calibrtaion of autosampler loop

[ziyads]

how to calibrate the volume of autoinjector loop
[/b]

[k66]

what for?

[tom jupille]

In HPLC, sample volume accuracy is much less important that precision. Because the calibration and sample runs are done on the same system, it doesn't matter (within reason!) how much you inject, so long as you inject the same amount each time.

[danko]

Besides, it’s not the loop that determines the accuracy or the precision of the injected volume.

Regards

[k66]

Because the calibration and sample runs are done on the same system, it doesn't matter (within reason!) how much you inject, so long as you inject the same amount each time.

Exactly. Systematic error.

[HW Mueller]

Overfilling the loop at each injection usually gives the best precision. If one still has a systematic error it most likely has its source elsewhere.

[Mark Tracy]

If you want to calibrate the "loop" that is easy: remove it from the system, remove all the liquid from the loop, weigh it, fill it with water, weigh it again. If you want to calibrate the HPLC system, that is a different matter.

[DR]

^ agreed - but, as alluded to, the LC injector has some volume above and beyond that of the sample loop (channels in the valve, tubing between the valve and the needle - depending on the injector design etc.).

A semi-good way to look at accuracy is to do a linear regression using a series of injections of different volumes of the same solution, a single component standard of something that has a large extinction coefficient and good peak shape. If you compare your regression results for "forced through zero" to "not forced through zero", you will get an idea of the magnitude of your injector's fixed error.

If you need more than "an idea", you'll have to consult someone who is more comfortable with weighting and statistics than I am.

Personally, I wouldn't do it. I was once told that the internal volume of a certain series of manual injector rotor/seal combinations could (and frequently did) vary by ~17%. If you're really expecting accuracy in the sub 5µL volume area for an automated injector, you may well be disappointed.

[Mark Tracy]

The standard tolerances on steel and PEEK capillaries is such that 10% RSD of the volume is not unusual. The manufacturing spec will be 2 times larger.

[HW Mueller]

Since the volume includes plumbing, as DR pointed out, I see no alternative to strongly overfilling the loop (and thus the rest of the plumbing) with a well characterized standard for a "calibrating the loop" sequence. This is the most consistent injection method I have experienced. (It has been discussed before, no or low overfill can give problems due to laminar flow, syringe variations).
Of course, this assumes that there is no syphoning.
To eliminate injection problems I use a combination of external and internal standard for quantitative work, +, of course, overfill.