Inconsistent RSD

jrich

<p>A method I'm working on is giving inconsistent injection repeatability. I've run it on 3 different days with 3 different results for RSD over 6 injections: 0.5%, 1.7% and 3.1%. I'm sure this inconsistency is from the method rather than instrument. The variance in area does not have a trend over the 6 injections, and I have not observed any carry-over issues. I've included some method parameters below. Any ideas? What components of the method are most likely to give rise to high RSD? Do you think the injector could have a hard time consistently drawing correct volumes of and IPA-based sample solvent? Thanks for any info! </p><p></p><p>Column: 2.1x100mm BEH Shield </p><p>Injection mode/volume: PLNO, 2 uL </p><p>Loop size: 10uL </p><p>Strong wash: 90% ACN (600uL) </p><p>Weak wash: 10% ACN (600uL) </p><p>Flow rate: 0.4 mL/min </p><p>Mobile phase Buffer component: 12 mM phosphate pH 6.0 with 0.2%TEA; Organic component: 50/50 MeOH/ACN. 18 min run time. </p><p>Analyte concentration: 0.4mg/mL </p><p>For the stability of my primary analyte, I'm using a sample solvent that is 75% IPA. </p>

dougm

If you injected a stable analyte such as a phenone or a phthalate or acenapthene (prepared in mobile phase to test your solvent theory) do you experience the same peak area variability? Have you tried another injection mode?

lizh

Hello,

I agree with Doug's suggestion, but also can you share your gradient table (or confirm it is isocratic if not) and please confirm what the organic is.

Are there any other SM settings enabled?

These are my first thoughts...

If the sample is soluble in IPA, then to eliminate any solubility issues I would have up to 75% IPA in the strong wash, to fully solubilize the sample and wash it away. In PLNO mode if the the SW is really strong that is OK and I would use a different selectivity for my weak wash, closely related to your initial conditions. If that works then I would reduce the amount of SW to abe a third of the WW to save time.

I will be rather bold and ask if this is a validated method and whether it was derived from an older HPLC method as I must say neither TEA and phosphate would be at the top of my list of things to use if they could be avoided.

Liz

jrich

Thanks Doug and Liz for your ideas and responses! Doug, your suggestions were helpful, and I'll give them a try.

Liz, to address your questions, my method is not isocratic but gradient. The organic is a pre-mixed 50:50 methanol/acetonitrile solution. No other SM settings are enabled.

I really appreciate your question about phosphate and TEA because it brings up another unresolved question. I was originally using acetate, but I have to detect at 227 nm, so, as expected, there were a lot of baseline artifacts late in the run. Besides those artifact peaks, with acetate I observed a periodic pattern in my baseline that interferes with accurate detection at trace levels. Because it is periodic, it looks pump or otherwise instrument related. I checked it on a different instrument (brand-new UPLC installed by Waters rep that week) but still observed that periodic baseline with acetate. However, when I used phosphate in my buffer, the baseline just looks like normal, random fluctuation. I attached a document showing the baseline with acetate and phosphate. That is why I'm using phosphate. Have you ever seen anything like this baseline issue before? As for TEA, I'm trying to separate a complex mix of degradants, and I observed that TEA selectively pulls one of them away from my critical pair in a concentration dependent manner.

With respect to strong and weak washes, I've been reading some of the discussions on this website about it. If I understand the posting by pcb (Dec 29, 2008 9:37 AM in response to: It (PLNO) allows a free hand when it comes to wash solvent selection. With PLNO you can go as strong as you need to prevent sample carry over. PLNO even allows the two wash solvents to be exactly the same and as strong as is necessary), with PLNO, the injected sample never actually comes in contact with the weak wash, right? So is benefit of having the WW similar to initial conditions related to how the needle is characterized (my understanding is that it is characterized using the weak wash)?

lizh

Hello again:

The acetate was showing I think mixing issues, it absorbs and again (I still do not have your gradient table) I do know the rate of change of your gradient against your flow rate, but a mixer could help, but you on phospahet now so best leave alone.

One of the chemists actually made a really good point IPA is viscous so you might simple reduce your syringe draw rate by at 50% and see it that helps.

Its is good to have your WW the same as the initial conditions, yes for characterization, but it should be the same as the gradient initial conditions which assume solubiloity for the sample. In PLNO it just keeps things simple.

Try speed first and then send the gradient

Liz

jrich

Hi Liz,

Reducing the syringe draw rate seems to have solved the problem. Thank you!

I've attached a document that shows my gradient and more details about the composition of my mobile phase system. As you can see, the gradient in the range of the baseline I showed is quite shallow. I'd be interested to hear any further information or ideas on this issue. Thanks.

manuel4146

Hi Liz

Can you clarify that point?

During the characterization

is the weak wash used?

Thanks Manuel

lizh

Absolutely, the weak wash is used to characterize the needle, and it should match the mobile phase and the gradient starting conditions.

However, as sometime happens rules need to get bent. I generally would avoid having a salt in the washes if I could (acid or base are fine) avoid it even if its in the MP.

Hope this helps

Liz

jrich

Thanks for your help!