Chromatography Forum

We are developing a method which uses Waters UPLC system. Initially carried out linearity and other tests and everything passes. However repeatability gives a high RSD (anywhere from 2.8% to 1.4% the lowest). We are trying to achieve lower than 1% but so far by changing gradient, speed of injection, the solvent the product is dissolved in. The injection volume is 0.2 (UPLC can go as low as 0.1%). Is there anything obvious that I am missing? What are the recommendations of what to try next?

Thank you,

Tanaz

What size loop are you using? To get good injector repro, you need to be doing injections between 20-80% of the loop volume. So, you should be using the 1 uL loop....

That's a good point. I will check that. I don't think it is a 1ul loop as other methods that run on there are higher injections.

Thanks,

Tanaz

The Acquity injector may work slightly different from what you are used to. Weak and strong washes play an absolute critical part in the injection process itself. Try to optimise these solvents as a part of your method development. These wash solvents should be fully compatible with your solutes in terms or you could get losses. As a generic solvent wash we have added TFA to both the strong and weak wash that reduces these losses for some compounds.

Also for the best possible reproducibility you could try a full loop injection but you will then be limited by what loop size you go with.

For Acquity we only use full loop injection for quantitative Work. The precision of the "partial loop needle overfill" option is not good enough.

I have also experienced that a validation failed for accuracy when using partial loop. The reason was that the actual injected volume was higher for the standards than for the samples. And the reason was that was that the samples had a slightly higher viscosity than the standards. This problem does not occur in full loop mode!

There are some good things about Acquity, but also some things where you wonder how they were thinking (injecting the needle wash with the sample is one of them).