Sample solvent intolerance of injections

[Hollow]

Hi

we observed the following phenomena:

We adopted a method from a project partner, and there the reference was dissolved in 100%MeOH, while the inital conditions of the gradient were 15%ACN/acidified water; column is a Nucleodur C18, 5µm, 4.6x250mm; InjVol was 20µL.

We first run the method on our Waters Alliance 2695 and on the 600-system with 717 injector (both low pressure gradient), and got "normal" chromatograms with good peaks.

Now we did a transfer to another lab which is using older Varian (low pressure) and new Agilent 1200 RRLC (high pressure gradient) systems and observed very bad peak distortion when injecting >10 µl of the reference solution on both systems.

Diluting the reference solution to 50%MeOH solved the problem but I wonder why we don't get this phenomena on the Waters HPLCs? (maybe they are simply too good

Does anybody have an idea why the different systems are behaving like this?

[tom jupille]

(maybe they are simply too good Wink )

Actually, it's more likely to be the opposite.

Peak distortion from a "too strong" diluent is a common phenomenon. It's the chromatographic equivalent of a skid mark as a fast-moving peak in a strong injection solvent comes to a screeching halt as it mixes with the mobile phase.

I suspect that the Waters system has a fair amount of mixing volume between the injector and the head of the column, which means that by the time your injection reaches the column, it's been diluted down to a lower MeOH content.

[Hollow]

I actually thought about this, but up to now I just keeped the gradient delay volume in mind which are similar on the RRLC and the Alliance.
Now I realize that I have to evaluate the "injection delay" volume of the systems.

[tom jupille]

Actually, it would be far better to avoid using a diluent that is stronger than your mobile phase. Or, if you absolutely must dissolve in a strong solvent, keep the injection volume to a minimum.

I just finished a teaching "Practical Troubleshooting" at a pharma company here in the US in which we talked about exactly this topic. I've seen one case where the separation looked fine until the instrument was serviced, after which the distorted peak shape problem came up. As near as we could tell, some of the connecting tubing was straightened out, which resulted in different mixing characteristics (coiled or kinked tubing sets up a "racetrack effect" which increases radial mixing).

[Consumer Products Guy]

We see this routinely, even a 28-year veteran here experienced this a month ago with a standard made up in methanol and samples in 50/50 water/methanol. I showed him that decreasing the injection volume actually increased the sensitivity due to better peak shape, and he changed his standard preparation to match up. Some think "more injection" will get them better sensitivity, not always true. Another analyst here experienced wandering retention times recently; again, my "second pair of eyes" immediately saw that the pressure was fluctuating too much, and was due to the vacuum degasser not being turned on. In that case ALL lab instruemntation had been turned off during an electrical repair, and we must've turned off a thousand power switches, and that one was missed upon powering up.

Never underestimate the power of asking a co-worker to take a quick look, the second set of eyes. Everyone overlooks something simple in their careers, maybe that should be a separate thread, sometimes quite amusing. I can start:
once, needing a long-prepared filtrate, while filtering I forgot to put my receiving beaker under the funnel, and my precious filtrate spilled out onto the bench in the exhaust hood. And a co-worker, a PhD Analytical Chemistry, once acidified his sample with HCl to do a chloride titration; yes, he used a lot of AgNO3 before he realized that. We've all done stuff like that.

[Uwe Neue]

Sometimes, you can INCREASE sensitivity by DILUTING the sample. The details if this is possible depend on the analyte (solubility and organic concentration needed for elution) and the equipment (maximum injection volume).

Assume the analyte elutes in 50% organic, and it is dissolved in 100% organic. Making a 3:1 dilution of the analyte with water (or buffer) gets you to 25% organic. You can inject very large volumes (mL range) of analyte dissolved in 25% organic into a mobile phase of 50% organic without getting any peak distortion, because the analyte enriches at the column top in a very narrow band. You are effectively executing a step gradient.

Works like a charm! Limits: if the eluting mobile phase has a very high water content (limit probably around 15% or so) or if the analyte solubility is too low (this is rare in analytical chromatography). Another limit is the upper limit of the injection volume imposed by the injector.

[Hollow]

@Uwe:
Do I remember it right (once attended a course at Waters DE), that on the Alliance and 717 injectors, I always have 5% of the flow beside the needle and not through it? So this could be enough to dilute my injection's eluting strenght, and that's why our Waters HPLC are a bit less sensitive to this overload effect?

Both the Agilent and Varian uses some Rheodyn valve, so the sample are entirely transfered as a plug from the loop to the column.
I think the UPLC now uses some equivalent injection technique and will also be more sensitive to this effect, right?

[Uwe Neue]

You are correct: all the older injectors used a bypass loop that was always active, i.e. some flow was always going through this bypass loop. Indeed, it did serve to dilute the sample, but is 10 to 20% enough to explain the observation?

[Hollow]

Thank you Uwe!

in this case it seems to be enough

(The mass injected is about 300 ng, and the initial eluent is 15% ACN, compound elutes at about 30% ACN)