EPA 525.2 Calibration Issues

[Ostravaczech]

I am experiencing linearity issues with some EPA 525.2 compounds. Data from July, August and Sept is good. Then, in October, PM was done (oil changed, source cleaned by Agilent engineer) and since then I cannot get the curves to pass, specifically adipate, methoxychlor, butachlor, phthalate. I am using the same method to acquire the data that I used in July, Restek drilled cyclounilner hole on top, MMI programmed from 65 to 275, column program 50 to 320; the only difference that I can think of is that instrument was autotuned after the PM. Agilent 7890/7000C combo, full scan, cal range 0.02 - 10 ug/mL, same standards used, made the curve many MANY times over. Any suggestions will be welcome.
Thank you!
Ivo H

[Steve Reimer]

In what way do they fail?

[James_Ball]

If the cleaning or retuning caused an increase in sensitivity, it could be a problem with exceeding the upper end of the instrument linearity. Compare the area counts of the Terphenyl-d14 monitoring compound from before and after the cleaning. If there is a large increase you may need to lower the gain setting in your method to reduce sensitivity. If it is lower in area counts then try to increase gain until you achieve similar sensitivity to the older data.

[adoghaim]

From my experience, when the agilent engineer finishes the PM, they tune the instrument too quickly, like a soon as it says ready. Normally you should allow the instrument to pump down over night. Now if that's what they done, try doing another autotune, or at least a checktune. I am thinking your instrument is just out of tune.

[Ostravaczech]

In what way do they fail?

October adipate:
https://lh3.googleusercontent.com/AdssD ... 96-h460-no

July adipate:
https://lh3.googleusercontent.com/7teAN ... 20-h466-no

As you can see, it’s only few cal points, and few compounds (adipate, metribuzin, methoxychlor and phthalate), even quad fit will not pass r^2. I made the curve many many times so I doubt it’s the prep. Today, since I was running CI, I took the EIEX apart and it was DIRTY. Of course, the Agilent engineer says I’ve been running dirty extract. Not true - I have not been able to run any samples at all, just pure standards in clean solvents. Any ideas?

[Ostravaczech]

From my experience, when the agilent engineer finishes the PM, they tune the instrument too quickly, like a soon as it says ready. Normally you should allow the instrument to pump down over night. Now if that's what they done, try doing another autotune, or at least a checktune. I am thinking your instrument is just out of tune.

Did that. It did not help.

[Ostravaczech]

As I mentioned previously, I took the EIEX apart today and this is what I found:
https://lh3.googleusercontent.com/tyK3b ... 0-h1334-no
https://lh3.googleusercontent.com/a6PXy ... 8-h1222-no
I must point out that the PM was done by Agilent engineers twice - October 13, and October 25 when they cleaned the source. Since then, all I have been shooting is standards. No samples extracts have been put through the instrument since mid October. I was bit puzzled when I discovered the discoloration on the lenses. Normally, this should not be occurring, am I right? Repeller, btw, was clean.

[James_Ball]

I can't get the images to open on my computer, but if the discoloration is blue to gold in color then it could be from an air leak. Even shooting clean extracts will give a bluish or gold color to the lenses if air is present. Normally you will see it on half of the repeller, but often it will appear worse on the drawout plate just to the side of the hole and on the ion focus lens inside the cup beside the hole.

[Ostravaczech]

I can't get the images to open on my computer, but if the discoloration is blue to gold in color then it could be from an air leak. Even shooting clean extracts will give a bluish or gold color to the lenses if air is present. Normally you will see it on half of the repeller, but often it will appear worse on the drawout plate just to the side of the hole and on the ion focus lens inside the cup beside the hole.

Thanks! They had blueish / gold discoloration, as you described but there is no air / water leak (around 2% for water, 1% nitrogen).

[James_Ball]

Check in the settings for MS2, there is a mass cutoff for TTL mode which allows you to reject ions below a certain mass from reaching the detector when running in single quad mode using MS1. If that is set to something above 18, then you may have more air and water than it appears, but the MS2 is filtering it out.

[Ostravaczech]

Check in the settings for MS2, there is a mass cutoff for TTL mode which allows you to reject ions below a certain mass from reaching the detector when running in single quad mode using MS1. If that is set to something above 18, then you may have more air and water than it appears, but the MS2 is filtering it out.

Will do, thanks! I never touched anything in the MS2 for TTL filtering. Would that be set up in the factory? I’ve had this instrument for over 2 years.

[James_Ball]

It comes with a default setting but I don't remember exactly what it is. MS2 tab should show a check in the TTL button and the Cutoff will be the lowest mass that will pass through the quad, it rejects everything below that. It isn't exact as you will get a small response below what you set the cutoff to, but it does a good job of cleaning up that low portion of the mass range. When doing volatiles analysis I set it to 35, which is the lowest mass of the scan range and it helps with filtering out the methanol peak that comes through from the desorb of the trap. The down side is it will mask any air leak you might have. When working on tunes and looking for air leaks I set it to 18, so the air and water will pass through better.

[Ostravaczech]

It comes with a default setting but I don't remember exactly what it is. MS2 tab should show a check in the TTL button and the Cutoff will be the lowest mass that will pass through the quad, it rejects everything below that. It isn't exact as you will get a small response below what you set the cutoff to, but it does a good job of cleaning up that low portion of the mass range. When doing volatiles analysis I set it to 35, which is the lowest mass of the scan range and it helps with filtering out the methanol peak that comes through from the desorb of the trap. The down side is it will mask any air leak you might have. When working on tunes and looking for air leaks I set it to 18, so the air and water will pass through better.

Hi James - played with the TTI today. Default value was set at 50, ran H2O check. Then changed it to 10 and ran the H2O again. Surprisingly, there was no difference in water/air leak values. Remember - I am in AZ where humidity is very low especially this time of the year so I generally get H2O of less than 2%.

[James_Ball]

That at least rules out an air leak causing the problem.

[Ostravaczech]

That at least rules out an air leak causing the problem.

Yes, we got that out of the way. As for the curves, for some compounds (adipate, phthalate, butachlor, methoxychlor), they are convex (concave upward) so the detector is not getting saturated. For others (propachlor, HCB, BaP), the curves are perfectly linear. Changing the linear (4mm with wool, without wool, drilled uniliner, 2mm ultra inert dimpled), changing the injection technique (hot SL, cold SL, pressure surge SL), using different column, etc., makes no difference - the compounds behave the same. The DFTPP tune is passing every single time, breakdown well below 20%, depending on the liner, it can be less than 1%, most of the time it is around 5%. So basically, today, I gave up because I do not know what else I could change…. As I mentioned before, I had successful runs July through end of September (all compounds linear, with RSD<10% for many), then in mid October, annual PM was done by Agilent. Any ideas / suggestion?