Dear colleages,

Can anybody tell me why the signal of a compound with a retention time of 4.5 min show a shoulder on a side and sometimes a clearly split?
The sum of the area of the two splitted peak is equal to the total area of the unsplitted peak.

No shoulder or splitting was observed for a compound with a longer retention time (7.85 min)

Thanks for your help

We need more info to help you. What are the compounds, the mobile phase, the column? Is this gradient or isocratic.

Most likely you are injecting the sample with a solvent that is not compatible with the mobile phase - too strong, different pH, salt ....
Yes, more info would be better...

Please find enclosed more info :

a narrow bore column C18 2.1mm * 125 mm
mobile phase: KH2PO4 30mM pH=2.5 : ACN (52:48 %v/v)isocratic mode
Flow rate 0.2 ml/min
Compound:
Naproxen showing shoulder and splitting with retention time 4.5min
Flurbiprofen excellent peak (Assymetry factor 1.05) shape with retention time 7.8 min
Fluorescence detection with cell volume of 2µl
The chromatographic conditions are similar that those previously used with conventional column C18 RP 4.6mm * 125 mm and a flow rate of 1ml/min

Special considerations concerning the inner diameter of the tubing were taken into account to minimize peak broadening

I presume a void volume of the column! But how to check it?

Thanks for your help

What is the solvent the sample is dissolved in, and how much (volume) is injected?
If it were a void, both peaks would be lousy.

If you're sample is made up in mobile phase, it's possible that you could have a void and have normal peak shapes for a later eluting peak while having split peaks for an earlier peak. To check this, make up the sample in a buffer solution containing 10% less acetonitrile than the eluent. If the splitting problem goes away for Naproxen under these conditions, it's a good bet you have at least a minor void.

Sorry, I forgot in my previous mail to specify the composition of the sample:

the analyte was dissolved in isotonic Phosphate buffer solution at pH=7.4

and 10 µl were injected onto the column

Chris Pohl,

That is interesting! Why is it that the splitting problem can be overcome by injecting a sample solution containing 10% less ACN?

Mathy FX,

What is the concentration you are injecting (both the compounds the same conc.?)? Is there a possible overload occurring? I reckon these acids not to be much dissociated at pH 2.5 having an aqueous pKa of about 4.3 and 4.7. However, the pH of your mobile phase and the pKa of your acids most likely increase due to the organic modifier. This could lead to a partial protonation of your compounds if the pH shift of the mobile phase buffer is more pronounced than the pKa shift of your sample compounds. Charged compounds can overload much more readily than do neutrals.

Also, your sample solution has got a pH of 7.4 (acids mostly deprotonated) whereas conditions in your mobile phase are totally different. The later eluting flubi gets more diluted (system internal) than does naproxen, which could compensate for this for the former compound relatively more than for the latter?!
Thus, indeed as mentioned above by the others, try analysis with samples solved in mobile phase.

What application did you column find prior the present use?
How old is your compound? - Is there a possibility of degradation for Naproxen, e.g. due to acidic hydrolysis and probably cleavage of the methoxy-group?


Regards,

Stephan

To elaborate on anon. and Chris: This looks like a classical pH inconsistency. The Naproxen is apparently partially "washed" into the column until it is no longer influenced by the pH=7.4, while the Flurbiprofen is retained at both pH. It looks like you got lucky to have only splitting, etc., and that in only one of the peaks. There is an easy way out: Remove the solvent with a centrifuge ultra-filter, add mobile phase and inject.

Stephan,

Of course it won't always work because the problem might be due to a pH problem as suggested by Mueller above, but typically if the sample composition is "focusing" (i.e. analyte concentrates from the sample band onto the head of the column because the sample solvent is a weak eluent relative to the mobile phase), the effects of a column inlet void will be minimized if not completely eliminated. This works best if the void is a uniform recess rather than angled toward one side of the column. Anyway, the thought was that since the effect was retention time related, it might well be due to this effect since if the sample solvent is identical to that of the mobile phase this focusing effect will be more pronounced for late eluting solutes. This is the reason that I almost never make up samples in mobile phase. Often, I hear that this is necessary in order to get the analyte into solution but typically one can arrive at a sample solution which dissolves the analyte using a higher solvent content and then dilute the solution to the appropriate analyte concentration in a solvent containing matrix with the solvent content significantly below that of the mobile phase. The same applies to ion exchange although in this case the simplest solution is to dissolve the samples in deionized water. Sometimes it's necessary to adjust the pH before hand but it's almost never necessary to make up the analyte in eluent.

Thank you very much for your interesting answer and discussion.

For my application dealing with on-line microdialysis sampling I cannot dissolve my standard solution in mobile phase. Indeed the perfusate medium of the microdialysis probe is the PBS.

Atfer reversing the column, the peak shape of the naproxen is perfect with a assymetry factor of 1.02 and the Area is equal to the sum of the splitted peaks which were previously observed when the column was connected in the other direction

the concentration range injected onto the column is 1 to 1500 ng/ml

so at the view of these last evidences, I'm afraid that my column (new and free for trial) presents a void volume

Chris,

Very interesting - thanks. Only one more :) - what about buffered mobile phases for example working with weakly basic compounds at low pH?

My concern here would be the change of pH with organic modifier (%B) compared with the aqueous pH. Thus, the sample solution might have a significantly lower pH (due to less %B) than the actual mobile phase has, assuming an acid buffer (e.g. phosphate). If pH of mobile phase is close to pKa of the base couldn't this have a detrimental effect on peak shape, according to above (e.g. Mueller)? Have you got any experience whether this can affect peak shape 'a lot'?

Stephan

Stephan,

You raise a good point which may well be the case in this instance. In essence, you are always better off if the sample solvent would result in an increase in retention if it were used as an eluent. In reversed phase, this typically means reducing the amount of solvent and in ion exchange this typically means reducing the ionic strength of the sample solution but certainly adjusting the pH of the sample solution would fall into this category. For example, the analyte in question will be more retained at low pH since the protonated form of a carboxylic acid will always be more hydrophobic than the salt. It would follow that, all other things being equal, focusing could also be accomplished by preparing the sample at a lower pH than the mobile phase (at least in this case, although the right direction to change the pH would be dependent upon whether the analyte was an acid or base and things get even more complicated for compounds with multiple ionizable groups).

Mathy FX,
Though your result with the reversed column indeed looks like a void problem, it may not be prudend to give up on the pH possibility as yet. Since you are probably injecting small volumes a slight increase in mixing efficiency due to the other hook-up could eliminate a pH effect. What happens if you inject much larger amounts of standard in PBS?

One other quick note, I have seen the effect mentioned above when the inlet frit is partially blocked. The injections sometimes look like a column void. Just turning the column end-for-end removes the inlet block and "restore" the performance. You might try replacing the inlet frit if possible and see if this restores column performance.

Regards,
Mark

Thanks Chris,

Stephan

What column are you using?

To minimize pH effects, you could collect into perchloric acid which is very common in microdialysis.

Also, please try injecting 1, 2 & 5 microliters at the same time as the 10 uL injection to test for column overload.

What are the possible causes of shoulders in affinity chromatography