Basic question on validation samples

[PeteMartin]

I use my high standard right now as my validation sample.

I do mainly soil chemistry by way of water extraction and am using the DX-600 system. I'm getting troubles with reproducibility of the validation sample concentration. It can be off by as much as 20 - 50% sometimes!

Could this be because I'm near the upper limit of my calibration curve? Would using a lower standard give a more reproducible reading over several injections? Or would that just be cheating myself out of solving the real problem?

I've tried nothing and I'm all out of ideas. Just kidding...I've troubleshooted just about everything I can think of. I just want some opinions before I take the seemingly easy way out and just use a less concentrated standard as my validation sample.

Thanks for any help.

[tom jupille]

If your high standard varies by that much, then you're waaay above the upper end of your linear range. There's another discussion going on right now that may have some bearing on your situation:
http://www.sepsci.com/chromforum/viewtopic.php?t=2857

If it were my problem, the first thing I would do is to check the repeatability of a lower level standard. If that looks OK, then the problem at the high end is an overload issue.

If the lower standard also has a lot of variability, then you have to dig deeper. If you have more than one peak, look to see if the ratio of the peaks is more precise than the individual areas. If so, look at issues like injection or sample prep. If the ratio is less precise than the individual areas, then look at things like integration, excessive tailing, poor resolution, etc.

[PeteMartin]

Thank you very much Tom. The other thread was very helpful as well.

My primer contains approximately 7.2ug total anions/10uL injection. I'm not sure what the loading capacity is for the AS18 column. I will try a lower standard as well as some rough gravimetric analysis of my injections.

Thanks again

[Chris Pohl]

Pete,

The amount you are injecting (assuming you are using a 4 mm column) is 5-10 times the amount generally injected on this column assuming you are trying to quantitate the major components in the chromatogram.

[PeteMartin]

Thanks Chris,

Very rarely do we have that much of each anion we can detect in a sample (as in the primer) so I think as far as sample anoin concentrations, we are within about the correct range. I am going to use a lower standard as a primer and see how reliable that is.

Is there any place I can get column loading information for Dionex columns? I've looked online and in their manuals but I've found nothing. Maybe I'm looking in the wrong places.

Thanks

[Chris Pohl]

Pete,

Indeed, Dionex does not publish dynamic capacity values for their columns. You will find static capacity information in the column manual but including dynamic capacity is problematic in that it is both eluent and analyte dependent. Actually, I argued for years against including even the static capacity information since in my opinion this value is both largely irrelevant and misleading but I was overruled by the marketing department. Furthermore, there is no generally agreed upon criteria for assigning a specific value for dynamic capacity as it depends upon the application to which the column is applied. For example, you might consider the dynamic capacity to be the concentration at which the efficiency for an analyte, say chloride, decreases by more than 10%. However, area based calibrations will generally give linear results well beyond this point. So, using peak area linearity as a basis for the dynamic capacity will clearly give you a different value for dynamic capacity than the chromatographic efficiency threshold method will. Our problem is: which method should we use and which analyte should we use as our test probe. In the absence of any clear answer to these questions, we have chosen not to include such information in the column manual. Nonetheless, as a general rule of thumb I generally opt for maximum total sample concentration in the 100-200 ppm range with a 25 µL sample loop when using a 4 mm column.

[PeteMartin]

Thanks very much for the info Chris, I'll use that as a guidline from now on.

These are my current standard for anions.



I realize now that the 8 and 9 standards are probably overloading the column, but I just ran them all and even the low ones are about 30% low. I've run about 1400 samples through this column since the middle of June. I replaced the guard column at about 700 samples. Could this be a sign that the guard column needs to be replaced again? My ASRS is brand new, I have just run the standards in the method I'm quantifying these samples with, and I don't have any backpressure problems.


Just a little background on my sample prep:

-5 (sometimes 8 ) grams of soil mixed with 40 mL of chilled distilled H2O
-soil and H2O mixed on an orbital shaker @ 200rpm for 1 hr @ 4*C
-1mL sample taken from bag
-sample centrifuged @ approx. 20,000xg for 15 mins
-approx. 400 - 800uL of supernatant collected and analyzed

[Chris Pohl]

Pete,

The calibration range you are using is higher than I would generally recommend. When you say "I just ran them all and even low ones are about 30% low" what do you mean by that? Do you mean that the area counts are lower than you are used to seeing with the same calibration mixture by 30%? If so, perhaps you need to do a quick chemical regeneration of the suppressor. Even though the suppressor is brand-new, the cation exchange sites in the suppressor materials might have gotten partially in the sodium form by one of the number of possible routes. If you follow the acid treatment protocol in the suppressor manual and the area's return to what you are accustomed to than this is a likely cause of the problem. Generally this treatment completely eliminates the problem if this is caused by the outlet end of the suppressor being contaminated with eluent cation (ideally the outlet in the suppressor should be 100% in the hydronium form).

[PeteMartin]

Hi Chris,

Yes, the area counts are 30% low on all standards, even the low standards that are within the acceptable range of the column.

I think I've narrowed the problem down to either the sample or the autosampler. I made new standards and they are almost right on.

It seems that the concentration of the standards is going down.

Before you laugh, I'll tell you my reasoning. I think that this is a problem with condensation. I have made 5 or 6 10uL injections from my standards, and the initial volume in the vials was 800uL (800mg). I weighed the vials and found them to contain approx 788mg (788uL) when they should have 750uL or less.

I've already tested the injection volume this way by measuring the mass of the standard between 3 injections and it was right around 10mg (10uL) so I don't think that the autosampler is injecting the wrong amount. As well as the fact that fresh standards are reading as the correct concentrations.

I see a LOT of condensation all over the rack and even sometimes on the tops of the vials. I also see condensation on the insides of the vial walls. Is it possible that due to the cooling design of the AS50 that cools the samples by way of conduction in the presence of the warmer air above that this condensation is being pulled into my standards through the slit septa during injection and effectively diluting my standards?

In my opinion, the AS50 has a terrible design for sample cooling. It can't be good to have all of that condensation there all the time. I've even seen it where I had to take the rack out and wipe up all the water from underneath just so the autosampler will recognize that the rack is in the tray.

I have the autosampler set a 4*C and the room temp hovers around 20*C. I think I'm going to set up a datalogger over the weekend with probes in the room, the space in the top of the autosampler, in a vial with water, and against the tray. If you or anyone has any ideas on what might be going on here other that this please speak up, I'd like it to be some other problem that I can control more easily.

Thanks for reading the long post

[Chris Pohl]

Pete,

To a certain extent I can see your point but it seems hard for me to believe that you could get enough condensation into a vial even if it was wide open to account for the effect you are noting. After all, you're talking about a 30% deviation in response. Could you really have had a 30% dilution via this method? In fact, the condensation you are reporting on the vial wall should have the opposite effect. So it would be even harder to imagine a 30% dilution in this scenario. Did you manage to correct the problem by making up fresh standards? This would be a simple proof of your hypothesis.

[PeteMartin]

Chris:

As noted in my post above, fresh standards corrected the problem. I'll do some more investigation into this because I'm rather skeptical of the dilution effect as well but it's the only explanation I have right now.

Thanks

[PeteMartin]

I've done a little more investigation into this and I'm still at a loss.

The new primer that I made is now reading low. So I made another one and the first run from it gave me almost perfect results. The next 5 runs from the same vial in the same sequence were 20 - 25% low.

I agree with Chris that this is just too much of a difference to be a dilution via condensation problem. Any other suggestions? I think my next step might be to get a more indepth look at injection volume using a Metler scale.

Thanks

[KarenJ]

Pete,

About a month ago I had weird results with my standard and sample areas going down considerably after the first injection. I was using screw cap vials and it seems that I was getting the vial caps too tight and creating a small vacuum in them. The autosampler was then not able to draw out the correct amount. I finally gave up on those vials and switched to a snap-cap vial and my problem went away.

Are your samples stable enough to run them without a refrigerated autosampler? If you are worried about the autosampler condensation causing a problem, why don't you try repeated injections of a stable sample with the temperature control turned off?

karenj

[PeteMartin]

Thanks for the suggestion Karen, I tried unscrewing the lids a little and it didn't seem to make a difference.

Since this happened right after I made fresh standards on the second injection of my primer, I don't think that condensation could possibly be causing this particular problem. I don't think condensation would have an effect after only 15 minutes and as Chris said, it's hard to believe that enough condensation would be getting into the vial to cause such a large variation.

I'm now getting some more help from Dionex in my troubleshooting so I'm sure I'll have a solution soon, however simple or obscure it may be.

Thanks for all of the replies