Molecular Sieve Packed Columns and Fixed (Permanent) Gas Analysis

For some reason, I find molecular sieve packed columns interesting.  Maybe it’s because they have a niche as the go-to columns for fixed (permanent) gas analysis.  Maybe it’s because they are old school technology that a new generation of analysts are always asking questions about.  Maybe it’s because using them successfully is a function of skill, art, and some luck.

As I researched molecular sieve packed columns over the years, I realized that for something that has been around more than 50 years, there wasn’t a whole lot of published information about them (at least not that I could find).  Application chromatograms were few and far between, and troubleshooting information seemed even rarer.  As a result, I decided to write this post on molecular sieve packed columns to help our customers.  This information came from books, the internet, coworkers, or was provided by our customers.   If you have any additional or contradictory information, I would love to hear from you.

I have written two other posts on molecular sieve packed columns.  One was titled “ShinCarbon columns – Installation / Conditioning / Helpful Hints”, and the other “Molecular Sieve 5A & 13X packed columns – Installation / Conditioning / Helpful Hints”.   So what’s the difference between these columns?  Let’s start with Molecular Sieve 5A and 13X packed columns.


Click on Chromatogram to Enlarge


Molecular Sieve 5A and 13X

Molecular Sieve 5A and 13X are referred to as zeolite molecular sieves.  These crystalline materials are synthetically produced and have very uniform internal cavities.  I am not going to discuss structural/composition details of Molecular Sieve type A materials verses Molecular Sieve type X materials, but rather focus on helping new analysts understand these packing and their applications a little better.

So what is the difference between 5A and 13X?  Well, as I already mentioned, material structure is one difference.  Pore size is another; a 5A has an effective pore size of 5 angstroms and a 13X has an effective pore size of 10 angstroms.  In order for a compound to be able to pass through these pores, it must be smaller.  For the compounds I plan to discuss (He, H2, Ar, O2, N2, CH4, and CO), each is able to pass through these pores.

So which compounds won’t pass through these pores?  Well, I’m going to answer a different question and list several common gaseous compounds which should not be introduced onto these columns.  These compounds include CO2, H2O, SO2, H2S, Cl2, HCl, NH3 (and other corrosive vapors), compounds larger than (and including) C2 (ethane) and/or compounds with boiling points greater than approximately 50C (please note: there are several high temperature hydrocarbon applications, but I will not discuss them here). Larger gaseous compounds may actually make it through the column, but many will have no retention because they cannot fit into the pores (like SF6 when using 5A).  Other larger compounds may simply not make it through the column.

So what are some of the advantages of using these columns for gas analysis?  The primary advantage is baseline separation between O2 and N2 with a relatively short column and at room temperature.  Another is the separation of CO from air and CH4 (a difficult separation for any non-molecular sieve packed column).

So what are the disadvantages of these columns?  The primary disadvantage is that these packings will trap CO2 and H2O, which is why these columns are typically used with a pre-column (stripper column) and valve switching.  However, if CO2 and H2O do get trapped in these columns, the effects are reversible (see “Molecular Sieve 5A & 13X packed columns – Installation / Conditioning / Helpful Hints”).

Although zeolite molecular sieve packings are commonly used for argon/oxygen separation, a disadvantage of using a packed column is that a very long length (approximately 24ft from what I have read) is needed to separate oxygen from argon at room temperature.  For (O2/Ar) separation, we usually recommend one of our Molecular Sieve 5A PLOT columns (click on chromatogram below to enlarge).



gc_pc1182Finally, one needs to be cautious when using any carrier gas other than nitrogen to dry/condition these columns.  Because these materials contain surface metals that cannot be easily removed, it is believed that other carrier gases like hydrogen, argon and helium (at high temperatures) can remove the oxidation layer on the metal’s surface, which can lead to activity (especially for CO).


ShinCarbon ST

ShinCarbon ST is a synthetically manufactured carbon molecular sieve material with a very high surface area of approximately 1500m2/gm and a pore size of 10-12 angstroms.  Its surface is very inert, and very non-polar.  It is typically used for separation of fixed gases and light hydrocarbons, but also can be used for the analysis of % levels of SO2 (usually elutes after ethane) and ppm / % levels N2O (usually elutes after CO2 but before acetylene).

So how does a carbon molecular sieve column (ShinCarbon) compare to a zeolite molecular sieve (5A & 13X) column?  Let’s start by looking where it doesn’t quite match-up with a 5A or 13X.  First, it can be difficult to get baseline separation between O2 and N2 using a carbon molecular sieve column (customers report that a 10ft x 1/8” OD column is needed).  In addition, it is even more difficult to separate O2/Ar (I have read that cryogenic cooling is a necessity).  And just like the 5A & 13X, a ShinCarbon isn’t meant for H2S or any corrosive gases like Cl2, HCl, NH3.


Click on Chromatogram to Enlarge

So why would you use a ShinCarbon?  First, unlike zeolite molecular sieves, CO2, SO2 and H2O will not get trapped in the pores (Notes: it will be difficult to quantitate H2O because the peak will tail; there are better column choices for SO2).  Plus, you can easily chromatograph larger compounds (up through and including pentane; hexanes will elute but tail).   If H2O does condense inside the column, 30 minutes at 250⁰C will dry a ShinCarbon (unlike the 5A & 13X which needs 300⁰C for 3 hours), and using a carrier gas other than nitrogen does not seem to add surface activity.

To view several ShinCarbon chromatograms using 80486-___ (our 1/8″ OD column), click here Shincarbon 80486  To determine the complete catalog number to purchase a 1/8″ OD Shincarbon column (including the required suffix number), please review this link Things to Consider Before Ordering a Packed Column


I hope that you have found this information useful.  As I mentioned, if you have any additional information, or you had an experience which is contradictory to something here, please let me know.  Thanks for reading.

19 Responses to “Molecular Sieve Packed Columns and Fixed (Permanent) Gas Analysis”

  1. Dear Alan –

    I was introduced to the MS5A and MS13X packed columns back in 1989, when I started working with gas chromatography. At my company we now have 50 GC’s configured with packed columns for gas analysis, including lots of moleculare sieves. Easy to build. Easy to operate. Excellent for separation of permanent gases, also in combination with other packed columns, like the Porapak and HayeSep. Precolumns for backflush and valve options for MS column isolation are mandatory. We actually wash the MS13X material for the RT change of CO and CH4! Total syngas analysis in seven minutes at 40C isothermal are easy achieved, using the nice Porapak-N and MS13X in combination. You can use He/H2, Ar, N2, H2 and He as carrier gas, depending on the application and component response at the TCD.
    From time to time, we think about changing the packed columns to PLOT columns. But we stick to the packed columns, due to larger column capacity on packed columns and lower TCD detection limits.
    I am wondering why we not see any PLOT-N and PLOT-13X capillary column? Are they difficult to build?

    With kind regards –
    Lars Kurstein
    Copenhagen, Denmark

  2. Jaap says:

    Hi Lars:

    The PLOT N is basically a Rt U-BOND. Its a high polar porous polymer and it is available on fused silica nd also on MXT.

    The 13X is used for gases , but for this separation, the 5A can do everything that is required. The 13X only adds only to a shorter retention for CO, which can also, be achived by running the 5A at 15-30C higher temperature

    The 13X has its own unique application for N-P separations of C1-C12 hydrocarbon mixtures.

    I do not know what you mean with ‘Washing” 13X columns. If you hydrate molsieves, they get less polar and CO will come off fast. The column is however containing a lot of water which is a recipie for reention time changes.


  3. Ivy Soon says:

    Hi, this blog is really good that I’m starting to read its contents and learning a lot. This is the closest to what I do and I have a question. I’m have Molsieve 13X with TCD (in tandem with Hayesep N with FID) for TOGA, using direct injection with valve timing and I get a ghost peak after O2, There’s quite a time difference for O2 and N2 RTs and this ghost peak almost always has the same area with H2. I don’t know how to fix it. I used a calibration gas to do test runs. Could it be standard contam? or I have to adjust the valvel timing? Any input would be much appreciated. Cheers!


  4. Alan Sensue says:

    Just to let our readers know, I’ve been working with Ivy for the past few days on understanding her instrument set-up and trying to determine what is going wrong. I believe valve timing is part of the issue, but column selection is another. Her 8ft HayeSep N will not separate O2, N2, argon and CO under normal analytical conditions, but she wants to send CO (and methane) to a FID to obtain lower detection limits (she has a methanizer which converts CO to CH4), and nitrogen & O2 to a TCD. Unless cryo cooling is incorporated, or a different pre-column is selected, I don’t see how she is going to get this to work. My best guess is that she may need to use a ShinCarbon instead of the HayeSep N.

  5. Scott McFarlane says:

    This looks like a good candidate for a simple two column injection with a 10-port. The 10-port can be plumbed with two sample loops to simultaneously inject onto two different column/detectors. MS5A/TCD for O2, N2, and Ar; porous polymer/FID for the CO and CH4. It’s probably not a bad idea to add a pre-column backflush setup on the MS5A channel, which would require an additional valve.

    If this is trace analysis, the ghost peak might be a result of leakage in a valve. I don’t know what the setup is here, but we’ve seen that sort of thing when there is a 4- or 6-port valve in between two columns. Unlikely, since it’s a TCD and such leaks are usually in the PPB range. Typically there should be two ghosts as well, one prior to N2 and one after. Standard contamination can almost certainly be ruled out. Nothing that I’m aware of should be eluting between O2 and N2 on a MS column.

    Hope that helps!

  6. Alan Sensue says:

    Hi Scott:

    Thanks for reviewing my post and the comment. Ivy had sent me a follow-up email showing me a much improved chromatogram. It appears the analysis is working much better. I don’t remember the details, but I do believe there was a small leak which was causing the ghost peak. I do remember that the balance gas was nitrogen, and the carrier gas was helium. I usually caution our customers that when using a TCD, it is usually better to use the same carrier gas as the sample/standard balance gas. If anything, it eliminates one very large, unnecessary peak from the chromatogram.

    Thanks again for the input.


  7. QB says:

    Hello Alan,

    Thank you for your post. It helps a lot for someone like me who is new to packed columns.

    I have a question in terms of Helium analysis by GC-TCD using a packed column. Zeolite molsieve vs. ShinCarbon, which column is a better choice? Currently, I can detect 1ppm Helium in Argon by using a Zeolite Molsieve 13x column. I am wondering if I can achieve lower detection limit with a shincarbon since it has large surface area and slight larger pore size?

    Thank you,

  8. Alan Sensue says:

    Hello QB:

    I am glad to hear that you have found my blog post helpful. Thank you for letting me know.

    I am surprised to see that you can detect 1ppm with a standard TCD. Usually detection limits for these detectors is triple-digit ppm, and not single-digit. I have seen however, that there are µTCDs which have much lower limits (some advertise single-digit ppm levels).

    Since compound detection limit is primarily a function of the detector for this type of analysis, as long as your peak shape looks good, I would not expect you to observe much difference in sensitivity by switching columns.



    Hello QB,

    I am also surprised you can quantify 1 ppm of helium, using a TCD/Argon.

    I analise helium in several gases, by TCD/argon carrier/MS 5A 12’x 1/8″ column.

    Carrier flow is 30 mL/min, 60°C/ loop=1 mL.

    I got a LOQ=4 ppm(2 x noise).

    But, take care: Neon peak elutes very close to helium , and may be mis-identified.

    In my analysis above, RT for Helium is 0,82 min, and Neon elutes at 0,86 min.

    I think Shincarbon column is not recomended for this analysis, because resolution for helium,neon and hydrogen is far less than MS.

    Even MS 13X may be not suitable, because resolution is not as good as MS5A.

  10. Sue says:

    Hi Alan
    Thank you for the helpful hints on Molecular Sieve Packed Columns.

  11. Alice says:

    Hi Alan,

    Really useful blog – thank you. I am relatively new to gas analysis by GC (I previously worked with liquid samples and GC-FID), but am trying to improve our system a bit as we have problems resolving O2 and N2 peaks in emissions from soil samples – I would be very grateful if you have any thoughts. We have a fully enclosed system that enables us to replace the soil atmosphere with a He/O2 (80/20) mix and determine soil N2 production. We currently use a molecular sieve 5A PLOT column, He carrier gas and an HID detector. We do not have any pre-column scrubber installed. Whilst we can resolve O2 and N2, the separation is quite poor and declines over time (presumably due to moisture build-up in the column?) until the N2 peak is eventually absorbed into the O2 peak. Chromatographic performance can then be improved again by column conditioning. It does not help that O2 concentrations are high compared with N2 concentrations. The only other gas we occasionally wish to detect in this (part of the) system is methane. I am wondering whether there is anything I can do to improve things (e.g. perhaps by using a different column) and would be very grateful if you have any suggestions? Please let me know if any further details would be useful.
    Many Thanks,

  12. Nat says:

    Hi Alan,

    In this blog you say ‘So which compounds won’t pass through these pores? Well, I’m going to answer a different question and list several common gaseous compounds which should not be introduced onto these columns. These compounds include CO2, H2O, SO2, H2S, Cl2, HCl, NH3’

    Could you please expand on why these gasses shouldn’t be introduced to the column, in particular I am interested in CO2.

    Many thanks,

  13. Alan Sensue says:

    Hi Nat: I always assumed it was the shape of the molecule that caused it to get stuck (trapped) in the pores of the particle. Maybe one of our readers knows for sure. Alan

  14. Alan Sensue says:

    I had asked Alice for more information, and she sent it to her local Restek Office. It appears that high amounts of CO2 were to blame.

  15. Dave says:

    Hi Alan,

    I am quite new to the use of molecular sieves and want to follow up on Nat’s question (23 Mar 2017) and your response (24 Mar 2017). If CO2 gets trapped, will other gases still be able to travel through albeit slowly, or does the trapped CO2 effectively prevent any other gases from making it through?

    Thanks in advance (and thanks for a very informative blog).


  16. Alan Sensue says:

    Hi Dave:

    Thank you for your question.

    My understanding is that the trapped CO2 prevents other compounds from entering the pores, which reduces the separation power (theoretical plates) of the column. However, I cannot say with absolute certainty that this is what happens. If you need a definitive answer, I suggest discussing this with a materials scientist; I would think they would know, or at least know where to look for the answer.



  17. Dave says:


    Thank you!


  18. Charles says:

    Hi Alan

    Can I use the Molsieve 13X I want to test for H2 in ppm Level my samples will be 99.5% Propylene and Ethylene.



  19. Alan Sensue says:

    Hello: Because I normally advise against putting C2 hydrocarbons (ethane, ethylene, acetylene) onto a 5A or 13X column, I would suggest a carbon molecular sieve column for this particular analysis, like our ShinCarbon. Regards, Alan

Leave a Reply