Chromatography Forum

I think that is only correct if you have the same number of acidic and basic sites. In Molly's case it might be not true. If you have more basic sites than acidic you can consider compound as basic. If you have more acidic sites you can consider it as a acidic compound. Also if charges are separated by space you can introduce a stronger acid or base on the surface to interact with zwitter-ion. You also need to look at pKa values of your amines and acids.

In addition to that in 100% water you cannot adjust retention by decreasing amount of ACN because negative concentration of ACN does not exist. In case of mixed mode (Primesep)-you still can use ACN to adjust retention plus you have buffer concentration which will affect your selectivity. Why use RP mechanism if is not working in the first place. On the best RP column (Atlantis T3 as you say) you will have for example retention of 3 minutes vs. retention of 2.1 minutes but what’s the difference? If you have other compounds you don't have to much room (no room at all) to adjust selectivity/resolution.

First of all, thanks to Uwe Neue, Sielc_tech, Yury, Bryan, Patrik, HW Mueller and Einar for your reply and suggestions. This is really a great chromatography community.

The topic seems to generate lots of interests, and there is probably more than one answer to the separation challenges of polar molecules. But back to my specific case, here is the original problem: The bottom trace is the mixture of compounds, and I am particularly interested in peak eluting at 6min (compound 1), which is prone to degradation. Top trace showed the acid stressed solution of mixture of compound 1 and compound 2 (eluting at 10min). Notice that compound 1 is completely degraded into three early eluting peaks with the condition and column listed on the chromatogram. Here MP A is pH 6.8 buffer and MP B is ACN.


The first column I tried is Zic-HILIC column since I am really interested in the reverse of elution order and hope the polar degradates can be retained. The column dimension is 150X4.6, and with 90%B to 10%B in 15min. Bottom trace is the mixture of compound 1 and 2, and top trace is the acid stressed solution. The elution order did switch as compound 2 elute earlier at 5.10min, and polar degradates are eluting after compound 1. However, one of deg seems to be overlapping with compound 1. Also, baseline starts to drift after 10min.



Next I tried to flatten the gradient (70%B to 40%B in 20min), hope to pull the peaks apart, but the separation seems to get worse. Not too familiar with HILIC, I am not sure how to further modify the conditions to improve separation.


The 2nd column I tried is Primsep 100 (5cm). I am a little skeptical since Primsep 100 has embeded COOH groups with pKa of 1, which will only enhance the retention of amine groups. The degradates I am trying to separate have 2 COOH groups and two amine groups, so I am not sure if they will be more retained or repelled. A trial run of 0%B to 50%B in 15min showed that compound 1 and 2 elute quite early.



With more 0%B hold, I can separate compound 1 from 2 as shown in the bottom trace. However, in the top trace of acid stress solution, you can see the degradates elute very early, one even before the solvent peak. Also because the compounds are eluting in the isocratic portion of the chromatography, peaks were quite broad. Since the degradates have one more COOH group than compound 1, using Primsep one might not be a good choice. But due to the pH constrains, I cannot use Primsep columns that have embedded basic groups.


Uwe Neue, I will order atlantis T3, but in the mean time I also tried Atlantis dC18 (since I have it), and get similar chromatogram as YMC ODS-AQ. Do you know if Atlantis HILIC will work here?


It's my first time posting pictures, so sorry about the mess.

I hope to try some more RPLC columns (love the sharp peaks), and also understand more about method optimization for HILIC columns. Ion pair still remians as an option as well.

Again, any suggestions are welcome.

Thanks, Molly

Einar, I mentioned twice now that the Risely, Pack paper is suspect to me, and that I have evidence that suggests that the ZIC-HILIC column shows typical, and thus strong, silanol interactions.
Thanks Uwe for the answer I was, sort of, seeking: Stationary zwitterrions as in the ZIC-HILIC would then also not interact with ions (ionically that is, polar interactions must be possible). That could mean Risely and Pack misinterpreted their data, which I suspect.
Vlad, if you can influence the behavior of zwitterions via pH, than only if you can remove the status of the zwitter via removing one of the ions as in amino acids, for instance, by protonating the carboxylate (weak ions). Now the Zic-Hilic has strong cations and anions, a pH effect would have to be via silanols or, of course via infuencing the analyte if possible.
Molly, I am not sure that I understand all of your data (would take more time....), but meanwhile I can only repeat: Do it with RP unless there is absolutely no retention.

Molly,

You are comparing 50 mm Primesep to 150 mm columns. You can try to use 150 mm, which will give you much sharper peaks and longer retentions Also I would try isocratic mode with ACN/water/phosphate buffer-60/40/10 mmol pH 6.8, 70/30/10 mmol pH 6., and ACN/water-60/40
Void on 50 mm column is 0.5 minutes at 1 ml/min.
Please look at th links I published above even without buffers zwitterions retained with K'=5-10. (ACN/water) (check newsletters)

Hi Molly,

Thank you very much for sharing your chromatograms with all of us. It's very helpful when trying to trouble shoot.

From looking at your data and knowing your experimental conditions / pH instability (and polarity) if your analytes, I think that sticking with an "aqueous-type" C18 column under neutral pH conditions is your best option.

My suggestion is a 2-prong approach - I suggest increasing your efficiency (increasing your column length) - while changing your selectivity by trying a different "aqueous-type" C18 column.

I would suggest our Unison UK-C18 column (250 x 4.6mm, 3um) column.
This is a nice 1-2 punch - it can retain the polar analytes, won't collapse under 100% aq conditions, and will significantly increase your efficiency (N) over a 150mm length column.

Here is an example of this column under aqueous neutral conditions
(please note the high column efficiency/low column back pressure):

http://www.imtakt.com/TecInfo/TI088E.pdf

Once the desired separation is achieved - we can then evaluate your
options to see if a reduced run time is possible - but separation must
first be achieved.

Feel free to contact me if you have any questions, thank you!

With more 0%B hold, I can separate compound 1 from 2 as shown in the bottom trace. However, in the top trace of acid stress solution, you can see the degradates elute very early, one even before the solvent peak. Also because the compounds are eluting in the isocratic portion of the chromatography, peaks were quite broad. Since the degradates have one more COOH group than compound 1, using Primsep one might not be a good choice. But due to the pH constrains, I cannot use Primsep columns that have embedded basic groups.

Peaks that are that broad, that early in an isocratic run always prompt me to ask how far the detector is from the column. If you have excess tubing & volume between the column and the detector, you can decrease your resolution and sensitivity as a result of diffusion. To a lesser extent, excess volume between the injector and the column can add to this.

What sort of LC are you running and do you need to trim some tubing lengths?

:: backs away from the column jousting in progress ::

Dr: I used Agilent 1100, and all the chromatograms were generated on the same instrument. I think the broad peak might due to the fact that I used 5cm column (from a method development kit). Also, is it possible that the ion exchange sites were overloaded and cause tailing (like the silanol interactions).

Yun

Molly, this is just beautiful with all this data and now you have all on the hook. To quote myself in a previous post “..I am sure that with all the support available here the method development step will be “piece of cakeâ€

Thorough work! The Atlantis dC18 gave more retention than the ODS-AQ. This is exactly as predicted. The Atlantis T3 will give only a marginal improvement, I think.

So now that you have increased retention, the question is if within the constraints of your compound stability you can resolve the early triplet (or even if you need to care about resolving it). Is it expected that small manipulations in pH will change the elution pattern of the early triplet?

Dear Molly,

Would you please specify column length and flow rate for each of your runs? it is hard to make conclusion (comparison between columns) without geometry of the column and flow rate.

Regards,

Vlad

Molly,

Any new results?