Newbie HP5890 EPA Method 8260

[ardin825]

Good afternoon,

I was just hired to start up an organics testing program (EPA 8260/8270) at an environmental lab. While I'm an organic chemist, I don't have experience running GC/MS beyond the one or two times injecting and pressing start as an undergrad. We are using HP 5890 Series II with a very old version of ChemStation (G1701BA). I have all the manuals I need, but I'm still having trouble setting up the calibration curve and getting sensible response factors. If someone could give me a quick set of instructions for how to use the software to do this, it would be a lifesaver. The instruments were set up and run by a previous team of scientists, but they're no longer available for help, and no one else in this lab has any experience whatsoever with this software.

Here's what I have done so far:

1. Ran my calibration samples through the P&T and GC/MS
2. Clicked "Quick Levels Update" under InitCal
3. Clicked "Yes" (clear responses first)
4. Clicked "Yes" (requant files before update)
5. Chose the files to use for calibration.
6. Clicked "Response Factors to Screen" under InitCal

The response factors are completely useless--although viewing the several calibration spectra overlaid on top of each other shows that the intensities increased with increasing concentration, I'm getting RSDs well over 100%. I have no idea how to fix this problem, and when I read the software manual, it takes me through adding all of the compounds, but they're already set up in the system and we don't want to change the library. Any advice would be super fantastic. Thanks so much!!

[Yama001]

A useful check on your standards - if you use chlorobenzene-d5 as an internal standard, the chlorobenzene response factor should be very close to 1.0. If it is not, you may not have set the quick levels update correctly, or your spiking may in fact be off.

[MichaelVW]

I'm using a 6890 with software version G1701DA - so I don't know how different this is from what you have.

Do you have the old method saved? Are you able to pull up one of the calibration standards you ran using the old method to see if you get decent-ish results? I'm not sure what maintinance you did... it's not possible, for example, that your RTs have changed significantly and it's not properly IDing compounds, is it?

Usually after I run cal standards I QEDIT my way through each one to make sure each compound is identified correctly and integrated well. (Quant -> Qedit quant results). Then I do the quick levels update as you did, except I say "no" to "requant files before update?". I then look at each individually by going to InitCal -> Edit Compounds-> View. "Page 3" shows me the responses that I got for each level and "Plot" shows me a graph of the response factors and an RSD.

Maybe try playing around with some of that. If you're still having trouble, try showing us what your resposne factor table looks like - maybe there is a clue there.

[ardin825]

MichaelVW, thanks so much--this is very helpful. Indeed, the QEdit shows that the software is integrating at the wrong locations. How do I fix this? I can change the retention times in the library, but it seems like doing some manual integrations in QEdit makes more sense. Would you be able to walk me through your procedure for when you run into this problem?

Thanks again!

[James_Ball]

Run a blank with only your internal standards, then use the menu item Tools/Easy ID which will let you manually integrate your internal standards(just double right click in the area where your response shows up and it will take you to the next internal standard) and when you are finished it will update the rest of the compounds relative to those changes. It isn't 100% accurate but it usually gets them close to the window you are looking for. Then take a mid level standard and manually integrate each compound. After saving the changes, go into the calibration update window and choose recalibrate and the retention time option and replace. This will set the new retention times. Save the method then re-quant the files and check each one manually to see that the peaks are correctly identified. Then you can do the update of the calibration and the curves should look better.

Note: Make sure the internal standards are all set to the same concentration. You should be spiking the same amount into each standard and sample you run, and if you have varied concentrations in the calibration table it will throw your response factors way off.

[ardin825]

Thanks everyone! I'm still working on it; not great yet, but getting better.

[whdees]

Thanks everyone! I'm still working on it; not great yet, but getting better.

just my 2 cents...I would not use "Quick Levels Update", I have had it fail in the past. It will look like it is working, but the changes aren't made correctly. Shoot your cal standards, examine each peak in each shot, and update the cal curve one level at a time... it takes a while, but if your data is good, the curves will be good. Enviroquant has some strange quirks and odd bugs that pop up from time to time... good luck!