Nitrogen Carrier Gas for GC – Is it Feasible? – Is it Practical?

Jaap de Zeeuw and I discussed nitrogen carrier gas for GC today, its potential benefits, its drawbacks, and how it might be used “successfully” in more laboratories given the finite supply of helium and helium’s higher costs than nitrogen or hydrogen.  As sometimes happens, we don’t always agree on the finer points, so maybe we can get our blog readers to comment, too.    

I start with a slide I have for my GC training course that indicates some general points about helium, hydrogen, and nitrogen carrier gases, followed by a dimensionless van Deemter plot that demonstrates the performance of those carrier gases in a graphical way.  We want minimum H for the highest efficiency (the best chromatographic separation), which makes nitrogen the most efficient gas, but we also want the highest average linear velocity for the carrier gas while not sacrificing efficiency, which gives the edge to hydrogen.  The other thing we’d like, but has become less critical given that we can operate in constant flow or constant linear velocity mode these days due to electronic flow control, is a relatively flat van Deemter curve.  This allows us to stray outside of the optimum linear velocity and not lose too much of the separation, important, especially if we’re going with a higher linear velocity for faster analyses.  Hydrogen has the edge in this criterion.

Next, I show three organochlorine pesticides separations for helium, hydrogen, and nitrogen on a 20m x 0.18mm x 0.18µm Rtx-CLPesticides column where the carrier gas linear velocities and oven programming rates are close to optimum for maximizing peak capacity.  The peaks from left to right are: heptachlor epoxide, trans-chlordane, endosulfan I, DDE (4,4’), dieldrin, endrin, DDD (4,4’), endosulfan II, DDT (4,4’), endrin aldehyde.  Generally speaking, our expectations are met where we see a faster analysis for hydrogen versus helium, and a much slower analysis for nitrogen.  However, we do get a better separation for endosulfan I and DDE when using nitrogen.  Surprised?

Finally, I had blogged on this subject a few years back and showed that even with nitrogen carrier gas you don’t always have to settle for slow analyses, as long as you have a highly selective GC column like the Rtx-CLPesticides.

12 min GC-ECD analysis of EPA Method 8081 organochlorine pesticides using nitrogen carrier gas

9 Responses to “Nitrogen Carrier Gas for GC – Is it Feasible? – Is it Practical?”

  1. Jaap de Zeeuw says:

    I am not surprised as N2 will be the best concerning plates. However, which lab manager will use nitrogen if the run time is 2.5x longer?
    Also on the example it shows another challenge: we may have better efficiency, but what happens with the response? You loose a factor 2 in peak signal also. Also not very attractive.

    You can use N2 instead of He for simple separations. As soon as separations are getting critical any loss in plates will cost you separation. Users out there will not accept longer run times, they want the same. Nitrogen will get you there, but peaks are broader if same velocities are used. That means we have to live with the consequences of “broader peaks”, no escape.
    Again if you are lucky that you have plenty of separation and clean, defined samples, you can start using N2 tomorrow. If not, you will experience that it take less injections to make efficiency drop.

    Other then that, If one is willing to change ALSO the column diameter, one can compensate for the analysis time and for the loss in sensitivity. That may be the best approach for applications where we need our efficiency..

  2. Jack Cochran says:

    Thanks Jaap for the good discussion. As we discussed in e-mail today, this is all relative. Someone who only has 20 samples per week (or month) to analyze may not care if the run time is 2.5x longer. And, one thing we always forget is that sample preparation is the big time bottleneck anyway. it takes longer than 48 min (in fact, it can take several days) to generate one EPA Method 8081-type extract for analysis.

    The chromatograms shown are “translated” to show the differences in analysis time and separation, but as noted in the blog, I published a 12 min separation for organochlorine pesticides using nitrogen carrier, which is much shorter than some analysts already do, and means that we can preserve relatively narrow peaks that don’t cause the detectability issues you mention.

    I agree with you that most analysts won’t agree to go 2.5x slower, but confusingly, those can be the same analysts who you can show how to go 2.5x faster and they don’t want to do that either! Look at the US aversion to using hydrogen carrier, which provides speed and solves the helium problem for many separations (GC-MS with hydrogen is another story). I believe this is the C term in the van Deemter equation!

  3. We’ve done quite some nitrogen implementations recently, due to inconsistent helium deliveries with several of our customers in Belgium and Holland.
    You can take a look at these results at:
    Some fruitful discussions with Jaap in this respect can be found here:


  4. Aviv Amirav says:

    According to the theory the sensitivity loss is not linear with the peak width but it goes down according to the square root of the peak width decrease.
    Thus, if the vendors will allow as they should to have flexible FID averaging time contstant than while the peak height will go down with the width the noise will also go down by the square root of the peak width since longer averaging time constant can be selected and the S/N will go down only by the square root of the peak width increase. With GC-MS one can reduce the scan speed linearly with the peak width and get exactly the same signal in terms of number of ions per scan plus the same column related noise such as co-eluters or ghost peaks and only the vacuum background will increase by the square root of the peak width. However, N2 is bad for the sensitivity of GC-MS due to ion source space charge and hydrogen has other issues from safety to ion source activation.
    Aviv Amirav

  5. Jack Cochran says:

    Thanks for the great (and educational comment), Aviv, especially as regards alternate carrier gases for GC-MS.


  6. Michele Sanders says:

    We are in the process of looking for alternative gases, since our Helium supply will become unreliable. However, one of our chemists indicates that we must use the more expensive Argon, since Nitrogen would burn out the GC/MS filaments quickly. Is this true?

  7. Jack Cochran says:

    Nitrogen is not preferable as a GC-MS carrier gas for two main reasons: it is very slow, chromatographically speaking (but efficient!, i.e. good separating power under optimal linear velocity, which is slow, but I said that already), which means separations take much longer. That said, if you are not doing isomeric or isobaric compounds and you have a mass spectrometer as your detector, you can live with a faster, but less efficient, nitrogen carrier analysis, PERHAPS. I say that, because the biggest reason we don’t use nitrogen as a carrier for GC-MS is that MS sensitivity is seriously compromised. If you are not doing trace level work, perhaps nitrogen is still an option.

    While I haven’t used argon personally as a carrier for GC-MS, I would expect that MS sensitivity would suffer similarly to nitrogen, and the separations would be even slower yet. I am not sure about MS filament lifetime for argon versus nitrogen. Although filament cost is not trivial, I think you have to weigh the analysis time and sensitivity issues first before considering alternate GC-MS carrier gases.

    Let me know of your further experiences, please.


  8. Jack Cochran says:

    Absolutely a great question, Ben! The easiest way to shortcut this for you and other blog post readers is to recommend the Method Translator that was formerly on Agilent’s website. For some reason, they seem to be making it harder to get these days, but I found it at:

    The critical things we do when translating a method for a different gas where we want to maintain our previous separation is (1) establish the comparable (not the same!) linear velocity for the new gas, (2) adjust the GC oven program rate such that compounds elute at the same temperature as before. When going from helium to hydrogen, this is going to be a faster separation for the same chromatogram, and when going from helium to nitrogen, the separation will be slower. BUT, the chromatograms, if you didn’t have the time scales, would look almost identical! Get the Method Translator and try it.


Leave a Reply

Restek Domestic Customer Service



Your Full Name

Your Email

Company Name


Spam Block (Please leave this blank)

all fields required

Thank you

Your message has been sent. We will be in touch shortly.

Message not sent

Sorry, your message could not be sent at this time. Please try again later, or contact Restek or your local Restek representative via phone.