Ion pair chromatogarphy

[Rowena]

Hi,
I'm developing an impurity method for bisphosphonate using ion pair chromatography. The ion pair reagent used was
Tetrabutylammonium hydrogen sulfate , pH = 7.0, 7.2 and 7.4 using K2HPO4 (900ml: 100ml MeOH)
Flow rate 1ml/min, Temp = 30
detector= 215 nm
I'm not able to achieve sensitivity at 0.05ug/ml
Please suggest
Also tried the tetrahexylammonium hydrogen sulfate at pH=7.9.
Thanks.
Rowena

[tom jupille]

What injection volume?
What is the concentration of the main peak?
What is your sample dissolved in?
What is your actual LOQ?
How much baseline noise do you have?

What I'm driving at with the first three questions is that detectability is more closely related to mass-on-column than it is to concentration. If your diluent is weak and you don't have a huge main peak to bother with, simply increasing the injection volume can improve your sensitivity (in terms of concentration; your mass sensitivity remains the same).

The last two questions get at the signal/noise issue. LOD and LOQ are fundamentally limited by S/N. You can have very good detectability with a very small peak IF your baseline noise is low. The manufacturer's performance spec for your detector should give you an idea of how much noise to expect (take it with caution, however; the spec is usually for a longer wavelength and under no-flow conditions); you should be within a factor of 5-10 of that value. If the noise is worse than that, you may have a lamp problem, a dirty cell, erratic flow, etc.

[Rowena]

Hi Tom,
The sample is a FP solution of 0.05mg/ml, so I'm limited by the sample concentration. The sample is injected neat with placebo in the formulation. I'm trying to achieve sensitivity peak for the active bisphosphonate peak in presence of placebo (that's what will be needed for the recovery experiment). The LOQ required for this dose is 0.1% which is 0.05microgram/ml. The sensitivity requirement at this level is S/N ratio > 10. The detection wavelength is 215 nm (maximum is 210nm, but I'm trying to work little away from the 210 because of baseline noise).
Initially I tried with injecting 100ul which gave me good separtion from the known impurity and the excipient peaks with good peak symmetry using the tetrabutylammonium IP reagent. Also tried the tetrahexylammonium IP reagent to have a better retention. Both the IP reagent mobile phases had EDTA added to eliminate the interaction of the bisphosphonate with the metallic component of the HPLC system But since no sensitivity was achieved, I tried increasing the injection volume but the peak shape of the active started to get distorted especially for the stressed samples of acid and base.

[tom jupille]

OK, 100 microliters is as far as you can push the injection volume. That translates to about 5 nanograms for your LOQ (0.05 micrograms/mL = 0.05 ng/microliter). That's getting down there for UV detection.

So, the next question is: "what is your LOQ right now?" (i.e., how much improvement do you need?). Then you have look at how to get that improvement ( more signal!? Less noise?). That leads you to issues of extinction coefficient and baseline noise at 215 vs. 210 nm, how wide is your peak and can it be sharpened,etc.

[Rowena]

Hi,
I have achived sensitivity (S/N of about 7) with 150 ul injection volume. I'll have to try how to improve the peak shape (sharper peak). Can IP reagent conc change/pH help me?
Will flow rate have an improved peak shape in this case?
Please suggest.
Thanks

[tom jupille]

Until I have some idea of what your chromatography actually looks like, it's hard to make any specific suggestions.
What is your LOQ (S/N = 10) with a 150 microliter injection?
What is the retention of your peak (k' value)?
What is the noise level of your baseline (in AU or mAU)?
What is the baseline width of your peak? (or, if you prefer, how many plates are you getting)?

[carls]

Is the EDTA necessary? It likely increases the noise significantly at 215nm.

An experiment I have been wanting to try with bisphosphonates is adding a small amount of Ca++ to the mobile phase. If you can use a lower pH, Ca phosphate monobasic Ca(H2PO4-)2 has reasonable solubility. I suspect a low concentration such as 0.1-1mM should be enough and you could use 0.1 v/v% conc H3PO4 to control the pH to ~2.1. You might also experment with another pH, such as pH 3, to see where you get optimum retention and peak shape. You'll still need an ion pairing agent, 10mM is usually sufficient.

Not sure if the Ca++ will help but it may be worth a try.

[Rowena]

Hi Carls,
The finished product sample strength is as such low (0.05mg/ml), so its better to inject the neat solution without any dilution. The pKa of this molecule is 5.9, hence I'm trying to work at pH=7.9. Also, being a bisphosphonate it is neccessary to eliminate any metallic interactions from the C18 phase as well as the HPLC component. I'm trying to achieve sensitivity @ 0.1% of this dose.
Any ideas.

Rowena

[carls]

In my post I was suggesting adding Ca++ to the mobile phase, not the sample solution. The idea is that the Ca++ will bind (chelate) with the analyte and thus reduce interaction with any metal components and hopefully improve peak shape.

Is the pKa of 5.9 the second pKa or a pI? Typically the first pKa is close to 2. Bisphosphonates typically have at least 4 pKa's, 2 for each phosphate group.

Removing the EDTA should reduce baseline noise and also allow you to use 210nm

[carls]

Rowena,

Any progress or updates?