Rising Baseline in Gas Chromatography

[kristind858]

Hello all,

I was wondering if you can guide me on the first thing I should try to solve the following: Our GC carrier gas (He) was too low to maintain the pressure set point therefore causing the front inlet to shut down. I installed a new tank and I loaded up one of our GC test methods to allow the detector to stabilize for about 4 hrs. When the samples chromatograph, the baseline begins to rise with increasing temperature. Should I clean the detector first or cut the ends of the column and rinse with solvent? Do you think the stationary phase is decomposing perhaps? I am using a ZB-wax 100% PEG column. Temperature program is 170 deg C for 0 min, then 0.6 deg C/min to 207.5 deg C for 0 min, then 5 deg C/min to 225 for 0 min. I have tried a bakeout with the oven temp to ~10deg below column max temp, inlet @ 325, & det @ 325 for a little over 2 hours. I took over for another chemist so I am not that familiar with how to troubleshoot.

Also how to we add an image to a post. I want to add a chromatograph so you all can see Thanks !!!

[Don_Hilton]

How hot was the GC oven with no carrier flow after the inlet shut down? And, how long was the column without flow.

Wax does not like the mix of elevated temperature and air - it begins to unzip.
And once the polymer begins to depolymerize - conditioning the column does not help. The hotter the column was and the longer the column was without flow, the more likely that the column is toast.

As long as the detector is set at a reasonable temperature (sufficiently above the oven tempreature to avoid condensation) there should be no change in signal resulting in anything from the detector as a function of oven temperature.

Having destroyed a wax column or two along the way, my first reaction is to suggest installing a new column and see if that fixes it. If you put some more details - instrument temperatures, type of detector, inlet type and temperature - we might be able to make better guesses.

[kristind858]

Thanks Don for your reply !! I'm not sure how long the oven was on for and at what temp I came back for the holiday weekend and saw the message. Let me check the logbook to see the timeline of events . I did a bake out of the column with oven temp ~ 10 deg below the column max, injector and detector at 325 deg C for two hours. After, I loaded up our testing method and made one standard injection consisting of EPA, DHA , and C23:0 ISTD. All three peaks were present but the peak areas were way off since system was integrating below the rising baseline. The next day, I injected another standard sample and slightly lowered the Helium flow from the second stage gage. Again the rising baseline was present but the peak areas were very close to values obtained before this issue occurred. The baseline rises by about 2 pA towards the end of my temperature program. Should I try adjusting the y-axis scale ? If so, how ? I'm using chemstation. Since increasing temperature causes an increase in gas viscosity, did my adjustment of gas flow help or is this regulated within the GC system itself ? I did a total of three standard injections the day after the bake out and all three of the peak areas from each injection were nearly the same as what I got when things were running well. Can someone explain what might be happening ? Also what forum am I allowed to post images in ? I would like to attach my chromatograms Thanks for all your help !!!!

[Don_Hilton]

Do not lower the pressure at the second stage of the regulator. You need to have that pressure set at least 20 lb above the highest hed pressure needed in your GC method. The messaged indicates that you have an electronic flow controller - and that regulates the pressure inside the GC. If the pressure to the back of the GC is too low, it will shut down on you again.

For posting possting chromatograms, there are instructions on the site - it may be on in the LC discussions? Look around. it is at the top of the list of one or another set of discussions.

[Peter Apps]

A 2 pA drift with a wax column does not sound too bad. How high are your peaks in pA ?

Peter

[kristind858]

Hello Peter !

The baseline initially starts out around ~10 pA then with the temperature program as stated above rises actually more like ~3 pA by the end of the run. The peak signals are as follows: EPA~30 pA, C23:0~17pA, and DHA~23 pA. Why is it that on my last three standard runs the peaks areas are similiar to what I got before this shutdown? I will look for a forum that will allow me to post my chromatograms My y-axis, by default, goes in 1 pA increments. How can I adjust this ? The baseline may actually appear more linear if I do so. Thanks for your help !!!

[Peter Apps]

You need to look at how the data software is integrating the peaks. Depending on settings the integrated baseline will either follow the actual baseline and the amount of drift that you have will probably not affect peak areas, or it will at some point drop the integration baseline below the actual baseline, and the peak areas will then be larger than they should be.

First step is to choose the option in you software that allows you to see the integration baseline. Then, if it falling below the real baseline you need to change integration settings until it tracks the real baseline. Try selecting valley to valley for the baseline. If you do not know how to do this then use the software help.

Peter

[kristind858]

Figure 1: Before bakeout. Isooctane with BHT. Temperature program: 170 deg C for 0 min, then 0.6 degC/min to 207.5 deg C for 0 min, then 5 deg C/min to 225 deg C for 0 min. Column: ZB wax 100% PEG.
______________________________________________________________________________________________________________________

Signal: FID1 A, Front Signal

RT AREA NAME
37.241 294.5889 EPA/DHA Std-#1
46.667 528.2255 EPA/DHA Std-#1
56.836 3321.5303 EPA/DHA Std-#1

Sum 4144.3447

Figure 2: Before bakeout. EPA, DHA, and C23:0 ISTD. Temperature program: 170 deg C for 0 min, then 0.6 degC/min to 207.5 deg C for 0 min, then 5 deg C/min to 225 deg C for 0 min. Column: ZB wax 100% PEG.
______________________________________________________________________________________________________________________


Figure 3: After bake out (Oven temp 240, inlet @ 325, & det @ 325). Chromatogram of Isooctane with BHT. Temperature program: 170 deg C for 0 min, then 0.6 degC/min to 207.5 deg C for 0 min, then 5 deg C/min to 225 deg C for 0 min. Column: ZB wax 100% PEG.
____________________________________________________________________________________________________________________

Signal: FID1 A, Front Signal

RT AREA NAME

36.586 290.2884 EPA/DHA Std-#1
46.032 73.2491 EPA/DHA Std-#1
56.018 187.3622 EPA/DHA Std-#1

Sum 550.8997

Figure 4: After bake out (Oven temp 240, inlet @ 325, & det @ 325). Chromatogram of EPA, DHA, and C23:0 ISTD. Temperature program: 170 deg C for 0 min, then 0.6 degC/min to 207.5 deg C for 0 min, then 5 deg C/min to 225 deg C for 0 min. Column: ZB wax 100% PEG. Peak areas have greatly improved from those in Figure 2. The peak area values are similiar to the peak areas I was getting before the shutdown occured.
______________________________________________________________________________________________________________________



Figure 5: Blank run-No injection. Temperature program: 170 deg C for 0 min, then 0.6 degC/min to 207.5 deg C for 0 min, then 5 deg C/min to 225 deg C for 0 min. Column: ZB wax 100% PEG.

_____________________________________________________________________________________________________________________


I disconnected the column from the detector and set detector at 450 deg C and oven at 230 deg C. Baseline was stable at 0 pA. Anymore advice on what is going on here Thanks for your help !!!!

[Stunt]

Kristin, that looks like a leak to me. Are you using the correct detector ferrules with the proper amount of torque? Although if your system had no flow through the column while the oven was on, you may have destroyed your phase/column.

[gserrano_84]

It looks like a degraded phase to me. Specially because of that jump in the baseline after you ramped it to 225 C. Wax columns are highly susceptible to oxidation (well, most columns anyway), and any leak will degrade the phase faster. I don't recall seeing a baseline drift like that in a wax phase after heating it to 225 C (maybe after 270 C). Since you mentioned that the column was heated with no flow, then, more reasons to suspect that it's a degraded wax phase

Gustavo
Supelco

[kristind858]

I haven't changed any of the ferrules. So the correct ones were previously installed by another chemist. Is this something I should change? Should I try cutting the ends of the column? Thanks !!

[Don_Hilton]

Graphite or graphite/vespel ferrules need to be checked and periodically. The heat cool cycle of the GC oven squeezes on the ferrules and it it is deformed. I have heard it jokingly put that the difference between graphite and toothpaste is the color.

The column nut should take a little bit of pressure on the end of the wrench and not move. Note - a little bit. If you try to tighten it down as tight as you can make it go, you will break something.

A check with a leak sniffer is helpful to know if the ferrule is sufficently tight at the inlet end. At the detector end, it will not be as sensitive to a leak, but it you can develop a feel for the pressure needed to keep the ferrule tight - be sure the nut feels tight enough and check it with the leak sniffer anyhow.

Looking at the traces, the baseline is flat enough to use. As long as you get good separation in samples - it looks to be in good shape - bleed or not. Integration parameters do need to be adjusted so the drawn baseline looks like a continuation of the baseline that shows between the peaks.

[kristind858]

I will check the inlet ferrule when I get back to work Monday I was speaking with tech support earlier and they wanted me to disconnect the column, insert a ferrule with no hole at the detector end and run my method. They want me to do this before I go ahead and switch out the column. If the detector is stable and I check the inlet ferrule to ensure no degradation and/ or replace it, that leaves mainly the column as the source of the issue ? Can a cut the ends of the column? if so, how much do you recommend ? I wouldn't need to replace the whole column right ? My peak areas of the standards are similar to what I was getting before the issue. Or is the column just done with ? All of you have been so much help !! Thank you Peter, Don, Gustavo, and Stunt !!

[kristind858]

The integration follows the rising baseline as seen in Figure 4. I clicked on show integration and integration line matched with the rising baseline.

[Don_Hilton]

I've run samples on columns that have looked worse than what you have shown (it may be that I shouldn't have, but...). If check standards or check samples (controls) are within tolerance and RSDs are within the expected range - you can continue. Note that check standards/samples should be across the expected range and in matrix - which is proper form for any method anyhow.

It would be nice to see a flat baseline. Typically the front end of the column is trimmed. That is where junk from injections sticks and where oxygen first hits the column. But trimmign the column does not always work. You have to try it to see. In some applications this will significantly increase the life of yoru column. In other applications you may get nothing - or extend the life of the column by only a week. Yes columns are expensive, but how many hours can you spend on "saving" a column before you have spend the price of a new column?

And when you cut the end of the column, all the retention times will move. They will probably move a bit with a new column, but the effort in adjusting the method for the larger change in retention time may be a consideration.