GC baseline problems - Agilent CP 7420 column

Hello, I have a Agilient CP7420 column connected to the GCMS-QP2020NX also equipped with a FID detector and the column is connected to the FID (carrier gas helium 1mm/min). The column and the chromatograph are new, but the column was purchased from 3 years ago but was not unpacked

In terms of FAME separation, I am very pleased with the column. However, I have some problems with an unstable line at higher temperatures which generally make it very difficult and sometimes impossible to quantify the analytes.
Above 175C, irregularities begin which, with the appropriate enlargement, look like 'saw blades'.
Basically, all procedures related to the chromatograph were performed (column conditioning for many hours at different temperatures, cleaning of the injector and FID, the split filter was replaced, the silica gel in the hydrogen generator was dried, the septa and liner was replaced, the column was cut several times. and reinstallation, filters are installed on both the carrier gas and the detector gas lines).

The manufacturer's service after performing all the above-mentioned steps found that these instabilities came from the column. My question is whether it is possible to deal with the fluctuations attached in the picture or Is it just the column's fault? Or maybe the injections of various solvents, such as methanol or acetone, can help. so far, n-hexane has been used

I would be grateful for any help and suggestions

  • Hi Domar,

    this is textbook behaviour.

    It is not related to the specific part number. What you see is the typical cyanopropyl-bleeding pattern of any stationary phase containing cyanopropyl groups. Normal bleeding is mainly hexamethylcyclotrisiloxane, a Si-O- six ring with 2 methyl groups on each Si (we call it D3). If you have a cyanopropyl phase one or more of the methyl groups is replaced by cyanopropyl groups (statistically the same percentage as cyanopropyl in the phase), the more CN-C3 the higher the molecular weight of the bleeding product and the lower is its volatility.

    Bleeding starts above 200 °C and you normally can see a rise in baseline. At the end of your temperature program you cool down (in most cases) ballistically, trapping the bleeding products in the phase until the next time you are above 200 °C. Because these CN-D3s are big they have quite a long retention and don't travel immediately out of the column, then you cool down again and trap them again. With your next run the game starts again, that gives you the toothy look. 

    The only possibility to get rid of the problem is giving the bleeding products enough time to leave the column after your run by temperature-controlled cool-down. Cool with (-2) to (-5) °C/min to 200 °C at the end of a run.

    For conditioning go to Tmax for 30 minutes and then cool down with -2 °C/min to 200 °C and then ballistically to your starting temperature. If it persists keep the column 60 minutes at Tmax. You need to experiment a little as the worst case is hexa cyanopropylcyclotrisiloxane, a molecule with molecular mass of 524 Da and a very long retention time.

    Hope this helps



  • Dear Dr. No, thank you very much for your answer and your time.

    Regarding CP7420 is 100% bonded and cross-linked cyanopropyl column. I will try your advice and write back if it helped. I already read somewhere that the cooling rate of the column can affect the baseline noise, and therefore I reduced it from the system setting to 15C / min but as you can see it was still not enough

    If you are so kind and clear up a few more things about high-polar column:

    - Do you think that the high-polar columns temperature from which the analysis is started may be of importance in GC? in my case, I start from 140C, using helium as a carrier gas, the analyzes take quite a long time, so I try to shorten them by optimizing the oven temperature, mayby
    Is it more advantageous to start the analysis e.g. from 60C (i.e. cool the column after each analysis to 60C) it would have some meaning
    - are there any rules for the rate of temperature increase in the column oven temperature program, e.g. not more than e.g. 5C / min.
    - the time of column equilibration before anlysis start may be of some importance? is it 1 to 2 min ok?
    - May carrier gas flow rate some influence on baseline stability with cyano phase columns? e.g He column flow  at 1 ml / min would be ok? however, in my case when using linear velocity as a carrier gas flow control mode at higher temperatures the flow clomn drops somewhat to about 0.9 ml / min.
    - Could  the FID temperature  be higher than the maximum operating temperature of the column. Opinions on this issue in the Internet are divided

    Best regards,


  • Hi Piotr,

    I assume to separate FAMEs. We saw that 0.5 °C change in elution temperature of a component or critical pair can make or break the separation for this column. 1 °C can switch critical peaks.

    Temperature rates for some instruments go up to 100 °C/sec and it works for some components, but might not for others. The advantage is less diffusion and therefore less ban broadening, but you will be out of the optimum in the Van-Deemter curve immediately and Nth goes down giving you more band broadening. Finding the right path here is the art of GC.

    Equilibration time: as long as you have a stable baseline go for it.

    Bleed is a function of column flow, twice the flow = twice the amount of bleeding, it still can be a stable baseline. It's more difficult when you use flow control in the GC. The flow is calculated, NOT measured and you use nominal column dimensions and not the real ones, with that you can have errors up to 20+% in flow.

    Rule of thumb for the GC detector is at least 20 °C higher than the max. oven temperature used. The same or lower temperatures can lead to condensation in the FID. So yes, the FID can have a higher temp than the Tmax of a column. I use 250 °C as FID temperature as a default, when high boiling compounds are present I increase it to 270 °C or 300 °C, that's normally enough. You will always destroy a little amount of phase at the end of the column (some centimeters).

    And finally speed up a GC analysis: better go for smaller IDs and the equivalent column dimensions than just use temperature. Temp is fine if you have a non-critical separation. Also here some experimenting helps. And use hydrogen as carrier gas if possible. Don't be afraid of hydrogen, it's at least twice as fast as helium and has a very broad optimum in the Van-Deemter curve (even with a little loss of plate count).

    And sometimes you really need 100 m of column if the analysis is very complex.

  • Hi Norbert,

    Thank you very much for your extremely valuable tips

    Kind regards,


  • Hi Norbert,

    I have tried your advice and there is slight improvement but generally only when 10C / min  cooling rate is used for each cooling cycle and preferably over the entire temperature range (e.g. 60-250C) the rate of 15C/min is already making the situation worse.

    However, I have another kind of problem, related to the low repeatability for early eluting FAMEs (first 5-6) between repetitions there are large discrepancies in their percentages, e.g. for C4:0 bresults etween repetitions range from 0.7% to 2.0%; for c6:0 from 0.6% to 2.0%.

    The columns are attached to the OPTIC4 GL Science injector. The syringe was replaced, great attention was paid to aitosampling, the liner and the spilite filter were replaced, but the problem remains,

    The chromatograph service claims that it is a problem with the column,

    Have you ever had such a case? can something really be with the stationary phase?

    Kinds regards


  • Hi

    Can you little bit trim the column from both end, means inlet and detector side then then condition column least 3 hour then after check...

  • Of cause the instrument guys tell you it's the column, very common method to get rid of questions. If they mean it's the method, then there might be truth to it.

    Trimming is standard process, but won't help.

    Just think about what's happening, the low boiling components with the least interaction time with the column show the deviation, the larger and higher boiling ones do not. That means you might loose the low boilers during injection.

    I'm not familiar with the OPTIC injector, but if there is some kind of "septum purge" it might influence your analyses. Low boiling components can be lost more by septum purging than high boilers. Same is true for split flow loss as the pressure build-up is the highest at the start of the injection and less later when the high boilers evaporate.

    Anyway you need to optimize your injection.



  • Thanks for the tips, I'll be in touch soon and tell you about the progress :)



Was this helpful?