Mobile Phase pH choice

[Apple_insci]

I will work on a new drug candidate. The formulator told me the compound has 3 pKa:

pKa1=4 (measured); 5.6(calculated)
pKa2=8
pKa3=11

If this is the case, what mobile phase pH should I try first? I plan to use H2O w/0.1%TFA (pH is about 2)and ACN w/0.1%TFA as mobile phase A and B. Column I plan to try Xterra RP18. Is it a reasonable plan? What other mobile phase can I try?

Thanks!

Apple

[tom jupille]

Are the functional groups acidic or basic?

[Uwe Neue]

Large changes in pH are often good. What detector will you use, and do you need to work with MS compatible buffers?

[Apple_insci]

The functional group is basic I believe. PDA will be used as detector. MS compatabile is not required.

Thanks!

Apple

[tom jupille]

The question of pH can get people into table-pounding arguments. Here's my 2¢ worth:

I like to start with pH couple of units away from a pKa. That minimizes the potential for robustness problems that can arise when you are close to pKa, and ionization changes rapidly as a function of pH. In this case the possibilities are pH 2.x, pH 6, pH 9.5 (13 isn't practical). In a reversed-phase system, pH 9.5 with a base-stable column would be my first choice. Your molecule will still be charged (around +1.5 units), and you may have a problem getting retention unless the rest of the molecule is quite hydrophobic.

Assuming you can get reasonable retention, you can manipulate the selectivity by playing with mobile phase strength, organic solvent type, temperature, column type, and pH. I tend to prefer the first three because they are more amenable to "tweaking" (changing column type is essentially reshuffling the deck and dealing a new hand; moving the pH can make big changes in selectivity -- so big that robustness suffers).

If you can't get retention, then you will have to use either an ion exchange column, add ion-pair reagent to the mobile phase, or use one of the "mixed mode" columns that incorporate both ion-exchange and reversed-phase functionality. Once you can get retention, selectivity control is pretty much as in the previous paragraph.

TFA is a possibility, but if I read your reply right, your analyte will have essentially a +3 charge, so retention would be quite a problem. TFA itself is a weak ion-pair reagent, so that may help (HFBA would be another possibility in that regard).

[Apple_insci]

I tried borate buffer at pH 9.5 and 0.1%TFA H2O(pH 2). The pH 9.5 bufferhas a little better baseline. BUt it seems both bufer worked. However, I met a very headache problem. There is a small peak coeluting with my drug peak. retention time is about 0.3min different. I found this coeluting peak in the blank (acn/h2o 50/50) too, so I thought it might be the solvent impurity. I changed the slovent to MeOH, but the same thing happened. I washed the column and purged injector several times. But it didn't help.

I also tried to separate the coeluting peaks. But not lucky so far.

Any suggestions or thoughts?

Thanks for any input!

[Uwe Neue]

I am not sure if I understodd the complete story. Is it correct that your coeluting peak remained close to your drug peak independent of the pH and the change from methanol to acetonitrile? If this is the case, it might be a closely related isomer with otherwise identical properties as your parent compound.

[Apple_insci]

sorry for the bad explaination. You understood correctly for most part of my question. Beside that, in the blank injections of solvents the coeluting peak appears too no matter the solvent is ACN or MeOH. Is this carry over problem???

If the coeluting peak is an isomer of the drug, should they have similar UV spectrum?

Many thanks for your help!

Apple

[ntruong]

1. I would call it a carry over if the coeluting peak doesn't appear on the first injection of the sample, but observe on the subsequent injections. Extend the run time should take care of it.
2. I have a same problem with one of our LC released method (mobile phase:phosphate buffer and ACN). I have to subtract the blank all the time because there is a peak present in ACN that somehow coelutes with one of our impurities. I have tried to switch to different maker of ACN and still couldn't get rid of this peak.
3. Since you switch the organic solvent (ACN to MeOH) and still see the same peak, could it be a solvent front?

[Apple_insci]

The retention time of the drug in ACN is shorter than it is in MeOH. But the coeluting peak always elutes next to the drug.

[Apple_insci]

By runing the blanks and stds on the brand new column, I confirmed that the coeluting peak is the impurity peak in ACN.

[Uwe Neue]

So it appears that in acetonitrile, you get a little bit of a separation, but in methanol you get overlapping peaks between this impurity and your main peak. If I also understand you correctly, pH does have no impact on this separation.

If this is the case, then you have two options left. One is to continue to play with solvent selectivity, the other is to play with column selectivity. I would play with solvent selectivity first and use the best condition that you have until now, replace some of the acetonitrile with THF and see if things improve.

Another and maybe simpler possibility is to run the best acetonitrile separation at another temperature, say maybe 40 degrees instead of room temperature. However, for reasons of column stability, I prefer to use the former option.

[Apple_insci]

pH 9.5 buffer and ACN/THF 80/20 finally solve the problem

Thank you all for help!
Apple