LVI glass bead-liner Siltek deactivated

[BMU_VMW]

When I replace a liner by a brand new one (LVI siltek deactivated with glass beads); I seem to lose some of my compounds such as atraton, parathion-Ethyl+methyl, pendimethalin, trifluralin. This will stay for up to 30 injections and than it improves. So I guess ther must be some active places on the liner. Problem is: it takes a lot of time to 'condition' a liner this way + as it starts to get dirty again, other components will start to get tailing.
Some liners seem to be worse than others, but there is no direct link to a batch.

Has anyone seen this before? Has anyone got a solution for me?
I can't seem to find other liners that are suitable to inject 50µl CH2Cl2 (at once).

note: We are using a Thermo Trace GC Ultra with PTV and a combipal.


kind regards
BMU

[cleh]

Inject some high concentration standards each time you install a new liner and that will usually bind up any active sites. I usually run three higher than my highest standard in the curve and then run blanks after so I know there's no carryover. Good luck.

[BMU_VMW]

Cleh, thanks for your fast response.
Problem is that I've already tried that, and it didn't seem to work (but I can try again).
Only after a good number of injections I managed to get a good result.

[BMU_VMW]

I've done a new test by doing 7 injections with a 1000ng/l std (highest normal standard is 500ng/l) befor starting my GC-run.
First 'real' samples are air, solvent blank and a regular blank, all in FS so the instrument should be ok after this (i.e. no cary over from the 1000ng/l std).

Than I run a QC50ng/l ; a drift-control 100ng/l ; 10 unkowns ; a drift-control 100ng/l ; 10 unknowns ; ...

these are the results:



QC 1 and 2 are about 25 injections apart, drift1 and 2 are 10 injections apart.

clearly it takes a lot of injections to get everything right. In this example all is back to normal after +/- 15 injections but sometimes it takes up to 30 and that is a lot of wasted time.

does anyone have another suggestion to solve my problem please ?

[Peter Apps]

What are the temperature and flow programmes for your inlet ?

Why do you think that it is the inlet liner, as opposed to the column for example ?

Why do you need to inject 50 ul quickly ? - the combipal has adjustable injection speeds.

Peter

[BMU_VMW]

I'm sure it's the liner that is causing the problem since it only occurs when the liner has been replaced.

the 50µl injection (25µl/sec), is the way my collegue has set up the methode, incl solvent vent times, etc, ...
It has worked fine this way, there are no other changes to this system exept for a new liner (same type and brand) so I didn't feel the need to put in a lot of effort to change the methode of injection.

Do you expect less problems when injecting slower, or is it just a way to be able to use e.g. a baffled liner?

methode:
PTV in large volume mode
start temp 50°C
splitflow 20ml/min splitless-time 1min
cte septum purge
injection-fase: 0.3min, flow 100ml/min
evaporation-fase: not used since there is a injection fase off 0.3min.
transfer-fase: 2.5°C/sec -> 240°C hold for 1 min
cleaning-fase.

Combipal is set to inject 50µl at 25µl/sec. So this is not realy at once, but I'd still call it rather fast.
GC-oven is set at 40°C , flow 1.2ml/min

BMU

[Peter Apps]

Two things that might need attention; your inlet temperature is 20C above the boiling point of the solvent (at atmospheric) so with 50 ul injected you may still be getting flashback of sample vapour into areas that it should not go. The rapid gas flow will cause localised evaporative cooling, which will help, but the amount of cooling will depend on the fine details of heat conduction through the inlet walls and so could well change if you change a liner. Do you have cryogenics in the inlet - if so try a lower temperature. Injecting slowly will reduce flashback but to get the full benefit from a LVI and a PTV inlet you need to inject below the boiling point of the solvent.

If I understand your inlet programme properly you have a splitless time of 1 min, but your temperature programme runs for about 100 s, so by the time the inlet reaches its maximum temperature - which is when the analytes should be transferring to the column-, the split has already opened and you will lose analytes.

Peter

[BMU_VMW]

We do not have a cooled PTV (nor AC in the lab) so it is really hard for us to get down to 40°C which indeed would be better. 30°C is impossible to reach (even in winter).

regarding to the splitless time: it is a bit confusing but this time only starts when the ptv has reached his max temp. So after reaching 240°C it will stay in splitless-mode for another minute. After that the split-valve opens again and the cleaning-fase starts.

cooling will depend on the fine details of heat conduction through the inlet walls and so could well change if you change a liner

I'm sorry I don't fully understand this. If a new liner changes the heat conduction, than shouldn't this effect remain as long as this liner is being used?
What I'm seeing is some sort of activity that has an effect on some components and last for about 25-30 injections.

ps: please forgive me for spelling mistakes in english. I try my best, but these messages are made hastily in between my regular work

[Peter Apps]

The evaporative cooling will not have a detectable effect on active sites, but it will affect the volume of vapour that is generated as the solvent boils during evaporation, and the volume of vapour might influence flashback into parts of the inlet that are catalytically active. Try doing a slower injection.

Can you clarify for me - the loss of analytes happens every time you put in a new liner, and the losses decrease after extensive conditioning ?. How many times has this happened ?

Are you doing any other maintenance when you change liners - trimming the column for example ?

What detector are you using ?

Peter

[BMU_VMW]

Yes indeed: The loss of analytes happens every time I put in a new liner. It happens every time, no exception (but it took us a while to notice the connection between the liner and the problems with these analytes).

In the methode we are using we try to add new compounds on a regular basis. The analytes that are affected had been added and first we thought there was something wrong with sample prep, SRM, ... it was only later that we noticed the connection with the change of liners. Our 'older' compounds are not affected. In the worst case we had; about 10 compounds were seriously effected. There are 112 compounds in the methode at this time.

It doesn't always take the same amount of injections to clear the problem. (at least 20 at most 35).

We are a drinking water company and use this methode on DM-extracts of raw waters (surface and ground) and on drinking water. matrix is relatively clean.

• We do one run a week (60-80 injections which take +/-40 minutes/sample)
• We replace the liner every 150-200 injections (i.e. after 2-3 runs)
• septum is changed befor every run
• we change our pre-column every month and than the analitical column is trimmed (+/-30cm)
• 2-3 times a year the linerspacer is changed.

detector is a Thermo TSQ quantum triple quad ms.

since we had problems with these liners loosing some off their glass-beads and jamming the splitt-valve we have a small filter installed just before the split-valve. This filter is replaced and the housing is cleaned about 4 times a year or when a liner accidently breaks in the PTV. but this has no effect on the loss off analytes.

Hope this info can help.
anyway thanks a lot for your effort!

[Peter Apps]

I think that you have to do some systematic troubleshooting to isolate the cause of the problem. You might have already done some of this.

First - try 1 ul injections of a 50 times concentrated solution - this can be a standard not a sample (in fact - do you see the same loss of compounds with standards as with samples ?) onto the LVI liner - if the peaks are OK the problem is with the large injection volume, not the liner.

Swap the liner for an ordinary split/splitless liner with a small wisp of glass wool. Inject the 1 ul 50x solution. Try it also without the glass wool.

Try an ordinary hot injection of the 1 ul 50x solution.

Can you post a link to the type of liner that you have.

How often do you tune the MS ?

I'm not a pesticide expert, are the "new" compounds that give problems known for being sensitive to inlet problems ?, and are any of your "old" compounds that give good detection also known to be sensitive to inlet problems ?

Peter

[BMU_VMW]

First - try 1 ul injections of a 50 times concentrated solution - this can be a standard not a sample (in fact - do you see the same loss of compounds with standards as with samples ?yes we do, standards, as well as spiked samples seem to have a loss in sensitivity) onto the LVI liner - if the peaks are OK the problem is with the large injection volume, not the liner.than why would it only occur each time we change the liner?

Swap the liner for an ordinary split/splitless liner with a small wisp of glass wool. Inject the 1 ul 50x solution. Try it also without the glass wool.
Try an ordinary hot injection of the 1 ul 50x solution. => I will as soon as I can find the time for it. the system has to be used now to do routine analysis.

Can you post a link to the type of liner that you have.
sorry I can only post an old picture. They are 12cm long and have a 2mm internal diameter. As you see we use Kalrez rings to seal.




How often do you tune the MS ? We tune when we change an ion volume. We check sensitivity befor every run and if needed we do a tune. The instrument is tuned +/- every second week.

I'm not a pesticide expert, are the "new" compounds that give problems known for being sensitive to inlet problems ?, and are any of your "old" compounds that give good detection also known to be sensitive to inlet problems ?
the old ones contain e.g. mevinphos, endrin, DDT which are all known to breakdown. These are fine. I'm not sure about the new ones, but they are normal ones the liner has been used for a while.

[Peter Apps]

First - try 1 ul injections of a 50 times concentrated solution - this can be a standard not a sample (in fact - do you see the same loss of compounds with standards as with samples ?yes we do, standards, as well as spiked samples seem to have a loss in sensitivity) onto the LVI liner - if the peaks are OK the problem is with the large injection volume, not the liner.than why would it only occur each time we change the liner?Agreed, the correlation between liner changes and the problem is compelling evidence for a link, but correlation does not imply cause and effect, and in troubleshooting taking a step back from what you are sure is the cause of a problem can often reveal new possibilities

Swap the liner for an ordinary split/splitless liner with a small wisp of glass wool. Inject the 1 ul 50x solution. Try it also without the glass wool.
Try an ordinary hot injection of the 1 ul 50x solution. => I will as soon as I can find the time for it. the system has to be used now to do routine analysis.Your strategic choice is between fixing the system or continuing to use the system in its present unsatisfactory condition.

Can you post a link to the type of liner that you have.
sorry I can only post an old picture. They are 12cm long and have a 2mm internal diameter. As you see we use Kalrez rings to seal.Which company makes these ?, do they have a wesite ? Are the glass beads loose, or sintered (their falling into the plumbing suggests that they are loose) If they are loose it is very easy to get uneven packing that leads to gas channels and strange behaviour of the liquid sample film. This is the kind of problem that might resolve with repeated injections. Where is the tip of the syringe needle in relation the the beads when the injection is made ? [/color]




How often do you tune the MS ? We tune when we change an ion volume. We check sensitivity befor every run and if needed we do a tune. The instrument is tuned +/- every second week. If +/- every second week for an MS tune coincides with 2-3 weeks between liner changes than maybe you are looking at the wrong end of the column. What happens if you retune without changing anything else ?

I'm not a pesticide expert, are the "new" compounds that give problems known for being sensitive to inlet problems ?, and are any of your "old" compounds that give good detection also known to be sensitive to inlet problems ?
the old ones contain e.g. mevinphos, endrin, DDT which are all known to breakdown. These are fine. I'm not sure about the new ones, but they are normal ones the liner has been used for a while.If the problem was due to liner activity then all the sensitive compounds would be affected

Peter