Tekmar ATOMX internal response problem

[lcgcscientific]

Hi,
We've been trying to determine an issue that is apparent in both ATOMX systems for the last few months. The problem is with internal and surrogate recoveries in standards versus blanks. Or, dirty samples vs clean samples. This happens for both water and soils. Standards will give proportionately higher values and tend to increase when run consecutively, when blanks follow the values slowly drop until they are about 60% of the mid point standard values (takes about 15 runs) Running more standards will bring the values back up. The first blank values are usually the same or even higher than the preceding standard, yet analyte carryover from a 1000 std is less than 1%.
Virtually the entire flow path of the Atomx has been replaced, except for the trap mount (that was cleaned). GC inlet maintenance performed, new column, source cleaned (and replaced), ETP Multiplier replaced, and method modified from 2 min desorb down to 1 min, and dry purge increased to 1.5 min with absolutely no effect. Flows have been verified.
The data suggests the problem would be active sites in the Atomx. The newest ATOMX is also showing this behavior but it is not failing the 50% limits yet.
Your help or suggestions are appreciated (or condolences).

[Yama001]

This issue seems to crop up a lot - the regulars here focus on water in the MS, which can often be the problem.

I had a similar issue on a 5973 - it went away after replacing the repeller and the draw out plate (6 mm hole).

The water rational fit our symptoms, but it pretty much never came up again after replacing those parts.

Its always nice to get a clean flow path in the P&T concentrator, so I would not consider that effort wasted. If nothing else, you can rule out flow path issues, at least until you get a foamy sample again .

[lcgcscientific]

Thanks, yes I have heard the larger drawout plate can help.
I will try that next. I have seen water build up in the MS cause decreasing responses through the run similar to this, but the values will be higher after allowing the system to rest or recover. I have baked (muffle furnace) the ceramics and found that to help. Here they continue to drop until a high sample or standard is run.

[cleh]

What trap are you using and what are your P&T parameters?