Internal standard ratios fluctuating - 8270, Agilent GC/MS

[625]

We are having the same problem the OP in the following thread had:

http://www.sepsci.com/chromforum/viewto ... t=perylene

Running 8270 on ocean sediment samples, on two different Agilent GC/MS systems, the recovery of the 6th and final internal standard compound, perylene-d12, will fluctuate significantly with respect to the response of chrysene-d12, the 5th internal standard.

The ratio of chrysene-d12 to perylene-d12 will vary from 1.2 to over 6. This diminishing recovery of perylene-d12 results in very high calculated values for compounds associated with this internal standard compound.

Here are some important points to consider:

These are internal standards, added after extraction and concentration, used to correct results for solvent evaporation, etc. The ratio of our target compound response to the internal standard response is compared to the same ratio in the calibration. They are not surrogate standards that go through extraction to monitor extraction efficiency.

This problem occurs both before and after cleaning source, cleaning the inlet, changing liners and liner type.

The final conclusion in the thread mentioned above is that it may have something to do with the acetone used in ASE extraction, which is what we are doing. However, the problem occurs in calibration verification standards prepared using only methylene chloride, no extraction and no acetone.

The problem seems to get worse as the run persists, suggesting that as the inlet or column gets dirty the problem gets worse. But this is not always the case.

Any ideas?

[Don_Hilton]

Assuming that the inlet is now well explored and maintained... And, the problem looks like it is chromatographic... I've run into this once before with PAH's and again with silylated sugars. (Ok it was someone on the next bench - their ulcer, but I learned...)

What is the temperature of the transfer line? If the transfer line is too cold, the high boilers will tend to be retained. And if the temperature has gone too high or the column has been damaged (like by air leaking into the system) the column will have active sites in the transfer line - and it will do some of the same ugly things as an inlet problem.

[625]

Thanks Don - I was thinking the same thing this morning - the transfer line Temp on our new 5975 is set at 250C, but our 5973 is at 300C. The same problem had been occuring on our 5972, which has no transfer line heater.

We will try this change and see if it helps. However, because the problem is intermittent, I'm having some doubts. Will post an update after this is tried.

[Ron]

I don;t think you are looking at the transfer line temperature, it sounds like you are looking at the ion source temperature. The 5972 has a transfer line heater, but no source heater, and the source was actually heated by the transfer line.

An ion source that is too cold can give some problems, but the transfer line from the GC to the MS is the first place to look for cold spots.

[625]

I don;t think you are looking at the transfer line temperature, it sounds like you are looking at the ion source temperature. The 5972 has a transfer line heater, but no source heater, and the source was actually heated by the transfer line.

An ion source that is too cold can give some problems, but the transfer line from the GC to the MS is the first place to look for cold spots.

You are right - I was thinking backwards in terms of the 5972.

As for the transfer line temperature, we use 300C on our 5973 but had 250C on the 5975; we are testing this now to see if it helps. It's the AUX Temp setting on both instruments.

[625]

Well, increasing the transfer line temperature did not help; we still see the recovery of perylene-d12 fluctuating a LOT relative to chrysene-d12, which stays relatively stable in terms of abundance.

The problem does not seem to be one that relates generally to our high-molecular weight compounds. Target analytes that have larger quantitation ions and have longer retention times do not show as large of a decrease in recovery, resulting in calculated results for these compounds that are way over 100%. As the liner / top of column get dirty, the large compounds all decrease somewhat, but the perylene-d12 shoots WAY down. The target compound perylene is fine.

While the ratio does change from sample to sample, there is also a general trend of decreasing ratio as the run goes on. We are going to try changing liners late in the run. However, this seems more like a band-aid than a fix of the problem.

Any other ideas out there?

[stewiegriffin]

what is your final temperature and how long do you hold at that temperature?
since you mentioned ocean sediments, i wonder if your extracts might contain a lot of high boiling material (biological compounds) that never make it all the way through the column.
i use a db-5 (30 x 0.25 x 0.5), and ramp up to 350C, holding for about 3 minutes. after running a number of soil extracts, i usually do see a drop in response for the last 2 ISTDs (and late-eluting PAHs).
as far as discrimination goes, my hunch has been that there is a critical run temperature at which compounds begin to be affected by high-boiling material that has accumulated on the column. i often have a problem with pyrene falling out of calibration because the critical point occurs between the r.t. of pyrene and its internal std, chrysene-d12. in other words, the pyrene result appears high because it was relatively unaffected while chrysene-d12 dropped significantly. (this doesn't really explain fluctuations from one run to the next, but ...)
usually routine injection port maintenance will correct this problem. i don't know whether you're using a guard column, but i have found this to be a big help. i use a single-gooseneck liner without glass wool, and the usual routine is to clip off a length of guard column depending on how dirty the extracts were in the previous run.
if this isn't sufficient, proceed to change the liner, gold seal, bake out the column (~ 30 min @ 350C), etc. if none of this helps, try clipping off ~ 1 m or more of the analytical column, and install a new 5 m guard column. if this doesn't work, replace the column. (i've tried rinsing columns, but it never seems to be worth the trouble)
anyhow, the upshot of all this is that i have struggled with similar problems, and it's still something of a mystery to me, but maybe this will give you some ideas.
good luck!

[AICMM]

625,

GPC cleanup. Depending on how you do it, you might increase your detection limit by two but I suspect your last IS will hold on much better.

I would be shocked if changing the liner mid-run would meet auditors approval unless you re-DFTPP and run another daily standard. If that sample were at the front of the very next daily run, would the IS pass?

Best regards.

[625]

Stewie -

We ramp up to 320C; the column literature states a max temperature of 325C. I'm guessing your DB-5 is the same; do you see that they can go up to 350C no problem? And is it a DB-5 or a DB-5MS? Not that there is really much of a difference.

We don't currently use guard columns, but do use Restek Siltek gooseneck liners with glass wool pre-inserted. They are working great for our recovery of the PITB compounds such as 2,4-dinitrophenol.

The other weird part is that for us the effect is only for perylene-d12; chrysene-d12 appears unaffected - its response looks consistent over time and consistent relative to the first four internal standards.

Any chance that something is happening to perylene-d12 as it sits in the matrix? Maybe one of the deuteriums exchanges with a hydrogen and completely changes the ionization behavior?

AICMM - I think changing liners mid-run is ok. In fact companies like Gerstel are selling automated liner changers that can change the liner at any numbered run interval you choose. It would appear to be a gray area, though, and I would prefer to find a better solution. Maybe 350C will do it.

We had an old GPC that died about ten years ago and we stopped doing it. Recently we experimented with a Varian GPC that did nothing. What manufacturer / GPC setup do you recommend for semivolatiles?

[stewiegriffin]

625,

the column is a phenomenex zb-5 (upper limit 360C). the zb-5ms is only rated up to 325C.

i guess i was thinking that perhaps due to differences in column (film thickness?) and temperature program, you were seeing a sharp drop in response somewhere between chrysene-d12 and perylene-d12, whereas i see the drop between pyrene and chrysene-d12. but it's basically the same phenomenon - some kind of interaction with high boiling material that has accumulated on the column.
it would seem that all the preventative measures (glass wool, frequent liner changes , guard column, frequent removal of short lengths of guard column, periodic bakeouts) are helpful to a degree, but nothing short of gpc will keep the crap from eventually messing up the column. (of course this is just a guess!)

also i am thinking mostly about our sim method for pahs (2 ng ISTDs on column). i'm not sure that i typically see such a pronounced effect for the full-scan 8270 method (20 ng ISTDs).

i don't think the perylene-d12 is de-deuterating (for lack of a better word), though i have wondered about this because i see very small traces of some undeuterated cousins of surrogates like phenol-d5 and nitrobenzene-d5 in method blanks.

i hadn't heard about the automated liner changers before. i could really use something like that for pesticides analysis by gc-ecd. we use restek cyclosplitters which are great for keeping stuff off the column, but they seem to be particularly bad for ddt breakdown once a bad sample is injected. (by the way i'm open to suggestions regarding the best way to deal with ddt breakdown)

[625]

Stewie - Thanks for the info. We're still working on this problem. We think there may have been some precipitation in the internal standard ampule before we cracked it open, and speculation is that the first compound to precipitate out would be the perylene-d12.

While this may have contributed to the problem, in that all would start with relatively low perylene-d12, the fact remains that perylene-d12 still fluctuates in ratio to chrysene-d12 across samples and standards all prepared with the same internal standard solution and injected in the same batch. You may be on the right track in that only GPC will fix it for good. However, we have not been using GPC for 625 (water matrix) for quite some time and the problem only recently popped up on that instrument. Blast!!!

DDT degradation for me appears to be tied to inlet cleanliness. Not too long ago I was seeing bad DDT degradation, and it persisted even though I cleaned or changed every inlet part multiple times, as well as the source and column. Finally I inspected the syringe (sometimes the obvious things escape you for a while), and it was a little dirty. I cleaned the syringe and that little plastic piece that the syringe rests on (for Agilent auto-injectors) and that did it. I think it was that plastic piece more than the syringe, but who knows. So maybe look for some obscure part that is getting gunked up by the dirty samples.

Finally, I'm still looking for GPC advice (on top of the perylene problem). Make and Model and how you like it.

[Peter Apps]

Vaporizing injectors are vulnerable to all sorts of problems, and there is an outside chance that small fluctuations in injection conditions can be causing the substantial changes in discrimination that you are seeing. You can try changing some of the autoinjector parameters such as pre and post injection dwell, but check also that you do not have a vacuum in your sample vials that is affecting how much sample gets sucked into the syringe (if there is variation it will give you changes in the areas of your well behaved peaks as well as erratic variability in ratios). Vacuum can be a result of your instrument lab being cooler than the sample prep lab.

Peter

[625]

Thanks Peter - will look into suggestions.

[AICMM]

Regarding GPC, OI makes a commercial one that used to be known as ABC. I believe Fluid Metering also makes (made?) one but I am not positive about that. You can always make one yourself as well with a pump, a Rheodyne valve and the appropriate column. Buying a used UV detector makes the system even more robust. If you would like more information contact me through my web site and I will try to find the folder I have on the subject.

Regarding the perylene dropping out of solution, I always used to warm the vial up to room temperature and shake thoroughly before spiking samples with IS. Helped minimize that problem but even if that was the problem you would see it across the board not just in a few samples.

Regarding breakdown, my suggestion is a single gooseneck with just a breath of glass wool. (Sometimes science is an art....) Just enough wool to hold up septa, not so much that you cannot easily see through it.

[625]

Thanks for the GPC tip, AICMM. Planning to look into GPC soon.

Regarding breakdown, my suggestion is a single gooseneck with just a breath of glass wool. (Sometimes science is an art....) Just enough wool to hold up septa, not so much that you cannot easily see through it.

Well said. I agree with the liner suggestion. One that we recently started using that is working GREAT for sensitive compounds such as 2,4-dinitrophenol is the Restek Siltek single gooseneck liner with glass wool pre-inserted; the one intended for injections of <2uL.

Also I have seen that deactivated / inert glass wool (if you insert it yourself) is only as advertised for so long; usually there is NO WAY you are going to use up all 10g (the smallest I have found to purchase) before it startes to turn, so you need to replace it every couple years.