This is more a combination question... of
both method development and microbore systems:

I'm trying to optimize a system to quantitate a
small polypeptide within a formulation; to
increase sensitivity & resolution.
The sensitivity is definitely an issue since
the analyte is at a 1.1 ug/ml concentration
within the formulation. And impurities of this
analyte need to be quantitated as well.
I've gotten to this point already... I just
want to increase resolution of the analyte with
a small peak on it's descending tail.

I've converted a standard hp 1100 system to a
microbore system, using a vydac 218MS, 1 x 250 mm, 5um column.
my gradient is 15% to 30% B over 70 minutes.
all the literature points to a shallow gradient
being best for small polypeptides.
A --> 0.15% TFA in H20
B --> 0.10% TFA in MeCN
flow rate=75 ul/min;
inj.vol=30-40ul, perhaps this needs to be reduced;
VWD, 215nm

What my question really is: what is the way
in this type of situation to resolve closely
eluting peaks?

A) I've tried a 55 min. gradient, 15-27%B, a
reduction of 10% organic at end. This seems
to give ok resolution.

B) increasing the ion pairing agent (TFA) wouldn't
help, would it?

C) In general, how much past the critical organic modifier concentration should you go when
establishing end times for gradients?

D) a .22%/min gradient had better resolution of the two closely eluting peaks than did .18%/min.
Is this generally the case... to not go too shallow?

----------------------------

If anyone has thoughts/comments/etc... I would
appreciate it.

Whew! you have raised a number of questions here.

The first question is "do you need sensitivity in *mass* terms or in *concentration* terms? In other words, do you have a strictly limited mass of sample (one tablet, for example), or do you have lots of sample, the problem is that it's diluted.

If you are in the latter situation, then, IMHO, the microbore approach may not be the best way to go. My temptation would be to stay with the "conventional" column and simple increase the injection volume (assumin, of course, that the sample is dissolved in something much weaker than your starting eluant.

If you are mass limited, then the microbore approach can help.

OK, now on to resolution. It may be helpful to go back to basics here. Resolution depends on the center-to-center separation between the two peaks in question. This means that you have to move them apart while keeping them narrow. Let's take each in turn:

Moving the peaks apart. As a first approximation, resolution in a gradient system depends on average retention (k*) in much the same way as does an isocratic system: Rs is proportional to k*/(1+k*) -- assuming that selectivity stays the same. This is why shallower gradients, *on average* give better resolution: the give a higher k*.

In practice, however, selectivity can change when k* is changed (just as selectivity can change when isocratic k is changed). Thus, it is quite possible for resolution for a particular pair of peaks to be better for steeper gradients. The only way to be sure is to do the experiments.

At the risk of sounding too commercial, we (LC Resources) make and sell a computer program called DryLab that lets you plug in results from two experiments and model the effects of changes in gradient time, flow, column dimensions, etc. This can save *much* time with this sort of problem.

Ok, now the peak width issue. Microbore columns are very, very, very, . . . sensitive to extra-column effects. If you check out the thread titles "extra-column effects" here on the Forum, you will find an Excel spreadsheet I posted in which you can enter various assumptions about column dimensions, resolution, etc. and see what kind of extra-column effects can be tolerated.

The "bottom line" is that, unless your samples are dissolved in a very weak solvent, 30-40 microliters is far too large for a 1 mm ID column.

If you want to e-mail me privately, I can take a look at your data and possibly make some more relevant suggestions.

Hope this helps!

-- Tom Jupille / LC Resources Inc.

Raj:

Microbore systems-in the right hands-can yield very good results in terms of enhanced separation and sensitivity. The problems you face as i can see it, are due to your method as well as the system itself.

I can offer the following suggestions:

1) As Tom mentioned above, microbore systems are very sensitive to dead-volume effects, particularly where it matters the most - between column and detector. The length of tubing here should be very,very short indeed. Not more than 5 cm., as far as possible. Also, the i.d. of this tubing, as also the tubing between injector and column should be the barest minimum - 0.009 inch or thereabouts. This means the system backpressure will go up, but one has to live with this.

2) I would also suggest you try reducing the flow rate. This usually results in an increase in the theoretical plate count and thereby better resolution. Not always though,( remember van deempter's curve!) so you need to experiment with different flow rates. If you have an automated system, it should be easy enough to set up, especially since your run-time is around 70 min.

3) Since you are using a converted HP system, please take a look at the mixing chamber. The mixing chamber volume usually has a direct bearing on system performance. Larger volumes here will have a negative impact on resolution.

4) I agree with Tom - 30 ul is way too much for a microbore system. It is too high for a conventional system too. It means that your sample loop is that much larger, and band spreading that much more. Are you using an autosampler with a 100 ul loop or something like that ?

A good microbore separation demands an injection volume of not more than 5 ul ( in my opinion). If your sample is too dilute to allow smaller injection volume, then you need to concentrate it using SPE or-since it is a polypeptide- perhaps an ion-exchange or ligand-exchange membrane filter. I would suggest you check out the Sartobind range from sartorius or similar products.

Hope you find this useful. If you need more help, do drop me a line at stimulus_99@yahoo.com or hplc_99@hotmail.com

warm regards,

s.k.srinivas.
stimulus/bangalore/india.