impure peaks when doing fraction collection

[xiuminchen]

Hi there,
I have a problem when I do the fraction collection. I use semi-preparative HPLC column to purify my sample, but I always get another unwanted peak in my fraction. I use Agilent 1100 series HPLC. The delay volume for the fraction collector is 74ul. These two peaks are quite close with Rt at 11.961 and 12.332 (unexpected peak). The screen always shows I only collect one peak, but when I re-run the collected fraction it shows two peaks and the ratio of these two peaks in the collected fraction is much lower than original samples, which means I lost part of my interested compound during fraction collection.So, can anyone suggest how to improve it?

thanks
xiumin

[tom jupille]

Without knowing things like your flow rate, which peak is the one of interest, what volume you are collecting, and how overloaded you are (how wide are the peaks, and how bad is the tailing), it's hard to make any specific suggestion.

As a general comment, collect more (smaller) fractions and then pool as necessary.

[skunked_once]

Tom's suggestion is useful. You can also skip the fraction collector and try to manually collect the peak, assuming that you can monitor the real-time absorbance rise and fall of the peak of interest. Don't rely on the screen chromatogram since there may be a delay due to signal processing, depending on your chromatography system.

Do you know that your peak of interest is stable? I tried collecting some pure peaks of a natural product and my pure fractions showed 2-4 peaks after sitting in methanol/water for about an hour.

[xiuminchen]

Thanks for all the suggestion. The flow rate is 1.6ml/min, column Agilent Zobax SB-C18 9.4*250mm. I am interested in the first peak.I collected from 11.72 to 12.00 min. In the original sample these two peaks retention time are 11.904 and 12.277 respectively; peak width are 0.1204 and 0.1156 respectively.

[tom jupille]

11.904 and 12.277 respectively; peak width are 0.1204 and 0.1156 respectively.

Is this an isocratic separation or a gradient?

[lmh]

It might be worth checking that the measured delay volume really is 74uL. As I remember, the delay volume is specified somewhere in the software, but the software cannot know if someone has subsequently replaced bits of tubing with different lengths or diameters without updating the delay volume!

Also you will always get a worse result at the fraction collector than you do at the detector, because the fraction collector is down-stream, and peaks get wider with each stage they pass through from the column onwards. This means that two peaks that are just baseline resolved in the measured chromatogram probably aren't baseline-resolved as they leave the tubing in the fraction collector. Similarly, if you set the window to collect a peak so that the peak is just contained within the window at the detector, although the 74uL bit of the software will move the middle of the window so that it centers exactly on the peak as it leaves the fraction-collector, the peak will now be wider, and its front and tail ends will be cut off, giving slightly reduced yield.

You should use as short, and as narrow a tube as possible between detector and fraction-collector, and make sure the volume is calibrated as accurately as possible.

[paul_b]

Even easier - use a dye mix like the one from Waters (part number 716000765 - expensive though at £116) and then visually check the fraction collection, you may be surprised by an unexpected dead volume. Contamination of a second eluting component is quite common at higher loadings but the first eluting peak should be pure with generally high recovery >95%.