Problem with negative peak in Ion Chromatography

[fatBrain]

Help needed. Thanks in advance.

We have a new IC system ICS-2000 from Dionex, which was modified to use external water to regenerate suppressor. The analyte is bethanechol Cl and we follow USP method exactly. The detailed method is:

MP: 29mg EDTA and 0.3 mL HNO3 in 1 L water. 950 mL the solution mixed with 50 mL of ACN.
Column: Waters IC-Pak Cation M/D, 3.9x150 mm
Cell temp: 35C
Column temp: 30 C
Suppressor: CSRS-Ultra II at 15 mA (for cation)
Sample and standard prepared in 0.1 N HCl
Bethanechol elute around 11 minutes

We had no problem running samples using the system until recently. The column appears degrading pretty fast and at certain point we saw the main peak getting a shoulder (same to other peaks from sodium, etc). After I repaced with a new column, we were surprised to see a well-defined negative peak at same retention time (11 min) but no bethanechol peak. The area count slightly smaller with similar peak shape. Sodium peak in the front is huge and positive as usual.

Have tried:
1. Injected known-good sample and standard;
2. Used known-good MP;
3. Flushed the system with water;
4. Flushed the suppressor with 40ACN/60water/1Methanesulfonic acid;
5. Put back the bad column, and positive bethanechol peak no longer there;
6. Inject 0.1 N HCl, and no peak at 11 min;
7. Replace the external water for supressor regeneration.

The only thing i am not sure is the working condition of suppressor, but even a weak suppressor shouldn't produce a negative peak at 11 min. When the system was good, total conductivity was about 0.6 uS and now it is about 1.5 uS. Where did the additional 0.9 uS come (if the reading means something)? When I turned off suppressor I can see base line gradually moving up, indicating it is at least working. There is nothing I can think of except getting the service tech coming to my place with a new suppressor. But they were quite confused on the phone.

A negative peak means conductivity at that retention time is lower than the back ground, which is not necessarily unusual, but how could that be the negative peak elutes at 11 minutes? What else did I miss?

Having working on LC for many years, I got really frustrated for the first time. Any suggestion is very welcome. Hope the new suppressor will solve my mystery.

[Uwe Neue]

What is the mobile phase used to ship the IC-pack to you? Maybe it is still bleeding slowly a consituent of the old mobile phase, and it does so less, when you are injecting your bethanechol.

If the negative response is stable (=repeat injections give you the same result), you may be able to run the analysis with the peak going either way.

[Chris Pohl]

fatBrain,

I think your problem may well be that you are using a method designed to be used without a suppressor, with a suppressor. The suppressor is designed to remove an anionic species from the eluent but there is a limit to how much it can remove in the case of polyvalent species such as EDTA. Most likely, given the low concentration present in your mobile phase it just took a while to convert the suppressor to the EDTA form. Until you get it back into the hydroxide form, you most likely will not see positive peaks again. Changing the suppressor will help temporarily but you need to regenerate the suppressor periodically to remove the contaminating EDTA if you want to restore the suppressor.

Are you stuck with this eluent system and column? You should be able to separate and detect this species directly using a simple methanesulfonic acid eluent on the CS17 column.

[fatBrain]

Thanks both you.

Uwe was right about this column. This column degrades relatively fast and we can get a few hundred injections at most. But it is still in USP, ouch!

Chris, it is hard to believe that the impurity method can be run under nonsuppression mode. For assay, it is just fine. Your explanation about negative peaks is reasonable. After the suppressors are converted to EDTA form, it just acts like a strong cation exchange column and retains my analytes. I have a question for you, Chirs:

Actually the regeneration approach provided by Dionex didn't work well, say, for organic contaminants. In your opinion, what is the best approach to get rid of EDTA or other negative ions retained on suppressor if it is true? Use NaOH or TBAOH? And acetonitrile? Thanks in advance.

[Chris Pohl]

fatBrain,

I think you may have misunderstood my last post. I wasn't trying to say that the nonsuppressed mode was good for the impurity peaks (i.e. better for trace analysis). I was only trying to say that the eluent looked like it was designed for use in such a system. You will definitely get better detection limits with the suppressor. If you are stuck with the aforementioned eluent system, you may be forced to live with the sensitivity disadvantages of nonsuppressed detection (your instrument is fully compatible with this operating mode).

There's another problem with your eluent system that I forgot to mention in my last post. The CSRS should not be used with nitric acid eluent in the electrolytic mode as this causes oxidative damage in the anode chamber of the suppressor (this is spelled out clearly in the suppressor manual). It's conceivable that you might get this method to work even with the EDTA present since it's present at such a low concentration in your eluent but you definitely need to get rid of the nitric acid if you want to use the electrolytic mode. In the chemical mode, you'll have a difficult time suppressing this concentration of nitric acid. Can you substitute methanesulfonic acid? If you can, you might be able to get this method to work in the electrolytic mode.

In addition, if the "0.3 mL of nitric acid" you are using for the eluent preparation refers to 0.3 mL of concentrated nitric acid, the current setting you are using is way too low for the suppressor in the electrolytic mode even if nitric acid wasn't problematic. Also, what flowrate are you working at? From what source of information did you arrive at the 15 mA current setting?

I assume there's a specific reason why EDTA is present in your eluent. Do you know the reason for this? If you are free to modify the eluent system, I would advise switching to a suppressible chelating agent in the place of EDTA. For example, oxalic acid will work reasonably well as a chelating agent for many transition metals and is readily suppressible.

Even though one could reasonably classify EDTA as an organic contaminant, the best approach to removing it from the suppressor is to treat it as a hydrophilic polyvalent ion (the organic contaminant removal protocol is designed for fairly hydrophobic molecules where solvent is necessary as part of the removal process). Removal of hydrophilic polyvalent ions is best done with a fairly concentrated electrolyte in the absence of solvent. To quantitatively elute a polyvalent species, you need use a high concentration of sodium hydroxide. You can elute the majority of the EDTA with a few milliliters of 0.5 molar sodium hydroxide but if you want to remove it quantitatively you should pass 0.2-0.4 L of 0.5 molar sodium hydroxide through the suppressor. My worry, though, is that the nitric acid eluent combined with the electrolytic mode of operation may have caused irreversible damage (the rate at which the suppressor membrane and screens are damaged is fairly slow with nitric acid so you may be able to restore the suppressor assuming you haven't operated for too long in this mode). In the case of extended operation in the presence of nitric acid, you won't be able to restore the suppressor.

[Einar Ponten]

At this low pH of the mobile phase the EDTA is not much of an effective chelating agent, while it will become one after the suppressor and during the neutralisation. Seems to be a great risk that it binds and deactivates the functional groups on the membrane or is it the bleed that is trapped?

Strange that this has qualified as an USP method. Anyway, HILIC would most likely be more suitable technique for the analyte.

[Markus Laeubli, Metrohm]

As Chris mentioned the method was defined for non-suppressed cation chromatography.
We recently ran a similar application (see http://www.metrohm.com/infocenter/downl ... f/c-80.pdf and http://www.metrohm.com/infocenter/downl ... f/c-81.pdf).
Whether non-suppressed or suppressed will give the better LOD is still under discussion.
Anyway Chris would use suppressed and I would use non-suppressed IC.