Frustrating Reproducibility Issues - Any help?

[mwbaele]

Hello all,

I recently started using an old Varian GC 3380 system with the aim of studying migration from recycled paperboard for my PhD research.
I am in the process of developing a detection method for a solution of surrogate compounds. However, i keep struggling with severe reproducibility issues.
Separation and peak shape all seem OK, but the relative standard deviation of repeated injections of the same sample goes up to alarming numbers.
It seems that the deviation of polar compounds is a bit higher than that of apolar compounds.
I have tried almost everything but can't seem to get it right.

Apparatus: Varian GC 3380 with a 1079 injector (single-tapered liner).
Temperature program: hold 0.5min at 100°C, 25°C/min, hold 6min at 280°C
Injection of 1 µl at 10µl/s
Injector at 340°C
Solvent: tert-butyl methyl ether
Split ratio 1/40

Below are some captures of chromatograms.
Full chromatograms:

http://picpaste.com/ales-wNNMvV8K.png

Solvent peaks:

http://picpaste.com/solvent1-4YjNpuiN.png
Crop of solvent peak:

http://picpaste.com/solvent2-VGh86fRI.png

Internal standard peak:

http://picpaste.com/piek1-TdcZo5Wi.png

Analyte peaks:

http://picpaste.com/pieken-vbxufUl0.png

I am rather new to chromatography and completely new to this forum, so feel free to ask for further info. Thank you so much for any suggestions.

[Bigbear]

Welcome.
Are you making multiple injections from the same vial? If yes, try 1 injection/vial.
First off I would lower your oven starting temp to 35 or 40C.
Make sure your syringe is drawing correctly.

[Peter Apps]

Welcome to the forum.

Although the peak areas are quite variable the ratios of areas on each chromatogram are probably not too bad - all the peaks get bigger or smaller to about the same extent. So there is no serious inlet discrimination. How repeatable are the ratios of analytes to internal standard ?

The two most likely causes of variable areas are variable injection volume and inconsistent splitting. Watch the autosampler in operation; are there bubbles in the syringe, inconsistent plunger positioning ?, is there any slack in the attachment of the syringe or the plunger driver ?

Try adding a little bit of loosely packed deactivated glass wool to the inlet liner to help evaporation and gas mixing. Check that the column is inserted the correct distance into the inlet.

Peter

[mwbaele]

Welcome.
Are you making multiple injections from the same vial? If yes, try 1 injection/vial.
First off I would lower your oven starting temp to 35 or 40C.
Make sure your syringe is drawing correctly.

The above results show 10 analyses, which come from two runs of 5 vials containing the exact same sample.

[mwbaele]

Welcome to the forum.

Although the peak areas are quite variable the ratios of areas on each chromatogram are probably not too bad - all the peaks get bigger or smaller to about the same extent. So there is no serious inlet discrimination. How repeatable are the ratios of analytes to internal standard ?

The two most likely causes of variable areas are variable injection volume and inconsistent splitting. Watch the autosampler in operation; are there bubbles in the syringe, inconsistent plunger positioning ?, is there any slack in the attachment of the syringe or the plunger driver ?

Try adding a little bit of loosely packed deactivated glass wool to the inlet liner to help evaporation and gas mixing. Check that the column is inserted the correct distance into the inlet.

Peter

In this analysis there is indeed no large difference between peak ratios, however I have noticed on other runs that peaks of polar compounds often show more variation (about 5% more relative standard deviation). Below is an overview of analyte to internal standard ratios. Peak 2 and the internal standard are apolar compounds. The others are apolar and seem to show a larger variation in their ratios.

EDIT: i forgot to mention that in the table below, the last value is the relative standard deviation and the one but last value is the average.

peak 1/IS 1,82 2,07 2,15 1,97 2,04 2,04 2,12 2,10 2,02 1,97 2,03 4%
peak 2/IS 0,97 0,97 0,97 0,94 0,96 0,94 0,97 0,98 0,95 0,95 0,96 1%
peak 3/IS 1,32 1,52 1,59 1,44 1,49 1,50 1,55 1,54 1,48 1,44 1,49 5%
peak 4/IS 3,27 3,65 3,76 3,50 3,59 3,61 3,71 3,67 3,58 3,50 3,58 4%

The autosampler has indeed shown some issues (air bubbles), but the syringe has been replaced by a fully functional one. Before injection, the syringe is rinsed twice with sample (plunger goes up and discards to waste), followed by ten flushes inside the sample vial (plunger goes up and down while needle is submerged in sample). During the first few flushes usually there is an air bubble above the sample, but this is usually gone by the last flush. However, I do not think that this air bubble, if present, poses any problem as there is never any air in the 1µl to be injected.

[tkubowicz]

Hello

It looks like you have problem either with sampler or inlet.
I'd recommend:

1.Check inlet for leaks - septum, split line, o-rings etc
2.Check syringe in sampler - is it tight? Fill it manually with solvent and pick it to septa. Then try to slowly move plunger to see if syringe is not leaking.

Also: Methyl tert-butyl ether boiling point is about 55°C so why your first ovent temp is 100°C?



Regards

Tomasz Kubowicz

[Peter Apps]

An air bubble in the syringe will spoil repeatability no matter where it is. Even if it is above the 1ul that gets injected it will be compressed as the needle pierces the inlet septum so that less than 1ul is injected.

The polar/non-polar difference might be a red herring if the polarity coincides with boiling point - what are the boiling points of each compound ? If the split is inconsistent it will affect different BPs differently. Adsorptive activity has more impact with smaller quantities - what are the concentrations of the components ?

Inconsistent splitting can be caused by inconsistent injection volumes - so you need to be sure that the air bubbles are eliminated.

Peter

[mwbaele]

An air bubble in the syringe will spoil repeatability no matter where it is. Even if it is above the 1ul that gets injected it will be compressed as the needle pierces the inlet septum so that less than 1ul is injected.

The polar/non-polar difference might be a red herring if the polarity coincides with boiling point - what are the boiling points of each compound ? If the split is inconsistent it will affect different BPs differently. Adsorptive activity has more impact with smaller quantities - what are the concentrations of the components ?

Inconsistent splitting can be caused by inconsistent injection volumes - so you need to be sure that the air bubbles are eliminated.

Peter

I will try installing a completely new syringe and applying some grease to the autosampler for smooth operation. I have noticed some rattling sound during operation, maybe these vibration cause some issues.

As someone suggested above, I have started some runs below the solvent boiling point, but this only seemed to make things worse. The peak area variation was now huge.

The compounds I use are listed below:
internal standard - bicyclohexyl - 0.04 mg/ml - BP 227°C
Peak 1 - triethyl citrate - 0.72 mg/ml - BP 294°C
Peak 2 - heptadecane - 0.11 mg/ml - BP 302°C
Peak 3 - dipropyl phtalate - 0.42 mg/ml - BP 317°C
Peak 4 - 4-methyl benzophenon - 0.28 mg/ml - BP 326°C

If low concentrations appear to pose a problem, I might be in trouble as the ultimate goal of this method will be to detect low quantities of these compounds...

By the way - thank you for your replies. Greatly appreciated.

[mwbaele]

Hello

It looks like you have problem either with sampler or inlet.
I'd recommend:

1.Check inlet for leaks - septum, split line, o-rings etc
2.Check syringe in sampler - is it tight? Fill it manually with solvent and pick it to septa. Then try to slowly move plunger to see if syringe is not leaking.

Also: Methyl tert-butyl ether boiling point is about 55°C so why your first ovent temp is 100°C?



Regards

Tomasz Kubowicz

Thank you for your suggestions. I will try them out.
As for the oven temperature - if i start at e.g 35 there is a large time gap between the solvent peak and the first eluting peak. This serves no real use to me other than prolonging the analysis time. Since I use split injection, raising this temperature should -I believe- pose no problem, since the high boiling compounds can still condensate upon entering the cooler column.

[Peter Apps]

Your are right that a lower oven temperature simply prolongs the analysis, with no benefit to repeatability - as expected. You have nice sharp peaks and plenty of resolution, and with a split injection the sample transfers to the column as a sharp band, so you do not need a low starting temperature to focus peaks at the head of the column.

However, a lower temperature should not make any difference to the repeatability, unless your are getting condensation of solvent on the column. I would expect that only with splitless injections - which makes me suspect that your split might be blocked so you are in effect doing splitless injections. The solvent peak is sharper than I would expect if that were the case, but as a troubleshooting exercise try halving the split ratio setting and see if the peaks double in size. If they don't your split is probably blocked, or restricted.

Peter

[mwbaele]

Your are right that a lower oven temperature simply prolongs the analysis, with no benefit to repeatability - as expected. You have nice sharp peaks and plenty of resolution, and with a split injection the sample transfers to the column as a sharp band, so you do not need a low starting temperature to focus peaks at the head of the column.

However, a lower temperature should not make any difference to the repeatability, unless your are getting condensation of solvent on the column. I would expect that only with splitless injections - which makes me suspect that your split might be blocked so you are in effect doing splitless injections. The solvent peak is sharper than I would expect if that were the case, but as a troubleshooting exercise try halving the split ratio setting and see if the peaks double in size. If they don't your split is probably blocked, or restricted.

Peter

Dear Peter,

The apparatus enables flow control through a split vent on the side of the GC. Using a bubble flow meter, I can measure a rather constant split gas flow, so I assume that a blocked split is not the problem.
I have ordered a new syringe and a glass wool packed injector insert. I will install these and see how they affect the results. Also I will further inspect gas flow rates, as these are not controlled electronically and might be subject to changes as they have been left untouched for quite a while.

Regards,

Maarten

[Peter Apps]

Hi Maarten

With back-pressure regulated inlets such as are found on all modern (i.e. past 25 years) GCs the flow out of the split vent stays on all the time during both split and splitless injections - all that differs is the route taken by the gas inside the inlet. So you cannot always diagnose a blocked split line by measuring split flow - that is why I recommended halving the split setting and checking the effect on peak area.

Peter

[mwbaele]

Hi Maarten

With back-pressure regulated inlets such as are found on all modern (i.e. past 25 years) GCs the flow out of the split vent stays on all the time during both split and splitless injections - all that differs is the route taken by the gas inside the inlet. So you cannot always diagnose a blocked split line by measuring split flow - that is why I recommended halving the split setting and checking the effect on peak area.

Peter

I just completed a run where I have halved the flow through the split vent -from a measured 10ml/min to 5ml/min- while keeping the carrier gas flow constant. This resulted in a slight tailing of the solvent peak overlapping the analyte elution window, but the peaks have indeed about doubled in size, as they should.
So I think the problem is either due to a leak in the injector or more likely a defective syringe/autosampler system.

Maarten

[Peter Apps]

Hi Maarten

The peaks size behaving as it should with a change in split ratio eliminates both a blocked split and a significant leak.

Inconsistent injection volumes and erratic vaporization are still on the table. The glass wool with help with vaporization, and might reduce the impact of inconsistent injections.

Peter

[mwbaele]

Hi Maarten

The peaks size behaving as it should with a change in split ratio eliminates both a blocked split and a significant leak.

Inconsistent injection volumes and erratic vaporization are still on the table. The glass wool with help with vaporization, and might reduce the impact of inconsistent injections.

Peter

Great news!

After replacing the syringe and changing the injection method, the relative standard deviation has finally fallen below 1% and it seems to stay that way!
The problem was partially due to a worn plunger leaking solvent and an erratic injection method.
The glass wool liner seemed to decrease peak areas though, I will not use it anymore.

Thank you for your help and suggestions!