Sciencemadness Discussion Board

air liquefier attempt, questions

condennnsa - 11-4-2012 at 16:32

So I've wanted for a long time to attempt to build an air liquefier using a fridge compressor.
I've done plenty of reading, trying to understand the best I can how these devices work, and what is in the
reach of my technical capabilities.
The setup I will follow is actually similar to what i believe Carl Von Linde built about 100 years ago, managing for
the first time to liquefy air.

Anyway, I bought a new 180W fridge compressor for $80 .
2800 rpm, 4.56 cubic cm displacement, air flow of 790 litres/hour, or 220 ml/second.


I also bought 30 meters of 6.35mm OD 0.8mm thick copper tubing and 3 meters 15mm OD 1mm thick Cu tubing.
A 0-40 bar manometer, will get also a 0-200 bar one.

I plan to use the 3 m 15mm tubing to make the first heat exchanger , right after the compressor, this will be cooled by running cold tap water, which should cool the compressed air to about 8-10C. I will do this by placing the copper tubing in a larger diameter hose , and running cold water continuously through it.
Now from what I've read in industrial settings, they cool the first heat exchanger by underground piping, which cools it to about 14C, so I reckon that mine will be a bit better, I may have to use more than 3m of tubing to bring the air all the way to the temperature of the water, though

The 30 m tube will be the place where the magic happens of course.
I still have to buy 30 meters of hose, 12mm ID at least, to slide it over this copper tube to make the regenerative cooler. And also a lot of pipe insulation to go over the hose, I want to put it 1 inch thick at least.

The expansion valve I will use is an all brass screw valve, that came from a household butane tank.

While I was thinking of my setup, i soon came to realize that by far my biggest problem will be atmospheric moisture , and to a lesser extent atmospheric CO2 which accounts for about 0.04% of air.
I will have to build the air desiccators myself, and plan on placing them at the intake of the compressor, even though these devices are usually placed on the pressurised side, I can't trust a homemade tube packed with desiccator running at 40 bar or more.

Do you have some ideas on building these desiccators? What desiccant should I use? CaCl2, or NaOH? I can get NaOH very cheap, and this also has the advantage of removing CO2 from the air, on the other hand from what i know CaCl2 is more hygroscopic than NaOH and will dry the air better. I also thought of building 2 desiccator tubes one with CaCl2 and one with NaOH, placing them one after the other at the intake.

Do you have any other ideas on how to get super dry air? to ensure that no water will condense, or freeze, at the expansion valve, which will be by far the coldest point of the system?

I played a bit with the fridge compressor , and the manometer, I took it to 30 bar, it had no problem whatever. I'm under the impression that it can do at least 50-60 bar.
What I don't like about it is that it 'spits' oil , but from what I've read all fridge compressors do this. I do have a couple of ideas on how to make sure this oil doesn't go into the whole system, though.

I also drew a Paint picture of my setup:
http://s13.postimage.org/e2stnlorr/untitled.png

Thanks if you've read all this and I'd really like to hear your thoughts!!

[Edited on 12-4-2012 by condennnsa]

[Edited on 12-4-2012 by condennnsa]

[Edited on 12-4-2012 by condennnsa]

watson.fawkes - 11-4-2012 at 19:39

Quote: Originally posted by condennnsa  
Do you have some ideas on building these desiccators?
Use another refrigerator and pre-cool your intake air. A two-stage autocascade with the right refrigerants should do you. It's also a good shakedown run for the more difficult problems involving even lower temperatures.

Do you have a pump lubricant that can operate at LN2 temperatures? It'd be a good thing to ensure, also, that the mechanical tolerances of the pump still hold at low temperatures.

Twospoons - 11-4-2012 at 21:27

^ what he said. Also post compressor and water cooling you could add an air tank placed in a domestic freezer. 2 benefits: the air should come out very dry, and also very cold before entering the regen cooler.
I'll be very interested to see how this works out. Keep us informed of progress!

Wizzard - 12-4-2012 at 05:22

Your compressor is going to heat things up considerably.
Instead of using just cold water, consider a salt/ice bath. That's a bit colder, and you wont need much cooling capacity for the air.
Keep the previously-dried air inside the clean loop, and have a low-pressure valve on the intake which allows air in, after travelling through the cleaning filter. Dry, cold normal dessicants will do nice to remove water, followed by calcium dioxide or sodium hydroxide for CO2 scrubbing.

Endimion17 - 12-4-2012 at 06:15

I hope you'll succeed, because this is quite a challenge. I'm not sure if your compressor can supply enough pressure for this to work with air, but I'd give it a try. However, have lots of quality insulation like styrofoam beads. I'd wrap the cooling elements with glass wool and put it in a box and stuff it with beads.
(if you don't have the beads, acquire beaded styrofoam and few kids; kids love destroying styrofoam, I know I loved it :D)

Consider using tubes of scrubbers and dessicants instead of large tanks. It's always better to let the air pass through a long tube will less material, than passing it through a small tank holding lots of it.

You'll also need few one-way valves, some exposed to normal temperatures, and some that can withstand extremely low temperatures. I recommend aquarium valves with a dab of inert oil for room temperatures, and home blown borosilicate glass valves can be the solution for liquid air, if insulated metallic ones aren't available.

Use the largest thermos flask you can find for the expansion vessel. I think the pressure inside is almost atmospheric, so bursting shouldn't happen. Avoid the stupid designer ones, with metallic casing. Use plastic ones, and stuff styrofoam beads inside. I did that to one of my flasks and it works like a charm, compared to a flask without the beads.

Remember to put safety valves, too. You don't want things to blow up. :)

dann2 - 14-4-2012 at 12:23


As said above the 'compressor' will heat things up. I put the compressor in commas because it made by putting the motor and actual compressor in the one sealed container. The motor will keep things warm. You would need to seperate the motor and compressor.

Book here:
http://freebooksearcher.info/usenetdownloadfree.php?q=liquid...

called Liquid Air by T. O'Connor Sloane

[Edited on 14-4-2012 by dann2]

BromicAcid - 14-4-2012 at 12:44

There are other threads on this. At the time of my research cascade cooling was the most reasonable to attempt at home. To do this in one stage seems a monumental feat. Even the refrigerants to use in the cascade are known although they would have to be procured or produced since they are not commonly used as refrigerants.

https://www.sciencemadness.org/whisper/viewthread.php?tid=15...

http://www.sciencemadness.org/talk/viewthread.php?tid=4333

Bikemaster - 14-4-2012 at 15:01

Remember that you need to be under the critical temperature of the air (Nitrogen/Oxygen) for the systeme to start producing liquide air.
http://en.wikipedia.org/wiki/Critical_point_%28thermodynamic...
If you can isolate the regenerative heat exanger enough it will probably work, but I will sugest you to use a multiple stage compressor (at least 2).If it's possible you can chill the second compressor output gas with frozen CO2. If will help you get these really cold temperature.

Good luck

watson.fawkes - 14-4-2012 at 18:08

Quote: Originally posted by Bikemaster  
Remember that you need to be under the critical temperature of the air (Nitrogen/Oxygen) for the systeme to start producing liquide air.
Not the critical temperature, which is about phase changes, but the inversion temperature, at which the Joule-Thomson coefficient is zero. They're different things.

The inversion temperature of N2 is 621 K and that of O2 is 764 K, both well above room temperature at about 300 K. There's no issue here for air liquefaction.

condennnsa - 15-4-2012 at 01:11

Thanks guys for the support. I've been really ill for 3 days now I don't what it is, but my back hurts like hell. It's ridiculously painful to perform basic taks such as walking :(. I suspect a herniated disc, but i'm not sure because i had episodes of back pain before which went away...

Regarding watson's and twospoons' recommendations , was also thinking to use cold to get dry air, but i was hoping i could get away without, just with desiccants. I searched the web the best I could for conclusive data about the dewpoint that each desiccant can achieve, but couldn't find any.
For example the wikipedia air dryer page states that:
"A regenerative desiccant dryer typically delivers a dew point of between -40°F(-40°C) and -100°F (-73°C)"

according to that page a desiccant dryer will dry the air the best. Of course if this data is for P2O5 as a desiccant, which is one of the most powerful, then it's irrelevant for me. Basically I will choose CaCl2 or NaOH or both , these I can get cheaply and in pellet form to allow easy air circulation. I also read on some other page that anhydrous CaSO4 is more powerful than CaCl2, so I consider that as well.

Wizzard, yes a fridge compressor can get quite hot after tens of minutes running, but I don't see a problem, as the watercooled first heat exchanger should cool the air pretty quickly to the waters temperature. Keep in mind that since this air is high pressure, it will move very slowly through the pipe, for example at 30 bar the flow will be 220/30 ml/sec, or 7.3 ml/s . The 3 meter tube with 13mm ID has a volume of around 400ml, so the air will spend about 1 minute in the cold tube.

Endimion, I'm on a budget here :). We have one refrigeration technology supplier in out city, and the safety valves they sell are ridiculously expensive, a 28 bar one is around 35 euro, as pressure goes up, price goes up exponentially.
I actually thought about using the compressors stalling pressure as a safety feature. 2 days ago I got a 0-400 bar manometer, I'll try to connect it directly to the compressor and start it to determine the pressure at which the motor is overloaded and stalls. I'm a bit worried though, what if the piping bursts... these fridge compressors, even if they look harmless, are mean little beasts

For the expansion tank I'll take your advice, i'll look into supermarkets for some thermos, as I don't own one, I also thought of just using expanded PU foam. Yes the pressure is just a tad over atmospheric , depending on the ID of the regenerative cooling tube and its length.

oh yeah, what also pisses me off is that copper fittings of small diameter <15mm are also expensive as hell with a mean price of 1.5 euro each. while a 15mm 90 degree curve is about 0.2 euro, a 12 to 8mm reduction is 1 euro, and weighs about 1/10 of the 15 mm one.

For the time being I set a goal in my mind and that is a temperature of -50C. If I will see that temp displayed on my multimeter screen, I'll jump with joy.


[Edited on 15-4-2012 by condennnsa]

[Edited on 15-4-2012 by condennnsa]

BromicAcid - 15-4-2012 at 07:11

Dewpoints of air provided by different desiccants from "Handbook of Preparative Inorganic Chemistry" available in our library and provided by by S.C. Wack.

gasdewpoints.jpg - 112kB

condennnsa - 30-4-2012 at 01:50

thanks bromic, for spoonfeeding me, i guess i should've been more persistent . I found another chart in "Purification of laboratory chemicals" it doesn't give dew point data, just puts them in order , it goes like this:

P2O5 >> BaO > Mg(ClO4)2, CaO, MgO, KOH (fused), conc H2SO4, CaSO4, Al2O3 > KOH (pellets),
silica gel, Mg(ClO4)2.3H2O > NaOH (fused), 95% H2SO4, CaBr2, CaCl2 (fused) > NaOH (pellets),
Ba(ClO4)2, ZnCl2, ZnBr2 > CaCl2 (technical) > CuSO4 > Na2SO4, K2CO3.

I tested the compressor and it is only able to go to 40 bar, it doesn't stall however, it seems that the compression ratio is about 1:40 so that the air eventually ends up being compressed and never leaving the cylinder .
here is a video I shot
http://www.youtube.com/watch?v=hygPGnjluL0

I will have to buy another compressor to pressurise the intake to a moderate 5-10 bar pressure that way I should be able to reach the 80-100 bar which I'm after. I know a guy who owns fridge repair shop that carries lots of them cheaply.

a few pics:






This is the 30 meter future to be regenerative cooler, the copper tube was a real PITA to push all the way through the PE tube.

The regenerative cooler is the part that's troubling me the most on how to build, I'm still undecided on whether to put piping insulation over the black hose, or to build a big box out of thick styrofoam boards, piping insulation is really expensive.





This is to collect oil from the compressor and prevent it entering the system, i can empty it when full and pour it back at the intake. It seems to work well.


[Edited on 30-4-2012 by condennnsa]

neptunium - 30-4-2012 at 16:41

this is quite a project ! i tried that a few years ago unsuccessfully. i have a book written in the late 1800 and early 1900 it explained everything about liquifying air .
and even today it is still a big tidious thing to do for a home scientist.
i didnt say impossible but costly and very tricky
best of luck to you!
every calorie count!! dont waste ANY!

condennnsa - 30-4-2012 at 21:31

Neptunium , what is the title of the book? Now you got me worried saying you tried this and failed... could you please post more details on your setup? like approximate pressures used, air flow, regenerative cooler design, insulation, and temperatures reached? thanks!

[Edited on 1-5-2012 by condennnsa]

Twospoons - 30-4-2012 at 22:12

Don't forget to put some kind of safety shield between you and those compressors. Just in case ...

As for insulation - have you thought about using 2 part rigid urethane foam? You could put your regen in a box and fill with foam. Downside is its very messy to take apart again.

[Edited on 1-5-2012 by Twospoons]

johansen - 1-5-2012 at 00:46

I have 4 large, 500 watt dehumidifier compressors and i have not even started on a liquid air project for a few reasons.

The first is that i'm interested in more than just 1% thermodynamic efficiency, and i intend to build a turboexpander in the near future (a few months to a year) ((either machining the impeller myself or using automobile turbos and a custom manifold))
The second is that to get joule-Thompson cooling from 10C down to -170C requires an insane amount of work.

You will need to insulate your heat exchanger inside a container filled with Perlite, and pull a vacuum on that Perlite, in order to get the heat input low enough to get a practical amount of liquid air.... you will need to pull 10-50 millitorr on said perlite insulation and you will need a heat exchanger around 30 meters long...

phlogiston - 1-5-2012 at 02:53

I was use conc. H2SO4 to get the last bit of moisture out of gasses (nitrogen mostly) which seems to work well.

babakfarhadi - 5-12-2012 at 03:16

Dear
Congratulation to you.About one month ago i try to build a linde cycle scale and this is a pic of project :

https://docs.google.com/open?id=0B-X7U-fPaR5hM0diSE50b21FX0k
---------------------------------------------
https://docs.google.com/open?id=0B-X7U-fPaR5hQjBzc29HSDhQck0
---------------------------------------------
https://docs.google.com/open?id=0B-X7U-fPaR5hVXVXSGhsM1p6d1U
---------------------------------------------
https://docs.google.com/open?id=0B-X7U-fPaR5hSC1XY1ZuRllpa1E
----------------------------------------
https://docs.google.com/open?id=0B-X7U-fPaR5hOFZPSlZ5ZkFPMGc
--------------------------------------------
https://docs.google.com/open?id=0B-X7U-fPaR5hRDA1Qm5KTkFWSEU
--------------------------------------------
https://docs.google.com/open?id=0B-X7U-fPaR5hb3czeTZVZ0hXTlE
--------------------------------------------

And test this project with nitrogen and temp below down until -10 degree centigrade.I think a ice build in expansion device.
Any one can help for get a better result?

Best Regard

[Edited on 5-12-2012 by babakfarhadi]

garage chemist - 5-12-2012 at 13:01

This will not work with a fridge compressor.
If you charge the compressor with 5-10bar on its intake side, it will not exceed 40 bar on the outlet, it is going to stall instead.
I found this out during my experiments with a fridge compressor.
What you need is a scuba breathing air compressor, these are designed to reach the pressures you need and also have a very good oil separator and drier at their outlet.

Oil separation is another issue when using a fridge compressor.
Any oil vapor in the compressed gas will instantly clog the expansion nozzle at -180°C. Complete oil separation from the gas is just as important as drying!

Please spare your fridge compressor this kind of torture and build a conventional refrigeration system with it. You can easily reach -40°C using propane as a refrigerant. Look at my thread:
http://www.sciencemadness.org/talk/viewthread.php?tid=16191#...


babakfarhadi - 6-12-2012 at 10:00

Dear,
Thank you for your reply.I Used this project with Nitrogen cylinder 150 bar (99.9%) and open cycle.Gas first enter to regulator and pressure go down to 40 bar and then enter to inlet pipe ,after process all gas purge to air.My guess is my nitrogen cylinder conatin vapor of water and ice make in expansion device.If have any suggestion plz tell me to re-test it with your idea and post a new data.

Best Regard

wg48 - 6-9-2016 at 08:11

Just got my third old pressure washer. The motor and pump works on all of them. So I am considering using the three in cascade to make liquide air or perhaps liquid nitrogen (less risky). Playing with liquid air or nitrogen is on my bucket list.

The setup it would be something along the following lines.

Air from my garage compressor at 6bar is dried then fed to

the largest throughput pressure washer then to a cooler output at 18bar is fed to the next stage,

second pressure washer modified to only use one cylinder out of three then to a cooler then the next stage at 62bar,

third stage smallest throughput pressure washer using only one cylinder out of three then to cooler then to the next stage at say 182bar,

final stage is the regenerative cooler and J/t expansion.

An other possibility is to modify the manifold of one or two of the washers to run the three cylinders in series with external cooling between each stage.

The problems may be:
the elastomer cylinder seals are running dry, the valves are also made from elastomer, the may not withstand the temperature especially air at temperature.

I will have to arrange the flow at each stage to get the correct compression,

Comments or suggestion anyone?










XeonTheMGPony - 6-9-2016 at 08:16

http://homemadeliquidnitrogen.com/

not my project but a very nicely built unit at the garage level!

wg48 - 6-9-2016 at 12:37

Quote: Originally posted by XeonTheMGPony  
http://homemadeliquidnitrogen.com/

not my project but a very nicely built unit at the garage level!


Thanks I have read that already. I am on a tight budget so no commercial compressor unless I find a scrap one I can fix. It will have to be repurposed and modified pressure washers for me. They are 100bar+ for water.

XeonTheMGPony - 6-9-2016 at 17:56

using one of those pots will work, but you need a way more sophisticated of a system and will take allot longer!

You need a cascade system to cool the discharge, colder the better, my plan was to use two to try and get to 500psi with a refrigerated inter cooler and a colder cascade for the 2nd stage discharge

Air intake/recycler > Compressor > Refrigerated inter cooler @ -40c > Compressor 2 > -80c gas to gas heat exchanger > expander

with aggressive oil scavenging using temp rite 340 series on to molecular sieve final scavenger.

but due to energy input scrapped that idea till I can get a real compressor, but the same system will be used for a natural gas liquefier.