gc tailing and bumps

[ece]

Good Afternoon,

I have been having the same problem with my GC for the last couple of weeks, and I was wondering if anyone on here can help me resolve it.

I am using an Agilent 5890 using helium (30 cm/sec) for biodiesel analysis (ASTM 6584) of mono, di and tri glycerides to determine conversion.

A little background...

We took our GC out of commission for several months. We left the column in place as it was. Recently, we recommissioned it.

We were consistently seeing lots of tailing and a relatively high baseline on our runs, with a bump in the range of 18-23 minutes during the run. We were using a Varian Select Biodiesel column, .32 mm x 16 m long with a 2m metal guard/retention gap .53 mm. I addressed the tailing issue by doing the following, one run at a time.

(prior to first run)

Installed the carrier gas with the appropriate oxygen/moisture trap
Checked the entire system and carrier gas supply lines for leaks
Ensured the linear velocity was adequate on the carrier gas (30-35 cm/sec)
Purged the column overnight and kept the detector at 200 deg C

(prior to second run)

Cleaned the inlet, the detector jet, the collector assembly
Replaced the septa

(prior to third run)

Used a brand new Agilent syringe
Prepped new standards and resilyated (1mL ampoules)

(prior to fourth run)

Made sure the inlet temperature was tracking the oven temperature (this helped significantly)
Trimmed the guard and column on both ends by about 4 inches.

At the end of all that, I determined it was highly suspect that the column was contaminated. I wanted to get a column rinsing kit from restek, but since we had no vacuum or no way to set it up we just replaced the column and guard.

So, new guard and column arrive from Varian (same spec as above). I did the following to commission the new column:

Put the column nut on the guard inlet end
Scored guard at inlet end
Installed to inlet
Replaced septa
Put union and ferrules on guard to column connection
Scored the guard on that end (by about 1 in)
Put union and ferrules on column to guard end
Scored column on that end (by about 1 in)
Overlapped the guard onto the column by about 1 mm
Tightened the union
Put column nut on the detector end
Scored the column end to the detector
Installed the detector end to the column
Replaced the septa on inlet side

then I turned the EPC to the unit on and

Checked the system for leaks
Purged it for 2 hours (that was what Varian said to do)
Conditioned for 16 hours (that was what Varian said to do)
Silyated calibration standards with brand new, Sigma Aldrich ampoules of standard solutions 1-5 and brand new, Sigma Aldrich internal standards
Blanketed everything I silyated or opened with Nitrogen

After doing all this, I am still seeing the same exact problem. A little tailing, a slightly high baseline (consistent throughout the run) and a bump I keep seeing throughout all the runs in a range from 18-23 minutes into the run. I am not sure why I keep seeing this bump, I even used brand new internal standards, installed a new column, as by the book as possible.

Attached at the bottom is a sample of my chromatogram which is consistent with all the problems I am seeing. This particular run is of a calibration standard solution 1.



The only other thing my boss and I can think of might be the problem is that, when he installed the helium gas lines, he used standard copper tubing from McMaster-Carr, and he believes that since we don't have a hydrocarbon trap that could be a possible source of what's happening.

I also believe it is possible that, due to the fact that the GC was out of commission for more than a year, that maybe water or dirt or something got onto the detector (though I'm sure we'd see much more problematic, erratic baseline noise).

Other than that, I'm completely stumped. Any comments?

[Bruce Hamilton]

What happens to the bump if you perform multiple injections of blank solvent after having conditioned the column?. Does it change in size?.

The tailing could also be attributed your detector gases/temperature or injector. I suspect the latter still has more activity than is desirable - possibly from historical gunk. Assuming your gases are OK, it's unlikely you've killed a new column, so I'd focus on the instrument.

It's possible that the copper tubing has some oil that has migrated, but I'd expect the problem to occur not long after initial commissioning, so I would work on the instrument first. If concerned, check to see if GC has Molecular sieve traps on carrier inlet inside instrument and replace those.

The good news is that your first column is probably OK.

Please keep having fun,

Bruce Hamilton

[Peter Apps]

Great - a post with enough detail to troubleshoot sensibly

What do you see with a temperature programme blank (i.e. no injection) ? If the hump is still there it is either carryover, or bleed from somewhere. If it gets smaller with repeated programme blanks it is probably carryover. If it is not there with a programme blank it is coming in with the sample.

When you cleaned the inlet did you replace the gold seal ?

Check the septum purge flow, and if possible increase it.

Peter

[Don_Hilton]

If you got copper tubing with oil in it, it can plague you for some time. A run without sample injected should still show the hump and high background - as noted above.

When there is oil in the gas lines, you can solvent clean the gas lines and bake them out (to get rid of residual solvent) - or obtain GC grade tubing and replace the lines. It will take a while for the plumbing in the GC to come clean - and you may need to replace some parts if the contamination is too severe. Unfortunately the heavy components in the oil migrate slowly -- like in chromatography.

Having been in a lab where contaminated tubing was a question (fortunately not on my instrument) I would immediately change out suspect tubing.

[ece]

Hello everyone!

Thanks for the responses. I have here an image of a blank run that I did (which is just heptane). As you can see, the bump is still there.

http://www.flickr.com/photos/39750479@N05/3656847973/

Then I did a dry run, where I didn't inject anything at all but just ran the column at the temperature selected. This was the result of that.

http://www.flickr.com/photos/39750479@N05/3657738104/

It simulates some kind of incomplete conditioning or something, but I don't see the characteristic bump. Does this mean the column, gases, inlet and insert are not the problem? It must be a) heptane i've been using b) syringe ?

So then I used brand new heptane (shipped this morning) without the bottle I've been storing it in and poured it into a clean beaker, rinsed the beaker with heptane a couple of times, and poured fresh heptane into an empty sample vial. This was the profile that I got (same deal)

http://www.flickr.com/photos/39750479@N05/3657984938/

Pending that, new syringe? Gotta wait til later on this week to get new one. If so, how can I prevent the syringe from getting contaminated? Usually I just keep it in the original box, but rinse it several times with heptane before re-using it.

As for the gold seal, I have never seen nor heard of one of those unless you are referring to the split/splitless inlet gold seal. We don't use that, we just use COC injection.

[Bruce Hamilton]

OK, My suspicion is that the nC7 is washing junk from either the injection system, or the column. I'd suggest performing a few larger-volume injections ( or decreased split ratio - if used ) to ascertain if the hump area increases.

I'd also perform a few injections in rapid succession to ascertain if the peak area diminishes. My guess is that your injection system is still contaminated.

I would want to thoroughly flush the whole injection system with 1:1 warm chloroform:methanol to remove all traces of lipids residues - especially the lines to the controllers and vents. I'd also replace/clean and re-deactivate the inserts.

I don't think is oil from gas lines ( did you look to see if the instrument had HP Molesieve traps on the carrier lines inside Instrument LHS cover?), or from the syringe. It's possibly from the septa, so ensure you use a good qualilty product.

The problem may be that you have moved some to the column, but I'd use a mixture of repeated large injection volumes and warm ( not Final temperature ) oven conditions to confirm the column has junk on it.
If the junk is lipid-derived, high temperatures might just cause it to degrade more.

Peter and Don may have better suggestions than mine.

Please keep having fun,

Bruce Hamilton

[Don_Hilton]

The run with no injection looks like it takes the issue of contaminated gas lines off the table (this is a good thing!). I agree with Bruce on the approach of cleaning the front end of the system. I would even take a cotton swab with the solvents to the inside of the inlet and scrub a bit (gently). Be sure that you get any fibers from the swab out of the inlet - or you add to the problems.

I doubt you have a contaminated syringe. A good cleaining with several solvents (like methanol, acetone, methylene chloride, in sequence) would clean just about anythign out of it. And it may be worth a try.

[Peter Apps]

One more thing to eliminate is the possibility of contaminated vials. Rinse the vial a few times before filling it (like you did with the beaker), and if your autosampler will tolerate it, leave the septum cap off, or take the septum out of the cap. You could also try a manual injection direct from the beaker.

Peter

[ece]

Thanks for all the comments, i'm sorry i forgot to clarify the following:

1) we have a COC direct injection port - no split/splitless
2) we don't currently run an auto-sampler
3) we use an OMI tube on our carrier gas (helium) that we vigorously checked for leaks everywhere from the cylinder to the regulator to the EPC inlet. but there is no molecular sieve that came with the unit.
4) i have cleaning wires for the inlet, and an ultrasonic bath. i clean them using a combination of those in acetone/methanol mixture, then dry the parts with nitrogen gas, and re-install using clean tweezers. i did this again this morning: same result.

Questions

How critical is it for the detector gases to be clean? Obviously, they don't have to be as clean as the helium but if we use standard copper tubing versus gc grade copper tubing is that bad?

We think it may be possible that, as someone mentioned, when the unit is not being used we still have helium flowing through it, so when we are out for the evening that oil from the lines is concentrating on the inlet, and pushing its way through the inlet when we make an injection. Would we see the concentration of hydrocarbons/oil varying or concentrating on the inlet? We see the bump as soon as it hits the high temperature (250-350 deg. C). Although, from what i see the bump doesn't seem to go up or down at all?

[Bruce Hamilton]

I still think the problem is in your injector/syringe/sample solvent - boring I know, but your blank run was flat.

You haven't commented about changes in the size of the hump if you repeat several injections quickly.

As I'm not familiar with the COC you have, I can't comment on how to clean it, but my experience is that acetone/methanol, even with ultrasonication, may not comprehensively clean all lipid components from metal surfaces. I'd definitely want any solvent to be warm to dissolve lipids.

It's critical for the detector gases to be free of organics, otherwise the detector baseline will shift up at all times, perhaps elevating even higher on higher oven ramp temperatures. Sensitivity will also decrease.

I had moisture and activated carbon traps on the detector gases. I used 1/8-1/2" Swagelok Fittings to make my own activated carbon traps from 2' of tubing for the detector hydrogen and makeup, with an Alltech ( now Grace ) large carbon trap for the detector air.

I'm still doubtful the hump is from your carrier gas lines, but you could always clean them to eliminate them. Note that you have to use a solvent that will quickly dissolve oils. I prefer multiple rinses using warm chloroform/methanol ormethylene chloride, but waste disposal can be an issue in some locations.

Bruce Hamilton

[Don_Hilton]

Oil in the detector gas lines will increase the background level - as they will be carried at a fairly contant rate into the GC and the flame of the detector. Note that I qualified the constant with fairly. As temperatures change, so does the partitioning of the oil. Also, if you turn off detector gasses over night, the oil will tend toward thermodynamic equilibrium while there may be some kenetic effect (rate of evaporation) with the gas flowing. And, if memory serves me correctly, you get increased noise in the baseline.

Reading your last notes, I am not clear on the run that you did with no injection and saw no bump. And, the conentration should vary depending on the time between runs. Let me suggest the following - and you may have it already in your data. 1) A run with no injection, just push the start button first thing in the morning. This will elute anything that has accumulated on the head of the column over night. If there is a bump, then 2) as soon as the instrument returns to the ready state, push the start button again. This will give you and idea of the difference in accumulation between runs. This may even be a good first thing Monday morning experiment.


Alternatively - if there is no bump after the overnight run, 2) take your syringe and be sure is thoroughly clean. Leave teh plunger pushed down and mane an injection just by pushing the syringe into the inlet and then removing it. This will look for material disloged from the inlet system.

Let me know if any of these ideas are of any help.

[Peter Apps]

In addition to the charcoal traps on the detector gasss I would add one to the carrier gas. Do not expect an immediate resolution to the hump since if it is coming from the gasses it will have contaminated the GC plumbing and will take some time to bleed out.

The hump's being very smooth argues against it being contaminants that are coming in immediatley prior to the current run, or with the current injection; in those cases I would expect something that looked more like GC peaks. If it is contamination it is moving very slowly down the column over multiple runs, and getting spread out into a smooth hump in theprocess. Did you see a hump with the very first run on the new column ?

How are you controlling carrier gas flow - constant pressure or constant (volume or linear) flow ? If constant pressure there will be changes in carrier gas flow during the run, which might change the detector response. I would expect it to give drift in one direction only, not a hump, and I would have expected the same problem with the programme blank, but it might be worth looking at.

Peter

[Peter Apps]

In addition to the charcoal traps on the detector gasss I would add one to the carrier gas. Do not expect an immediate resolution to the hump since if it is coming from the gasses it will have contaminated the GC plumbing and will take some time to bleed out.

The hump's being very smooth argues against it being contaminants that are coming in immediatley prior to the current run, or with the current injection; in those cases I would expect something that looked more like GC peaks. If it is contamination it is moving very slowly down the column over multiple runs, and getting spread out into a smooth hump in theprocess. Did you see a hump with the very first run on the new column ?

How are you controlling carrier gas flow - constant pressure or constant (volume or linear) flow ? If constant pressure there will be changes in carrier gas flow during the run, which might change the detector response. I would expect it to give drift in one direction only, not a hump, and I would have expected the same problem with the programme blank, but it might be worth looking at.

Peter

[Bruce Hamilton]

Just a note that it's not that rare to see "humps"in lipid GC. They are often underivatised polar lipids or triglycerides that elute with the injection, or the following run - depending on temperature profile, column film thickness, length etc. .

Another common source is any high end solvent-refined and/or vacuum-distilled hydrocarbons ( eg light lubricant base grades - such as in copper pipe drawing oils ).

I'm not saying the peaks here are similar, but both certainly appeared when I was doing lipid analysis for biodiesel in the 1980s, 1990s.

Please keep having fun,

Bruce Hamilton