Chromatography Forum

I have an Agilent liquid autosampler connected to a liquid HPLC. I've tried different methods but I couldn't reach the RSD% of 0,5. Is there someone who can tell me which method I have to use to obtain this value?
Thanks
Daniela

Different manufacturer's use different methods, I would contact Agilent.

However, you should not have difficulties in obtaining RSDs of 0.5% unless the autosampler is an antique. Thoroughly check the system for leaks.

Is the 0.5% the riteria for full or partial loop injections? Often Partial loop values are more variable.

How do you test the variation? Weighing vials? Or calculating area variation?
In either case, how much do you inject?
Where does the 0.5 % RSD requirement come from?
Might be a good idea to post your data.

Best Regards

For Danko
I inject 20 microliter. The mode injection is full loop and I calculate the RSD from the pick area.
Do you usually weight the vials?
If I weight the vials I have a good RSD but I hope to find a good RSD also with the pick area.

The problem might be of chromatographic nature. I assume you calculate the variation between multiple injections of some kind of control-sample or a standard that actually have been run on a column and so on.
If that is the case, try to remove/shortcut the column and carry out the test again. You might need to dilute your sample solution if the peaks get too high and go off-scale. Also, set the sampling rate of the detector up to 3 – 5 points per second, or higher.
If you obtain better/lower variation, then you’ll need to review and maybe revise the method of analysis.

Best Regards

Is your method gradient or isocratic? And are your retention times regular? These aspects could affect your peak area, so you might not even have a autosampler problem...

if you say that you use a full loop, then your injector type is a pulled loop or similar in type autosampler.
there are various settings in such type of autosamplers that are critical in order to achieve decent RSD. they generally involve wasting a certain amount of sample in order to wash critical tubing.
and when i say decent with those autosampler types (dependend upon the chromatography as well), it will seldom go under 0.3% to begin with.

Reproducibility issues with the Agilent autosampler (if it truly is an autosampler issue) are usually associated with needle, needle seat, rotor seal, leaks, or metering valve seal.

And if you want to check out the autosampler, run the same standard and HPLC check out conditions that Agilent uses.

To obtain better RSD value what do you think I have to set up about this parameters?
-HPLC flow speed
- HPLC sampling frequence
- air gap on the autosampler;
- flush volume on the autosampler;
- injection speed;
- washes;


The pump is isocratic

Thanks

What model autosampler do you have? 1100? 1090? 1050?

Agilent sells a restriction tubing you can use instead of a column. That way you can isolate problems to the autosampler. In the past, we have used caffeine to test autosampler performance.

20uL seems like a high volume for filled loop injection. To be accurate, the autosampler would have to withdraw and inject 3-5x that volume, or near 100uL. Weighing the vial is useless, if filled loop is really what you are using, as amount withdrawn should NOT match the amount injected. However, if you are using an 1100, it seems likely to me that you are not doing filled loop.

It's 1100.

20ul injections are relatively large injections. If you're using 3mm or 2.1mm i.d. columns, that's likely way too large. Also, what are all your HPLC conditions, standard concentrations, etc? The reason that this is important is that maybe you're overloading your detector. Try again with 5.0ul injections or something, that's what I'd do first. If you really need 20ul to see your peaks, maybe you're close to quantitation limit. We can't help if you don't provide full information.

I use naphtalene 0,1 mg/ml with a SEA18 coloumn

At some wavelengths you may be saturating your detector. What is the absorbance at the top of the peak? If it is larger than 1000 mAU, then saturation may be a problem. Also check the integration, to make sure the start and stop locations are the same for all peaks. Change the integration settings if necessary.

Is the reproducibility better for peak height?