Purge and trap problem

[JGK]

We are having an intermittant problem with an EST8100 purge and trap sampler and quite frankly we're baffled, so I'm hoping the board Guru's can help.


The sampler removes 10 mL of sample (from a 40 mL vial), meters in the surrogate, traps then releases in to the instrument.

The problem is that occasionally during sequences of > 10 vials we will get an analysis where the surrogate peak responses are normnal but sample/standard peak values are nearly double the expected values.

This does not occur at the same point in the sequence, different vial positions are affected
This does not occur at a single STD level.
The volume sampled from the vials are the same from an affected vial as an unaffecred vial.
We have even taken affected samples vials and run them again and cannot reproduce the abberation. Peaks are (qualitatively) as expected on a rerun.

Any help on this would be appreciated.

[Bigbear]

What do you mean when you say internal standard and surrogate esponses are normal? Do their areas match those of the calibration standards?
My first thought is there may be an intermittant issur with the internal standard loop. Can you run a series where you inject the internal standard manually? You could empty the IS vessel and inject the IS into the vials with a syringe ( adjust the volume so thar it equals what the autosampler delivers).

[JGK]

What do you mean when you say internal standard and surrogate esponses are normal? Do their areas match those of the calibration standards?
My first thought is there may be an intermittant issur with the internal standard loop. Can you run a series where you inject the internal standard manually? You could empty the IS vessel and inject the IS into the vials with a syringe ( adjust the volume so thar it equals what the autosampler delivers).

Yes, the surrogate responses are normal across all injections in an affected sequence, it is only the non-surrogate peaks in a particular injection that are affected.

Our short term solution is to inject the surrogate into the vials directly rather than via the sampling system.

In he long term we need to find the root cause and be able to use the unit as intended. The unit supplier is not being of much help on this.

[Bigbear]

Has the system ever worked properly?
What was the last thing done to the system just prior to noticing the issue?

[JGK]

this is a system we "had in reserve" for a while. When we first tried to get into operation the ingjection response was too variable. we had the system ovehauled and that problem appears to hve been settled.

Currently, given the symptoms we are seeing, I'm not sure if there is a fault in the vial sampler most of the time or just periodically. Since you can't make something out of nothing, I'm leaning towards the former. Performing a sequence with surrogates added to the vial should be definitive as, if the glitch occurs, they should be affected as well.

[Bigbear]

As I suspected earlier, it may have something to do with the way the IS is delivered. I'm not familiar with your system. Mine uses a "sample loop" that is filled via gas pressure. The "loop" is really one path in a solenoid valve. Prior to the sample being delivered to the sparger, the valve rotates 90 degrees and the sample passes through the IS "loop" part.
I have had problems in the past in filling the loop due to blockages in the tubing providing pressure and liquid to the loop.

[sidz28]

Hello, JGK.

Full disclaimer - I work for EST in technical support. I'm not sure if your difficulties have been with us, or with the service group that covers your territory for us, if you aren't located in the US. If it is with us, I sincerely apologize.

Reading through your comments, it seems to me that the 8100 is delivering the same amount of surrogate compound every sample. How are you measuring the amount? Is it quantified with an MS and an internal standard compound, or are you going by area count of the peak? What type of detector are you using if you're not using a mass spec?

It's an odd presentation, as the IS addition of the 8100 is a fixed-loop system similar to what Bigbear described. The 8100 is basically an 81-position version of the Varian Archon and OI 4552, so anyone familiar with those systems IS valves will know that you have a 1uL (or in some cases a 0.5uL) notch in the valve rotor that will fill with solution, then rotate back to be in line with your sample as it's delivered to the concentrator.

So for you to see that same amount of standard added for every run (and matching up with hand spiked amounts) leads me to believe that the standard addition of the 8100 is OK. Unless I'm misunderstanding, it sounds like the problem is that you recover twice as much as you expect of your spiked target analytes. Is that correct? I can't think of a logical explanation for that. Theoretically, I guess you can add twice the volume of sample, but you should notice that in the remaining volume in the vial.

Can you give a more thorough explanation as to what you're seeing? Specifically how you're measuring responses (detector type, measuring in qualitative area counts or quantified concentrations with an internal standard)?

If you prefer, you can email EST support directly at support@estanalytical.com. We can do our best to help remotely.

Mike

[James_Ball]

Are you quanting with an internal standard and is that internal standard in the same mix as the surrogates?

If yes above, does the area of the internal standard change on the bad samples?

If yes, then the problem is you are not getting a full 1ul into your loop when the valve fills with internal standard, possibly because of a gas bubble or a small particle of some kind occasionally blocks the tube that goes into the internal standard/surrogate vial. I have had this happen on our Archons before, (we have both models, the 8100 and the regular units). This will cause the surrogate to quant correctly but all other targets will be higher than expected.

Most other problems I can think of would be just the opposite, where you would have low recoveries of analytes versus internal standard/surrogates. These problems occur when the sparge tube fails to drain because the needle is catching gas bubbles coming from the sparger during bake and it pulls out gas instead of liquid. This leads to an over fill of the sparger on the next run which often will then completely drain on the next bake, but the carryover from that should also increase your internal standard response. The other problem I have encountered is when the switching valve in the concentrator does not completely turn which leads to a lower flow rate to the GC and it alters the split ratio and causes a higher than normal area count, or almost no response depending on how badly the flow is restricted.

[Yama001]

You folks are handling the issue well, I would just like to point out that EST is one of the best organizations I have worked with on instrument issues. If you are not dealing with them, they are well worth checking out.

[Bigbear]

How do you prep your standards? Adding too much MeOH can be a problem as it effects the trapping of compounds. There is a maximun percentage that for now escapes me.
We use a 5ul is loop for a 25 ml sample in our system. The highest amount of MeOH we use is for our high standard ( 50ul MeOH to a 50 Ml volumetric containing water).

[James_Ball]

How do you prep your standards? Adding too much MeOH can be a problem as it effects the trapping of compounds. There is a maximun percentage that for now escapes me.
We use a 5ul is loop for a 25 ml sample in our system. The highest amount of MeOH we use is for our high standard ( 50ul MeOH to a 50 Ml volumetric containing water).

I currently use up to 600ul/50ml for my highest standard and don't really see any problems with linearity on my waste water analysis. I used to see the classic problem of having the Internal Standard rise substantially as the standard concentration increased and thought it was the MeOH that was causing it. I now can run about any amount of MeOH without much problems.

I found a publication by Agilent the described the problem and a solution. Reducing the EM Voltage to the lowest possible setting while maintaining the sensitivity needed, even if you need to reduce split ratio, helps linearity a lot. I was having problems getting enough sensitivity to run UCMR3 samples by EPA524.3 and I was adding up to 500v to the EM but losing response. What was actually happening was I was losing signal to noise. Dropping the EM down and setting my Threshold to 10-20 counts and lowering my split to 30:1 I was able to get great analyte peaks at concentration as low as 15ppt while reducing my Internal standard response from 7,000,000 counts at 5ppb to 70,000 counts at 5ppb and instead of my 1,4-Dichlorobenze-d4 increasing as much as 100% from low standard to high standard it now increases maybe 10% sometimes less. The effects of the water peak also are less pronounced which surprised me with such a low split ratio.

I do still see the rising IS versus standard concentration when my Moisture Trap gets dirty, but I have started swapping out the tubes more often to help with that.

[JGK]

Sorry for the delay in getting back to you all but I've been discussing the issue with the technicians and scientists who are using the system.

First, and I apologise, but we aren't using surrogates (someone in SOP production wasn't using the terminololgy correctly). We take 10 mL of sample and meter in Internal Std prior to the trap. We've checked the sampler and it's taking up and using the full 10 mL.

Periodically, in longer sequences we will see an injection of a std/QC sample where the analyte responses in a STD or QC sample are 1.5 - 2 x expected values. The vial (10 mL sampled from a 40 mL vial has not been over sampled (visual check) and the IS responses are in line with IS responses all other QCs and STDs in the sequence.

At present we have only checked out STD and QC solutions on this injection system as we get it ready for the production environment. However, we have not seen excessive variation in area counts in the IS from the MS detector (variation in area counts is at worst case ~13%. Consequently, we don't suspect an error in IS delivery and based on the lack of variation we don't suspect a trap issue (although we have tested a second trap which did not cure the problem).

We have re-run the same sequence using the same vials (now containing 30 mL in a 40 mL vial) and the vial which prevously exhibited the analyte resonse increase no longer shows the same effect, peak responses are as expected Is responses are comparable to the other vials within the sequence.



When it occurs it appears to be random and not associated with specific STD/QC level, vial position or injection in a sequence. It may/may not occur in a sequence.

[Bigbear]

What IS are you using. Have you plotted the IS areas over time to see if there is a trend?
Does the area drop durring a run? Does it "come back" after the instrument sitting idle over nite?

[Peter Apps]

When you say "values" do you mean peak areas ?, or the analyte peak areas divided by the internal standard peak areas ? If only the analyte areas are higher than expected, with internal standard areas as big as expected, then the source of the problem is something to do with the sample. If the ratio between analyte and internal standard areas is high it can be because analyte areas are high, or internal standard areas are low; referring to "values" provides no information as to which it might be.

Peter

[Yama001]

James_Ball, if you do not mind, can you post a link to that Agilent pub? I think I have read that one as well but I would like to look it over again. Thanks.