Injection of aqueous samples in normal phase HPLC?

[paroma]

Do you know if it’s possible to inject a water solved sample in a HPLC system working in normal phase without sample treatment. This problem arises because I'm using a very well validated normal phase method for separation of impurities of an active substance and I wish to use it also for its eye-drops formulation (obviously in water as solvent).
Thanks

[danko]

It depends, very much, on the injection volume as well as the column volume.
If you inject 3 – 5 μL aqueous sample solution onto a 4.6 x 150 mm column or something like that, you have good chances for fine chromatography.

Best Regards

[Noser222]

and if that doesn't work, dilute with acetonitrile

[paroma]

Thanks very much, until now I didn’t try because I was afraid to damage the colum or the system. By your experience, what can happen if it don’t works?

[Noser222]

What exactly is your solvent system and column?
If it doesn't work, you'll probably just get distorted peaks.

[sassman]

Depends on the column. A bare silica column will take a very long time to re-equilibrate after you put a significant amount of water through it. Kind of like putting hexane into a C18 column. An amino column will be much less affected.

[paroma]

The validated method that I used for API impurities is: Silica column (LiChrospher Si 100 5μm), mobile phase Hexane: Ethanol 96:4 (isocratic elution), 2.0 ml/min, loop 50 μl (I used 20 μl), 25ºC, 210 nm.
I also know details of another validated method that make use of an -NH2 column. Perhaps it would be more suitable for injection of aqueous samples?
In reverse-phase, if one injects a sample solved in pure organic solvent in most cases takes place a band-broadening. Maybe it's the same with water as solvent in normal phase?
Thanks.

[Alex Buske]

I wouldn't inject water in your solvent system because:
-water is the strongest solvent in that context, in the best case broad peaks are expected
-water is inmiscible with hexane - Iam not sure about your mixture
-water could directly layer on the silica gel, changing the column properties.
On the other hand SPE with C18 would be possible, load the cartridges with the aqueous solution, dry and elute with your solvent mixture.

alex

[Kwet]

I agree Alex Buske,

I think you should absolutely avoid even trace of water on your column because water adsorbed will modifie the column propreties and give unpredicted retention time.

SPE would be a solution if you can recover all the impurities you want to analyse.

If using an evaporator is allowed (no volatile impurities to analyse) you can extract your aqueous sample with acetonitrile. If you reach 90% acetonitrile in your sample you can evaporate easily the water with acetonitrile to dryness. Then you can dissolve dry residue in your mobile phase.

To avoid evaporation, perhaps a small amount of your aqueous sample (10 - 100 mg for exemple) can be extracted with your mobile phase (hexane / ethanol 96:4). I know water will not be soluble in hexane / ethanol but your formulation perhaps…and if not increase the volume of exctraction solvent. Filtre on sulfate to remove water and inject.

[paroma]

Thanks very much to Buske and Kwet, I think you’re right.
About SPE C18, it’s perhaps the best and quicker choise in this case, but I don’t know how to validate the recovery of impurities although I think it would be very high because my sample is very hydrophobic.
About liquid/liquid extraction procedures I fear that the presence as preservative of benzalkonium chloride that is a strong surfactant could drastically reduce the recovery.

[Kwet]

If you have analytical standards of your impurities, you can validate the recovery by adding a known amount of impurities at the beginning of the method (standard addition method).

[danko]

my sample is very hydrophobic

If your analyte is very hydrophobic, normal phase was not the right technique in the first place

Best Regards

[sassman]

If your analyte is very hydrophobic, normal phase was not the right technique in the first place

Please explain why you think this is so. Most often, the choice between normal phase and reverse phase is dictated by the sample matrix. For example, a few weeks ago someone was analyzing impurities in petrolatum. Normal phase was an obvious choice since he just had to dissolve in mobile phase and inject. The sample prep for a reverse phase analysis would have been difficult in that case. It is, in fact, difficult to think of an analyte (other than aliphatic hydrocarbons) that will not be significantly retained on a normal phase column.

[danko]

Hi Sassman,

Please explain why you think this is so

Are you kidding me? If not, there you go: A very hydrophobic compound would clearly “prefer to stayâ€

[paroma]

Thanks Danko and Sassman, yours considerations are right in general, but perhaps I have to give you more information about my specific problem. The molecule I must analyse has a prostaglandine structure and it’s absolutely insoluble in water. In eye drops formulation, solubility is achieved by the surfactant action of benzalkonium chloride. I have to get the chromatografic separation of potential impurities that are isomers of API. I’d be very happy if I could use a reverse phase method and I tried in many different conditions, but in separation of isomers RP is much less selective than normal phase. Also using very efficient columns and very long analysis time, I never succeeded in resolving all of them, where normal phase gives always a very good separation, with good peak shape in relatively short time. Moreover the normal phase method is already validated for identification and quantitation of impurities in the raw material and maybe I cold have good results also in the formulated product if I could inject it.