Dionex ICS3000 problems in running EPA method 300.1

[yangz00g]

We are currently transferring the EPA 300.1 (disinfection byproducts) from a old Lachart IC to a newly aquired Dionex Reagent Free ICS 3000 system. Surprisingly, we had much worser sensitivity for bromate/chlorite than excepted and even than the old Lachart system.

The conditions: Dionex AS9-HC 4x250mm, 9mM sodium cabonate, 1ml/min flow rate, Ultra II 4mm suppressor, suppressed conductivity CD (not a CD 25), 200uL injection using a 200uL loop, basline is about 25uS, retention times are close to suggested value. At 5ppb, vitually no peak for bromate/chlorite, at 10ppb, just like a bumps, at 50ppb, there are broad peaks. We tried two columns, one is nearlly new. Since I used the exactly the same condition suggested by EPA300.1, except a different (or better ) suppressor and a CD, I cannot help suspecting that the CD is the culprit.

Anyone has any input? Thanks

[Chris Pohl]

yangz00g,

What baseline noise are you observing with your system? It shouldn't be significantly more than 2nS. If it is, you may have inadequate back pressure on your conductivity cell. Have you measured the back pressure on the cell?

Chris

[yangz00g]

Thanks, Chris.

The baseline noise is always less than 1 uS. I haven't checked the cell pressure yet, I will give it a try. Thanks again.

[Chris Pohl]

yangz00g,

The noise needs to be less than 2nanoS so it should of course also be less than 1microS. After I reread your message, though, it looks like your problem might actually be due to overload conditions. The depending upon the chloride and sulfate concentrations in your sample you may need to either inject a smaller volume or use sample prep to remove the chloride and/or sulfate. If you inject a common anion standard at low ppm concentrations, are your peak widths and peak shape normal?

Chris

[yangz00g]

Chris, for our instrument, it seems possible to reduce the nosie level to 2nU or even 20nS with a baseline signal at 25 uS. When it was installed, we were told a noise < 1 uS was OK.

we are having trouble with low concentration ions in drinking water such as bromate (usally < 10ppb) and chlorite (<200ppb), not high one like sulfate or chloride. We follow exactly the procedures of EPA 300.1 no any change.

Another thing troubled me is the peak is very broad with a 0.5 min baseline peak-width, if there is a peak at low concentration range (<50ppb). Thanks

[Chris Pohl]

yangz00g,

As I mentioned earlier, you should be able to achieve noise of 2nS under your operating conditions. If you aren't, there may be an issue with your cell back pressure. I'm not sure why anyone would have told you that noise <1uS would be okay since that is about 500 times higher than it should be.

Regarding the observation that chlorite and bromate peaks are broad at low concentrations, this is most likely an effect of overloading by the major ions present in the sample. If you inject them by themselves at low concentrations, you should observe normal peak widths. Most likely this is caused by relatively high concentrations of chloride and/or sulfate in your sample. This is the reason why mentioned this possibility in my earlier post. In such a case you need to either remove at least a portion of these ions from the sample or dilute the sample.

Chris

[yangz00g]

Chris, Thanks a lot.

I took a close look at the noise level, at 25uS background, the nosie is less than 2nS, so we can exclude the detector pressure problem.

The standards I used were in pure water, so there is no interference from chloride and sulfate ions, and they were also freshly prepared everytime.
Service guys from Dionex felt that the standards were wrong, after they saw the chromatogram I sent them. Bbut definitely it's not the case, I have asked another analyst in my lab to prepare all standards from scratch, the signals came out close. I really couldn't figure our what goes wrong.

[Chris Pohl]

yangz00g,

Well, if you're seeing broad peaks at low concentrations and your standard doesn't contain matrix ions at typical concentrations, you should check the column to see if there's something wrong with it (even though it may be brand-new). My suggestion is to inject the test mixture used to validate the column to see if you can reproduce the original chromatogram. That will allow you to determine whether or not the column is the root cause or if there's some other problem affecting the early part of the chromatogram.

Chris

[yangz00g]

Thanks for your suggestion, Chris,

That's something I am going to try, but the text mixture for this cloumn contains ions at ppm level, I am certain at this level, the peak will be fine (the surrogate DCA gives nice peak at 1ppm, but only 30% of the peak area I got from the old machine). If it still doesn't have improvement, I am thinking of a AS19 column, Dionext claims it has better sensitivity for bromate than AS9-HC.

[Chris Pohl]

yangz00g,

When you say that your area is 30% of what you used to get on the old machine, does this refer to a different, earlier model Dionex instrument or an instrument from some other manufacturer? Different vendors have different practices with regard to how area is expressed so the area might be different because of this. Also, are you using the same software in both cases? Different software products have significantly different practices in terms of units which can also affect area, at least the numerical value expressed in the report. Of course, what really matters is the signal-to-noise ratio not the absolute value of the area. Is your signal-to-noise ratio worse?

You are right, of course, that you will get lower detection limits if you use the AS19 because the carbonate background with the AS9HC adversely affects both signal and noise compared to what is achievable with a hydroxide eluent. Since the AS19 is used with hydroxide eluent, you will have a lower background, less noise and a larger signal. This combination of factors typically yields a five-tenfold improvement in sensitivity.

[yangz00g]

Chris,

the s/n ratio from a old Lachart machine is definitly better than ICS3000. At 5ppb, I can see nice peaks for both bromate and chlorite within the window of 1ppm DCA, but with ICS3000, the software couldn't recognize there are peaks, you have to zoom in to manually pick broad "bumps"and define them as peaks (the min area thershold is set at 0) , you can say it's just a basline fluctuation not even a peak. Same column, mobile phase. We bough this ICS3000, becuase we assumed it has better sensitivity with advanced (?) suppression technique, but seems we are wrong.

[Chris Pohl]

yangz00g

I can assure you that when you're a instrument is operating properly in that it will provide better signal-to-noise ratio than anything you achieved on the Lachat instrument. If you aren't seeing that improvement, you should contact the Dionex service department and have somebody take a look at your instrument. Something is definitely not right.

By the way, I never asked about your suppressor conditions. What current setting are you using and what is your operating flow rate?

Chris

[Chris Pohl]

yangz00g

I can assure you that when your instrument is operating properly that it will provide a better signal-to-noise ratio than anything you achieved on the Lachat instrument. If you aren't seeing that improvement, you should contact the Dionex service department and have somebody take a look at your instrument. Something is definitely not right.

By the way, I never asked about your suppressor conditions. What current setting are you using and what is your operating flow rate?

Chris

[yangz00g]

Chris,

I am talking with Dionex service depratment for a while, they haven't been able to figure out what's wrong.

I tried various flow rates from 0.8ml to 1.2 ml/min, the suppressor current was suggested by the software once you put in your flow rate and mobile phase infor. I pick a few values across the recommended range, no improvement.

I am certain this machine will do what the method requires since EPA specifies Dionex IC and column in their written method, I just could not figure our what setting went wrong or might not correctly been set duirng the setup. The machine is two month old (in my lab).

Thanks