps over the 1/4" stainless tubing.

The ends of the tubing need high-pressure, stainless steel fittings and adapters. I got mine from Swagelok, but there are other companies that sell similar parts.

You then insert the coil into the garbage can or container. Industrial manufacturers fill the container with perlite and create a vacuum to prevent heat from entering from the environment. I filled my container with alumina silica high-temperature wool. This is normally used to insulate furnaces to contain the heat. Well, guess what? It also contains the heat of the outside from getting to our coils. You need to loosely fill the container with the material so it stays fluffy. Air is also an excellent insulator.

The pictures above show me using the 36 gallon recycle container. At the end of this tutorial you will see the concrete form container which houses a bigger run of coil. The stainless steel tubing is surrounded by the Teflon tubing. This, in turn, is wrapped with polypropylene foam to further decrease heat penetration.

Step 7: Throttle

Picture of Throttle

The gas expands using a needle valve as mentioned earlier. One controls the pressure by finely tuning the orifice opening with a large lever/knob. I made a large lever because the standard knob handle is too small for fine control. Furthermore, as the temperature drops one finds difficulty turning the knob because it gets frozen in position. A large lever makes this fact mute.

I extended the stem so it would reach the outside. I insulated the connection to the stem with Teflon to reduce any heat transfer through the stem from the outside.

Step 8: Baffle

Picture of Baffle
baffle for insertion.jpg
baffle for insertion2.jpg

The expanded gas distal to the needle valve moves at a high velocity. This can agitate the liquefied gas in the reservoir and blow it back out into the gas stream. You need to slow the gas down to allow the droplets to collect and drop into the container. Until I figured out this simple step I was never collecting a lot of liquid.

Making this is very easy. I used a fitting and attached it to 1/4" copper refrigerator tubing. I drilled a few holes with a 1/8" drill bit on the other end and fixed some copper abrasive scrubbing sponge that I got at a food store. Make sure the copper wool (abrasive scrubber) does not come off. The other pictures show how this part relates to the needle valve and reservoir.

Step 9: Reservoir

Picture of Reservoir
reservoir ready.jpg

The reservoir holds the collected liquefied gas, whether that is N2 or air. A simple stainless steel vacuum bottle, or a thermos as many of you would know it, serves this purpose very nicely. Such a simple solution that alluded me for some time. The trick is to fashion an adapter that allows you to contain the expanding gas and direct it to the regeneration Teflon tubing that surrounds the high pressure stainless steel tubing. The details, along with all the other methods I tried, can be found at my extensive tutorial at

Basically, the baffle connects to the output of the needle valve. The inflow to the valve connects to a tube fitting on the end of the high pressure tubing. A Teflon cylinder surrounds this and just fits over the orifice of the thermos. Small screws keep this fixed so the pressure of the expanding gas does not push it off. A small tube of Teflon goes from the bottom of the thermos to the outside through an insulated jacket to allow the liquefied gas to escape. A small plug keeps the liquid contained until you are ready to drain the system.

In addition to these connections, there is a small opening for a RTD probe for monitoring the temperature. I am currently using a unit I bought, which is seen in the video at the beginning of this Instructable. I have developed a low-cost LCD cryogenic thermometer that is accurate to 0.1F. I plan on posting this as a new Instructable in the near future.

Step 10: Summary

Picture of Summary

This project allows an individual to do what only commercial industry has done in the past. You can generate your own liquid air or nitrogen with a high production rate. The cool-down period depends on your compressor, length of regeneration tubing and how well you can insulate your system. In the past, making liquid nitrogen in your garage seemed impossible. Not any more.

Good luck and stay cool.

Step 11: Making Pure N2 for Liquefication

Picture of Making Pure N2 for Liquefication

I have gotten a lot of comments suggesting that I am making liquid air, which contains O2, and not liquid N2. I made a pressure swing adsorper that removes the O2 from the air, leaving 98.5% pure N2. If I used a second stage I could get 99.999% pure N2. I feed this N2 into the liquid nitrogen generator. I am putting out about 30 L/min of 98.5% N2.

I have just finished writing a web tutorial on this subject which you can find here. This can also be used for having a pure source of N2 in your garage. I use it for filling my tires. I will write this as its own Instructable in the next few days.

You can click on the this link to see a video I made, showing my PSA device.

1-40 of 79Next »
imsmooth (author) 2 months ago

I am thinking of selling a low-cost, highly accurate cryogenic thermometer that I designed and built. Please see my webtutorial for details at

If enough people are interested I will make a run of 10.

Also, as an aside, my daughter made her first video and I would appreciate it if you could just visit her video and LIKE it. Thanks. It is here:

en2oh1 month ago

can you give any pointers on where you sourced you CMS?

Btw, I just got my Stirling Cooler up and running. Cooling down as I type.

imsmooth (author)  en2oh1 month ago
I'm still trying to set up my source to get the CMS for everyone. I have not forgotten your inquiry.

Also, I've developed a high precision cryogenic thermometer. You can read about it on my site at
Ugifer5 months ago

This is a fabulous instrucable, but I do have one concern: do you not find that you condense out liquid oxygen, at least to start?

Oxygen liquifies at a higher temperature than nitrogen and one of the classic mistakes people used to make with the liquid nitrogen traps in the lab was leaving the pump running and drawing loads of air through. You would end up with a big trap full of beautiful blue liquid oxygen and a terrible danger of sudden explosion or raging fire!

In spite of that hazard, you have achieved something that i wouldn't have thought possible at home, so great work there. I would like to say that I will be doing this but I have a feeling my wife knows that it's then only two easy steps from here to liquid-oxygen-barbeque-lighting and would instantly ban me from even trying!


imsmooth (author)  Ugifer2 months ago

It takes about 30 minutes to get to a point where O2 liquefies. By then the constant inflow of 99% pure N2 has washed this O2 out. I use an oxygen analyzer to measure the O2 coming out and it is LOW.

telstarpk Ugifer2 months ago

You could use the liquid Oxygen to make a rocket. Theoretically, you could also make liquid hydrogen, but it would take a better compressor, a better insulator and a better container. if there is a spark, though... 3... 2... 1... BOOM! Goodbye garage!

exploded garage.jpg
imsmooth (author)  Ugifer5 months ago
That's a good pick up there. I may have briefly mentioned it but I'm using a pressure swing adsorber to take regular air and make pure nitrogen gas. It Is this pure nitrogen gas from the PSA that I liquefy.
mattvdb imsmooth3 months ago

do you have any details on the calculation of qty CMS used for the PSA system.

is the calculation based on the fact that only 21% of air is Oxygen??

Ugifer imsmooth5 months ago

If you mentioned it then I didn't pick up on it. Does that mean that the molecular sieves were used as a PSA as well as taking out the CO2 & water, or are they separate units? I suppose the PSA must be on the high-pressure side of the compressor so I guess it's a separate unit but I don't have too much background in pressurised gases.

Once again: great method & amazing instructable.


imsmooth (author)  Ugifer5 months ago
I'll add a separate step explaining about the PSA

The PSA is a completely different device which I may list is another tutorial in the future. This is only to remove the oxygen from the air. I actually have a separate scrubber removing the water and carbon dioxide going into the PSA.

The scrubber described in the tutorial here is just for removing water and carbon dioxide going into the compressor for the ln2 generator.

The PSA has a working pressure of about 100-120 psi
KumailA2 months ago

Do you think this will work for liquefying air ?

imsmooth (author)  KumailA2 months ago

Did you ever get it?

This is awesome, and I really want to do it, but it seems super-expensive.

Is there to do this more cheaply (even if it means that it'll take longer to make less liquid)?

Please let us know about cost-lowering trade-offs. :)


imsmooth (author) 3 months ago
I asked the supplier what amount is needed for a flow of 1 scfm

I think these are derived empirically
Ωmega4 months ago

Great project, but is it possible to liquify other gasses (such as noble gasses), or is this "reaction" only applicable towards Nitrogen?

imsmooth (author)  Î©mega4 months ago
as long as the inversion temperature is above ambient, you can do this for other gases. You would need to check the value.
ViperSRT3g5 months ago

I'd love to make this, but I'm unsure of how I could put this liquid N2 to work.

liquid nitrogen is used in many shop/industrial applications. I need it for cryo-treating certain alloys, shrinking parts for interference fit, purging oxygen from containers. Freezing corpses...

Haha, aside from the fact that this would be a great learning project, I can't justify creating a nitrogen generator JUST to make ice cream XD

RayJN ViperSRT3g5 months ago

Very, Very hard ice cream

I've looked into it, checking professional cooking establishments around the planet and everyone who has a fair bit of experience at making this type all say the same thing in that liquid nitrogen ice cream is smoother and creamier than the normal kind. You can even take the same mixture used to make your ice cream, split it into two halves and using liquid nitrogen with the one half and a traditional ice & rock salt hand cranked ice cream maker with the other half. Making both batches on the same day, in the same weather conditions, choosing whatever conditions are the best for the traditional method. And I will bet you every time the traditional stuff is just not going to freeze fast enough to prevent ice crystals from forming. - and the liquid nitrogen ice cream is made in less than 10 minutes....

Maybe not, but to me ice cream is a very serious issue. I don't like any ice cream that blends air into the mix, uses cheap low-fat milk or maybe even adds a bit of water to extend the savings. And then there's that 'beaver butt' vanilla replacement I don't much care for. And I really like the idea of freezing alcohol into the mix, which is not so easy when you make ice cream the old fashioned way

MOST vanillin comes from the paper pulp processing industry as a byproduct of using wood in the process. There's more paper made than there are beavers available for industry.

Sure, but how do you confirm the actual ingredient used when the label only says "Natural Flavor"?

Sort of like winning the lottery, or getting hit by lightning. Millions of tons of wood pulp are processed and vanillin usually comes from that. Beaver hunting tags are quite limited and a thousand tons of beaver butt may not exist in the entire planet. Confirmation may be bit and miss, but it's a really BIG miss. (Just my thoughts on statistics, not trying to argue with you. I don't like the idea of eating it either so I use real vanilla beans and avoid synthetics as much as possible.)

"Beaver-butt"! Ha! It's actually the fake raspberry flavor that comes from beaver's butts. For those that think we're making this up, just google "castoreum".

i agree with u:)

rusticator5 months ago

A cheaper high-pressure compressor is It is available with 4500 and 3000 max psi for $650 plus an oil-less utility compressor for $50-150 as a front-end. Designed for airguns and paintball. Works well.

imsmooth (author)  rusticator5 months ago
Flow rate is too low
I had looked into yhese
Shkinball5 months ago

What do you do with all this Liquid nitrogen though?

scitch5 months ago

Wow! A SCUBA compressor is $1,500-$3,500 on craigslist!

imsmooth (author)  scitch5 months ago
If you get one make sure it is oil free and has a sufficient flow rate.
espdp2 imsmooth5 months ago

As some folks can get several fridge compressors, is there any good reason that you couldn't run two or more in parallel to produce a better flow rate?

There's a risk they produce significantly different pressures. That can
have unpleasant consequences. If you use several of the same make and
type, though, that can be an interesting idea. They also produce a lower
pressure, tho.

lperkins scitch5 months ago

Should be possible to use lower pressure, but you'll need more insulation to compensate for the slower heat extraction.

flashcactus5 months ago

Great 'ible. I'd build one if I didn't live two blocks away from a
company that makes all sorts of gases, both liquefied and not, and sells
LN₂ at ~ $1.60 a liter. Maybe I'll build one anyway just 'cause it's
cool (and I can liquefy other gases with it, too).

What's the lowest temperature that this setup could practically produce?

You've also mentioned that the steel tubing for the regenerative heat exchanger came in 20-foot sections. Did you weld them together or what?

By the way, the compressor page on your site apparently has a typo:

>a temperature, T, of 32C (273K)

You probably meant Fahrenheit, as 273K ≈ -0.15°C.

imsmooth (author)  flashcactus5 months ago
The temperature should be able to go as low as the boiling point of the gas in the system. Of course, if the inversion temperature is too low you will not be able to cool the gas with throttling. You would need to implement a different method like a turbo expander.

BTW, thanks for pointing out the typo. It should have read 0C like you said.

As far as the connections I think I mention somewhere that I use high pressure tube fittings. I got these from Swagelok.
wierd idiot5 months ago

simple amazing, would never of thought it was possible to do at home. Good instructable man.

1-40 of 79Next »