Dear José:

Might you use a higher pH or a longer column?

Remember that positively charged compounds usually elute very fast from analytical columns. Principles of HILIC are similar to those of HPLC.

Best,

Thank you, Aurea

I would prefer not to change the column.

Increase pH was one of the solutions I was considering, also because I think riboflavin is highly dissociated, since I found two peaks, one of them at m/z = 377 (M + H)+ and other (elutes before) at m/z = 359 ([M – OH-]+?) Or maybe I have an old standard…

Best regards,

Hi Jose:

Here is a very straight forward application with several vitamins:

http://www.kromasil.com/applications/view.php?app=g1024

Best regards:Cecilia

i had the same experience with HILIC once, i would suggest exchange some of the aqueous part of mobile phase with alcohols , hence try first 90 ACN 5 methanol 5 water (with formic acid 0.1%) , or other alcohols and/or other ratios ..

HILIC separations are not only dependent of pH. Do not underestimate the effect of ionic strength in your mobile phase.

Try lowering your Ammonium Formate concentration to 5 or 2 mM and observe if it has an effect on the retention times. If it has, another approach could be using a gradient with different AmmForm concentrations in mobile phases A and B.

I do not know your column exactly but usually manufacturers of HILIC columns advise against the use of alcohols in mobile phases.

Hi Jose

If I read your question correctly, your mobile phase may need adjusting. I understand your 'Bottle A' is pure acetonitrile and your 'Bottle B' is Ammoniun formate 10 mM, pH 3.2. 

Joachim has a point that ionic strength is important in HILIC, especially for ionogenic solutes. Low ionic strength buffers are associated with poor peak shape and mass overload, also enhanced retention for charged bases (see here for a paper on the benefits of salt buffers in HILIC http://www.sciencedirect.com/science/article/pii/S0021967314005780).

If your method is running at 95% ACN, the concentration of ammonium formate in the mobile phase will be 1/20 of that in bottle B, i.e. 0.5mM overall once mixed online. This is low, and could be contributing to the broad peaks you have observed. An overall salt buffer concentration of 5mM is quite effective in HILIC; perhaps using an overall buffer concentration 10 times higher can help.

Contrary to what Aurea has said, positively-charged compounds retain well on bare silica columns in HILIC due to ion-exchange with silanols. This effect can happen on bonded phases too (http://www.sciencedirect.com/science/article/pii/S002196731201895X).

However, salt buffers can precipitate above around 10mM overall concentration in HILIC mobile phases, so making your 'Bottle B' 10 times more concentrated may result in buffer precipitation when mixed with neat acetonitrile. This can be avoided by pre-mixing your salt buffer with acetonitrile, which is quite possible at 95% ACN with 5mM ammonium formate pH 3.

Perhaps making 'Bottle A' as 70% ACN with 5mM AF pH 3 and 'Bottle B' 95% ACN with 5mM AF pH 3 can help. You can then develop the method and find a blend of A to B that gives the desired separation.

If you're using a diol-type column, this may be an issue. They're known to be less retentive, and I agree with Aurea that changing the column can help (see the Kumar reference for more information).

Good luck!

Article Practical investigation of the factors that affect the selec...

Article Comparison of peak shape in hydrophilic interaction chromato...

you can using a gradient with AmmForm in mobile phases A and B. increase organic percent is good for separation of rip with sharp peak 

Have you considered just using a different column, e.g. C18 column? I tried HILIC for multiple B-vitamins but I found the Waters HSS column (50x2.1, 1.8u) worked much better for me in terms of peak shape and retention time. An alternative for HPLC rather than UHPLC would be the Waters Atlantis or the Agilent Poroshell C18.

In regards of your separation issue of riboflavin and pyridoxine, since you are using MS/MS as detection I am assuming you are running in MRM mode. Thus, each vitamin will be detected by its own trace. Parent and product ions for both vitamins in questions are very different in their MRM (e.g., R 377 > 243; PN 170 > 134) so your separation by time is not that crucial since your detection is based on MRM, not on retention time.

Good luck!

I thinink the Phenomenex: Kinetex 2.6 HILIC columns is unmodified silica, which is not bearing any of organic groups. Hence, the separation  selectivity only depets on the ınteraction bewtween silanol group of silica and functional groups of the analytes. ın order to piridoxale, pyridoxine and piridoxamine; piridoxale is an aldeyde,, pyridoxine is hydoxy and piridoxamine is an amine functional group. piridoxamine is more basic and can make strong interaction with silanol group. But in the strong asidic media, it can evaluate faster then pyridoxine. On the other hand, piridoxale is an aldeyde and les polar then other two; therfore it must evalute as first in ACN rich mobile phase. You can separate easyly by trying to work at other pH from 3 to 8.5. BUT the best option is that to change your column.You can use HILIC LUNA, Obelisc N or TSK-Amide 80 hilic columns. At the moment My group is working on HILIC separation of water soluble 8 vitamins by our new  column. I have seen clearly that the separation of water-soluble vitamins are strongly affected by structure of stationary phase and change pH of mobile phase.