Anion exchange chromatography problem

[Edde]

I guess you mean increasing the target analyte concs. instead of increasing the mobile phase salt conc.

[Edde]

With 30mM phosphate buffer in 10% ACN pH8, isocratic run, I got nitrite at about 6.7 min and nitrate at about 10.4 min.
Not much interference was seen in serum matrix. However, in blank urine there was coeluting interference peaks at the target RTs.
In order to separate the interferent peaks from my target analytes, should I try adjusting the pH, changing salt type and/or conc. and adjusting modifier (ACN) conc. ? Which parameter should I try first and which next?

Please comment. Thank you very much.

[tom jupille]

No easy way to predict without knowing something about the interference.

If you suspect that your interference is a weak acid, then changing the pH would be the most promising approach (change the analyte's degree of ionization).

If you suspect that the interference has a reasonable level of hydrophobicity, then tweaking the ACN concentration would be best.

If you suspect that the analyte is multiply charged, the adjusting the ionic strength would be the best approach.

In practice, probably best bet to try them all; it only takes three additional runs to see which parameter affects peak spacing:
- double the ionic strength (keep pH and % ACN constant)
- change the pH by 0.5 unit (keep ionic strength and %ACN constant)
- increase ACN from 10% to 15% (keep pH and ionic strength constant).

[Edde]

I already used 30mM KH2PO4 buffer. My question is doubling the ionic strength will be 60mM. I don't know whether 60mM will be too high so that there will be essentially no retention and the capacity factor is too low.

change the pH by 0.5 unit (keep ionic strength and %ACN constant)

Should I try both +0.5 and -0.5 pH unit?

increase ACN from 10% to 15% (keep pH and ionic strength constant).

The column insert recommends less than 12% ACN should be used. I think I should try 12%. And should I try lower conc., say 5% ACN, as well?
I appreciate your kind comments.

[tom jupille]

Remember that the idea here is not to "get it right" on the first try, it is to see which of those parameters has a significant effect on the spacing of the peaks. Once you have established that, you can do more experiments to focus on that parameter and "fine tune" its value.

If it were my problem, I would make one significant change in each parameter and see what happens. If the interference still coelutes, then that parameter is not effective and you can keep it at it's initial value.

In ion exchange, the relationship between retention (k') and ionic strength is essentially log-log; for fully ionized species, the slope (called the "Z value) is approximately the ratio of the charge on the analyte divided by the charge on the driving ion. In your case, both are monovalent, so Z should be about 1, and so doubling the ionic strength should cut the retention k' in half (okay, this is a gross oversimplification, but it should be in the ballpark. ). If you want to be on the safe side, increase it by 50% to 45 mM.

I'd only do one pH experiment (for the reason cited in the first paragraph). Toss a coin to decide on the direction.

Same idea with %ACN; you only need to go in one direction initially. Since you're pushing close to the recommended 12% maximum, in this case I'd go down.

The following rant is not directed specifically at you, but when we teach method development courses, we have to spend a fair amount of time disabusing people of the notion that effective method development consists of getting the best conditions on that first try, but rather of mapping how the system responds to changes in key parameters so that those parameters can be optimized systematically. The occasional lucky guess -- like winning the lottery -- is a tremendous emotional high when it occurs, but it's not a good long-term strategy!

[Edde]

Thank you so much for your clear explanation.
I can't describe how enormously I benefit from your invaluable comments and elaborations.

[Edde]

I tried 3 mobile phase conditions:
1. 10% ACN, 45mM KH2PO4, pH8
2. 10% ACN, 30mM KH2PO4, pH 8.5
3. 5% ACN, 30mM KH2PO4, pH8.0

There were a shift in target analytes' RT. But there were still interference peaks coeluting with the target analytes.

Is it appropriate to try gradient elution next?

I use commercial blank urine either unspiked or spiked with nitrate and nitrite standards. My question is commercial blank urine may contain preservatives or interfering substances which may not be present in normal urine. I don't know whether it is better to use fresh urine instead.

Thanks.

[tom jupille]

Ok, the question is "were they the *same* interferences coeluting (hopefully there was enough variation in peak sizes to tell)? If not, then selectivity changed and you should be able to interpolate (or extrapolate) to find acceptable intermediate conditions. I would be very surprised to find the same interferences coeluting with your analyte with all those changes in conditions. Are you sure that what you are interpreting as "interferences" are not actually endogenous nitrate and nitrite in your urine samples?

However, this brings up a point: are there any "clear" regions on the chromatogram where you don't have matrix peaks? If not, then you are fighting a losing battle, because no matter where you put your analytes, there will be *some* coeluting interferences. In that case, the only solution may be some sort of sample cleanup prior to the ion exchange chromatography.

[Edde]

Thank you very much for your inspiring comments.

the question is "were they the *same* interferences coeluting (hopefully there was enough variation in peak sizes to tell)?

At one condition, interferent A coelutes with nitrite and at another condition, interferent B coelutes.

Are you sure that what you are interpreting as "interferences" are not actually endogenous nitrate and nitrite in your urine samples?

The interference for nitrite is not endogenous nitrite. The UV spectra are different. The interferent for nitrate is endogenous nitrate + another interferent.

are there any "clear" regions on the chromatogram where you don't have matrix peaks?

You are right, there are no clear regions around the target analytes on the chromatogram.
I tried QAX1 SPE which is a quaternary ammonium anion exchange SPE. Procedures as follows:-
Filter urine through 3kD centrifugal filter. Filtrate used as sample.

2mL MeOH, 4 mL water conditioning
Load 1mL filtrate
Wash with 3 changes H2O
Elute with 2mL 0.5M NaCl

In blank urine, there was still interferent peak at nitrite's RT.
No nitrite, nitrate recovered in spiked urine.
I have an idea but not sure if it is true. Will the NaCl present in the eluate affect the RT of nitrite and nitrate so that I can't see them in their usual RTs?

Thanks.

[tom jupille]

At one condition, interferent A coelutes with nitrite and at another condition, interferent B coelutes.

That implies that there must be some intermediate condition where neither A nor B coelute with nitrite (of course, that still leaves interferents D, E, . . . ).

You are right, there are no clear regions around the target analytes on the chromatogram.

That pretty much defines your problem. One way or another you have to find or make room on the chromatogram to accommodate your analaytes, or you have to get a selective detector (e.g., electrochemical detector for nitrite) that will be "blind" to the interferences.

Will the NaCl present in the eluate affect the RT of nitrite and nitrate so that I can't see them in their usual RTs?

If you're injecting large amounts, there *is* a chance that you will see some "displacement" chromatography going on. The only way to tell for sure is to inject standards with increasing amounts of chloride.

[Edde]

That pretty much defines your problem. One way or another you have to find or make room on the chromatogram to accommodate your analaytes, or you have to get a selective detector (e.g., electrochemical detector for nitrite) that will be "blind" to the interferences.

Since I don't have another detector, I will try gradient elution to see if I can make room to accommodate the analytes.

If you're injecting large amounts, there *is* a chance that you will see some "displacement" chromatography going on. The only way to tell for sure is to inject standards with increasing amounts of chloride

I will also try injecting stds in NaCl.

Thanks for your suggestions.

[tom jupille]

I will try gradient elution to see if I can make room to accommodate the analytes

It can't hurt to try, but I would not be optimistic. Gradients are particularly useful to accommodate a wide range of retention, not to improve resolution within a given range.

The other option is to improve the efficiency (plate number, N), which means making the peaks narrower so that you can resolve more compounds in the same run. The catch here is that resolution is proportional to the square root of N, so that it takes a big increase in N to have much effect on resolution.

There is a question that I should have asked earlier, how large are the interferences relative to the amounts of nitrate and nitrite you will be measuring? If they are small, you may be justified in simply ignoring them.

[Edde]

The other option is to improve the efficiency (plate number, N), which means making the peaks narrower so that you can resolve more compounds in the same run

How can I increase the efficiency? I don't have another column with smaller particle size.

There is a question that I should have asked earlier, how large are the interferences relative to the amounts of nitrate and nitrite you will be measuring? If they are small, you may be justified in simply ignoring them

The interference for nitrite is huge. The nitrate interferent peak might be endogenous nitrate. But the UV spectrum look similar but doesn't match exactly. Although not very high, it can affect quantitation.

Actually, nitrite appears sandwiched between 2 interferent peaks. These 2 interferent peaks can't be baseline separated. When I adjusted the MP pH, this time lowered to 6.0, the 2 interferent peaks for nitrite were finally baseline resolved but the later eluting interferent peak still coeluted with nitrite. The pKa of nitrite is 3.4. I wonder if I should lower the MP pH further to resolve nitrite and the coeluting peak? Howerver, the nitrite peak shape is broad already.

I also tried filtering the urine and then carried out C18 elution as follows: Filtered urine was passed through C18 SPE and the flow-through was collected for analysis. The afore-mentioned coeluting peak diminished in size, peak area is ~350. I am not certain if it is small enough to be ignored.

In this pH, the nitrate peak is even broader. What can I do to make it sharper?

Thank you very much for your direction and guidance.

[Edde]

I tried 60mM KH2PO4, 10% ACN, pH6 on filtered then C18 SPE treated urine. The 2nd interfering peak disappeared. I don't have interfering peak for nitrite now.
However, the nitrate peak is not resolved into 2 non-baseline resolved peaks. I intend to try increasing KH2PO4 conc. to 100 mM to see if the peak can be further resolved.
The peaks are still quite broad.

[tom jupille]

Sounds like you're making progress!!!

How broad are the nitrate and/or nitrite peaks? What are the plate numbers?