Nitrogen Separator / Generator

Using simple parts and fittings to create a reasonably capable nitrogen source.

Similar projects worth following
Due to a need for lots of pure nitrogen gas for another project, it was decided to go about building a nitrogen separation system. This device is designed to compress atmospheric air, dry and filter the compressed air, and then separate out the nitrogen gas from the remaining atmospheric gases (mostly oxygen and argon).

General Details and Concepts:

The main concept behind the separator is to force filtered air through a nitrogen separation membrane, which is simply designed to separate atmospheric oxygen and argon from the nitrogen in the air. The rest of the project is consistent of gas compression, regulation, and filtering. 

The nitrogen gas being generated in this project needs to be mostly pure along with being clean and dry because it will be used in a nitrogen liquefier. Any water of oil contaminating the gas can freeze in the liquefier and clog pipes. The retaliative purity of the gas being generated can be adapted for different types of projects.  

To obtain dryer gas, add more Oil/Water separators and desiccant dryers. To obtain less pure gas, use a less accurate nitrogen separation membrane (in theory the only reason you would do this is to reduce cost). 

The whole project is deigned to be built with parts costing less then 400 dollars, but that depends on the availability of certain higher priced items. For example, I already had a old compressor that I wasn't using, so I chose to use that instead of buying a new one. A further note on the compressor, one can choose to use their shop compressor rather then a dedicated one, just keep in mind the noise level. Due to the slow output rather, the separator can be left on anywhere from a few hours to more the several depending on the amount of nitrogen needed. With that rate, a shop compressor might be going off every 20 minutes or so. Another option is to use a refrigeration compressor instead but this requires more filters and an oil return line due to the lubricant usually being mixed with the refrigeration gas. 

The project logs below will outline the building aspect of the project as it is built. Details for the design are directly below.


 The flow of the working gas through the system is as follows:

  1. The gas enters the compressor through a air filter such as an automotive air intake filter. This removes any large particulate from contaminating the air source and helps extend the life of the compressor.
  2. The compressed gas is then stored in a tank. This not only creates an available reservoir for the system, but also cools the gas after compression to prevent it from melting any of the regulators or filters.
  3. The gas then reaches a ball valve that controls the input to the separation system. This allows the user to quickly turn off the gas supply without having to lower the regulator to zero.
  4. After the valve, the gas enters a pressure regulator to regulate the compressed gas to around 80 PSI.
  5. The now regulated gas enters two "wet side" filters: a particle separation filter, and a oil/water separator. These are labeled as "wet side" filters because at this stage the gas has yet to enter the next stage: the desiccant dryers. 
  6. At this stage, the gas is dried in two cylinders containing a rechargeable desiccant such as molecular sieves or silica pellets. This removes a majority of the moisture from the gas.
  7. The gas then enters a single "dry side" filter: the coalescing filter. This removes the last of the oil and water that may have eluded the previous filtration devices.
  8. The gas now is regulated by a flow valve where the flow is lowered to a rate suitable for the nitrogen separation membrane.
  9. It now enters the nitrogen separation membrane where the oxygen and argon mixed in with the nitrogen is removed and expelled from the system. 
  10. The pure nitrogen gas now reaches its last stop, the output valve. Connected to the other side of the output valve is the user's storage tank of choice.

  • 1 × General Air Compressor Any store brand compressor will do as long as it has reasonable CFM.
  • 2 × Ball Valve 1/4" NPT (F-F) McMaster-Carr: 47865K21
  • 3 × Pressure Gauge 2" Dia. 1/4" (M) (0-200PSI) McMaster-Carr: 4089K61
  • 3 × Tee-Connector Brass 1/4" NPT (F) McMaster-Carr: 4429K251
  • 2 × Brass Pipe Nipple 1/4" NPT 4" Lg. (M-M) McMaster-Carr: 4568K137

View all 20 components

  • Initial Testing

    Leo Mahdessian08/15/2017 at 05:19 0 comments

    After receiving most of the filters and fittings, I connected everything of what was available to test for air leaks and filtering capability. At this time the only things not tested were the separation membrane and the molecular sieve dryers, both of which were not tested because I did have them yet. All the fittings and connectors are 1/4 inch NPT to keep things simple. Make sure to use thread tape or some sort of thread sealing compound. NPT was designed as an interference fit and it was designed to work only with pipe sealant. I used thread tape with about 3-4 wraps. The connections between the parts that are further away are done with 37 degree flared fittings for stainless steel tubing. I chose to use stainless steel because it has a better tolerance for lower temperatures and because it was cheaper then copper tubing.

    When I was done connecting up all the parts, I filled the system with air to approximately 40 PSI to test for leaks. I plan on leaving the system pressurized overnight and checking to see if the internal pressure has dropped at all. 

    Possibly the hardest component to procure in this project is going to be the nitrogen separation membrane. I'm hoping to get lucky and find one cheap on eBay, but otherwise I've been scouring the web for any supply of the membrane.

    The next step is to glue up the PVC chambers for the molecular sieve dryers and thread the flared fittings onto the end of the pipes. Hopefully the stainless steel tubing should be here by the end of the week which will allow me to test the functionality of the dryers.

  • Storing the Generated Nitrogen

    Leo Mahdessian08/05/2017 at 19:47 0 comments

    The current plan is store the produced nitrogen in a high pressure cylinder similar to what one would buy from a gas supply company. To do this, some sort of reservoir is needed on the output to create a buffer of gas that the high pressure compressor can empty. Without this, the high pressure compressor will pull a vacuum on the nitrogen generation system causing possible system failure or critical failure resulting in injury of those near by. Furthermore to prevent the high pressure compressor from creating a vacuum in the reservoir after it is emptied, a pressure control valve will need to be installed to control the compressor. 

    One does not have to store the produced nitrogen at a high pressure but for certain applications it is more efficient due to the increased amount of gas stored. In one scenario where high pressure storage is advantageous is in the production of liquid nitrogen. Due to the large difference in densities between liquid and gaseous nitrogen, a large amount of nitrogen makes only a small amount of liquid nitrogen.

View all 2 project logs

Enjoy this project?



jrball wrote 05/19/2021 at 16:42 point

So this isn't going to generate 5 9s purity N2.  It looks like 95 to 99.5 N depending upon the Pressure supplied/ and generation rate you need.  So you need to work backward to size the membrane based on your purity and volume before you size all the other components. 

  Are you sure? yes | no

jrball wrote 05/19/2021 at 16:31 point

Here is a link to a N2 membrane mfg.

  Are you sure? yes | no

Lars Herrnsdorf wrote 03/02/2021 at 07:37 point

Hi Leo

I have a need to use small amount of N2 flow in a reflow oven, I have already access to clean , dry air  but lack the  nitrogen separation membrane. Did you find any source for that ?

And how far have you got in the project

Best Regards Lars

  Are you sure? yes | no

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates