Sample Loop Sizes

[stormsct]

Hello,

I am still relatively new to HPLCs, and I have a question about sample loop sizes. Specifically, what effect(s) will different sample loop sizes have?

The reason I'm asking is that I am using an assay that was developed using a 20 uL sample loop. The HPLC I'm currently using has a 100 uL sample loop and it's the only sample loop I have right now.

[Mark Tracy]

You did not say if this was a manual injector or an autosampler. In general, a larger injection volume gives a proportional increase in peak area, but may have some disadvantages. If the injection solvent is strong compared to the mobile phase, you will see significant distortion of the peak shape with increased volume. If the solvent is weak, you will not see as much distortion, but you may easily exceed the linear range of your detector, or the linear range of your calibration. You can even exceed the loading capacity of your column.

Generally, overfilling the loop 3-5 times gives the best analytical precision. Both autosamplers and manual injectors are able to use less than the full volume of the loop. If you must underfill the loop, keep the injection <50% of full. Autosamplers can do this with acceptable precision, but manual injection probably can't.

If you are doing manual injections, it is easy to make your own loop. 98mm of 0.020 inch i.d. tubing holds 20µL.

One other (usually minor) thing is the gradient delay is a few seconds longer with the larger loop. At 1 mL/min, this is negligible.

[stormsct]

Thank you, Mark. I really appreciate the information.

I am using an autosampler, more specifically an Alliance 2695. So, if I've read this correctly, there should not be a problem with the sample loop being larger. Is this correct, or am I reading this too simply?

[Kostas Petritis]

It will also depend on your column ID (i.e. flow rate) and how retained are your compounds. For 100-200 uL per min elatively unretained compounds can experience significant peak boarding...

Briefly, if your analytes are retained enough, you won't have any problems due to the stacking effect...

[Mark Tracy]

If you are doing standard applications with standard-sized columns, you have no worries.

If you need to do something at a smaller scale (<3 mm columns) you will need more than just a new sample loop to optimize your system. When the time comes, just ask the Forum.

[Victor]

As the volume of sample you inject increases, then you will get larger peak areas and signal/noise ratios. Thus large sample loops may be essential in some trace analytical applications.
However, to be balanced against this is the effect of volume overload. The volume of sample injected must be small in comparison with the volume of mobile which contains your peak; thus the volume injected is more critical for peaks of low retention compared with peaks of high retention, as Kostas has suggested. As an approximation, a completely filled 20ul loop is probably about the maximum volume I would want to use to avoid seeing too much of this effects. It is not too difficult to calculate the consequences of sample volume on column efficiency for peaks of various k, but it might be easier just to do the experiments and see how your separation is affected by sample volume.
The situation with an autosampler is different. Here the loop is partially filled. When the valve is switched the loop is usually backflushed on to the column, which reduces sample mixing in the loop during the injection process. Thus, 20ul injected from a partially filled 100ul loop might give you very similar results to 20ul injected from a totally filled 20ul loop. I presume you can inject whatever volume you want with your autosampler, partially filling the loop with say 5, 10, 20, 100ul. If you do this you will observe degradation of the efficiencies of the lowest retained peaks, even if you are using quite a large standard bore column (e.g. 4.6mm). The effects will be worse as the column diameter decreases.
If you are doing gradient analyis, then you can inject much larger volumes without any degradation of separation. This is because the sample can be concentrated on the column during the initial stages of the gradient.