5973 GC Problem with relative Abundances

[jeffsmitheverybody]

I was hoping that I might be able to get your input on an issue that we are having here at the lab. We have a 5973 GCMS that we are running VOC’s on. The problem we are seeing is a rise in the relative abundances of the ISTD’s throughout the run of the curve.

At first I thought that this issue might be due to varying amounts of MeOH in the vials, but I made standards with the same amount of MeOH and still saw a rise in abundance as the higher concentration ran. For example, our 5ppb standard will run with area counts around 500K, 1Million, 1.1Million, 7200K. By the 200ppb standard exhibited 820K, 1.4Million, 1.6 Million, 1.1Million

We run with the following configuration:
4552 OI (Archon) Autosample
OI Eclipse concentrator (also tried with a 4560 Concentrator)
6890 GC Oven
5973 Network Mass Spec

I have tried the following in an attempt to distinguish where the problem might be coming from.

1) I replaced the concentrator and the XFER line to the GC with no change, under the assumption that there could have been active sites. Still saw the ISTD rise over the course of the curve.
2) Ran water samples instead of Soil samples (both directly through the concentrator and using the autosampler) and had the same results of an increase in the ISTD throughout the run.
3) I have replaced the XFER line from the Autosampler to the Concentrator. No change
4) I have replaced the Concentrator and still no change.
5) Replaced the 1uL ISTD loop
6) We have tried multiple traps
7) Direct injection: I did a direct injection of standards to represent a low point (5ppb) midpoint (25ppb) and high point (200ppb) and so no change in the ISTD area counts. They all remained constant 500K, 1M, 1M, 500K.
Spiked directly three points directly onto the trap (5ppb, 50ppb, 200ppb)

These problems have all arisen since we had maintenance on the instrument in which the following were replaced. The parts were:
Clean the source, replaced Electron Multiplier, Replaced Repeller, replaced DO, and replaced Repeller Insulators.

Any help would be much appreciated

[Yama001]

It sure looks like you have done all the right things. The behavior looks like something is active, disproportionally attacking lower concentrations. Does this behavior follow time of injection? Or does it follow the concentration of the standards (not sure if you have tried, for example, running from high conc to low)?

It looks like you have swapped out everything but the GC - perhaps it is activity in the inlet or the weldment? I had a case like this years ago where that was the problem. (though the direct injections seem to rule this out).

You replaced the drawout plate - is it the one with the 6 mm hole (the conventional wisdom in VOA at this point seems to like the 6 mm hole - no idea if this will bear on your problem though).

[jeffsmitheverybody]

It sure looks like you have done all the right things. The behavior looks like something is active, disproportionally attacking lower concentrations. Does this behavior follow time of injection? Or does it follow the concentration of the standards (not sure if you have tried, for example, running from high conc to low)?

It looks like you have swapped out everything but the GC - perhaps it is activity in the inlet or the weldment? I had a case like this years ago where that was the problem. (though the direct injections seem to rule this out).

You replaced the drawout plate - is it the one with the 6 mm hole (the conventional wisdom in VOA at this point seems to like the 6 mm hole - no idea if this will bear on your problem though).

We are using the 6mm Hole. I will replace the inlet and see if maybe that is it. I also noticed that the analyst was using a pretty wacky (sorry for the adjective) tune file. It seemed like the tune was jumping all over the place. I have since changed it to a more conventional VOA tune file, which is far more stable and am running a calibration curve today to see how it fares. Will keep this thread updated.

[James_Ball]

Rising IS with increasing calibration concentration is a problem we dealt with for years after we bought our first 5973 and on through the next two and a 5975. We tried everything. On the first unit it seemed that fresh diffusion pump oil fixed it for a few days, but we had the same problem when we got our first turbo pump model too. We tried the larger draw out plate, different filaments, traps, temps, and a lot of other things including methanol volumes. Nothing worked.

After running our ICPMS for about a year I returned to running Volatiles and just notices that one of the instruments no longer had this problem. After looking into it, the previous analyst had replaced the K&M electron multiplier with one if the ETP ribbon type multipliers. Not sure what the difference is, but we changed all the volatiles instrument to those and the problem went away. We then changed the semi-volatiles instruments over and a lot of the problems with the heavier PAH not giving linear calibrations went away also. It may or may not work for you, but we have not had that problem since switching multipliers.

[01310040231]

I have used ETP multipliers and problem was not resolved. I have started cleaning the water management assembly once a month and it has been working for us.

[James_Ball]

Another thing I have had to do recently on one of our drinking water instruments is to use the Gas Saver setting to boost the total flow up to about 40-45ml/minute after the desorb where is it in the low 30s, and hold it at the high setting throughout the rest of the run. It seems to help with the sensitivity drop on long runs. Sometimes we run 48 hours non stop and would have low IS recoveries by the end of the run, letting it sit 24 hours brought them back. Also running a larger vacuum pump and the inline moisture traps helps too.