Baseline ripple problem

[sdegrace]

One thing I think doesn't come across in the way I have presented my chromatograms is that the "ripple" is pretty much constant over the whole time the gradient is "shallow" (in other words, the first parts of the programs). In each, a window was chosen to get a good qualitative representation of the data, although in retrospect it would have been smarter of me to present the data for each method in exactly the same way.

The time chosen in the third chromatogram was selected because it shows the peak I was interested in for LOD. In that chromatogram, the "frequency" and "amplitude" are lower (the ripple is less perceptible), and significantly, the noise measurement and hence s/n (using peak to peak noise) is much more consistent from injection to injection for peaks at low concentration.

I fully expect the new mixer to resolve the issue to the degree that I need, but to me these data also suggest that the instrument on which the first chromatograms were collected is not functioning as well as should reasonably be expected to in this regard, which is an issue I plan to raise with the vendor.

[unmgvar]

sdegrace

the different results between your two systems could come from two more factors.
1. the "speed" at wich you collect your data, which can smooth stuff depending on how you set it.
2. clogged filters that can cause the creation of very tiny bubbls that can cause an increase in the drift and noise of systems.

again this is only regarding the differences between your systems.
your main problem is what other have already posted.

[sdegrace]

The instrument methods between systems are identical - but other factors such as filters could be contributing. The main user-servicable filter other then the reservoir filters within the instrument that I'm aware of is the pre-column filter, which has been recently replaced on both systems. There could be something else I'm not aware of.

[HW Mueller]

Looks to me as if your integration is too high by about 100%. I would continue with the setting of the first apparatus (or even more sensitive) and clean up the mixing act.

[danko]

Hi sdegrace

The first 2 examples/chromatograms are zoomed in at least 2 – 3 times more than the last one.
Maybe you should use the same absorbance scale for all tests.

Best Regards

[hookeslaw]

I think Bruce Hamilton may have a point, you can change the stroke volume on waters but i believe it depends on the software you have. you either have to entre it manually ( millenium) or it is altered (optimised) automatically through empower

[sdegrace]

I'm using a 50 uL stroke volume for all of these methods. I tried changing to a 100 uL stroke volume the other day but it didn't really improve anything. Anyway, I have the parts on rush order and should get them today. We'll rip the LC apart, plug in the mixer and see how she goes. Fingers crossed! I'lk post to let you know how it goes.

[sdegrace]

By the way, we are using Empower 2, not that it is especially relevant, but I think we do have pretty good control over the system

[sdegrace]

We received the mixer (150 uL static mixer) today and installed it right after the pre-column filter. The parts arrived today as promised and we got a surprise visit from our Waters rep right after who installed the mixer for us and stayed until we verified it was doing what we wanted.

The baseline is vastly improved and from my point of view the problem is completely fixed. I'll try and post an "after" chromatogram when I get a chance later.

Thanks so much to everyone who helped! It is very satisfying to not only have a solution but a rational explanation.

[Bruce Hamilton]

I'm using a 50 uL stroke volume for all of these methods. I tried changing to a 100 uL stroke volume the other day but it didn't really improve anything. .

I'm glad that your problem is eliminated by the mixer.

Just to correct a misperception. To reduce ripple, the stroke volume would normally be reduced, and the compressibility of the solvent also optimised. Most pump manufacturers that offer the ability to adjust these parameters would probably discuss the details in their manuals.

Pleaze keep having fun,

Bruce Hamilton

[sdegrace]

50 uL is the lowest setting available in Empower, and that was already set... I always pick the lowest volume that will allow the flow rate I need. So I tried higher since that was the only available option that was different than I was already doing. Like I said, it didn't help

[DCL11]

We had similar problems in our lab, and our investigation concluded that the check valve is the root cause.

I believe that your system has Waters Ruby ball based check valve. This type of check valve works well for most applications. However, for some reason, if the mobile phase contains TFA, the baseline ripple happens.

I don't think the absorbance of TFA caused this problem. You can check it quickly by adding ~100ppm acetone to your mobile phase and monitor at 265nm to see if the same ripple happens. I don't expect it to happen!

To fix this problem, simply replace both check valves with Waters ceramic check valve (P/N 700002761) or ASI check valve (also ceramic based).

Regards,

[HW Mueller]

DCL11, what is your mechanism for ripples via checkvalves?

[sdegrace]

DCL11, thanks for your advice. If we need a further refinement of our baseline we will definitely look into it. We are currently investigating a similar but less crippling problem on our UPLC, and while Waters is going to give us a quote on a suitabile mixer for this instrument, that instrument definitely has ruby ball based check valves, and we'll bear that in mind.

However, the mixer we were advised to obtain truly fixed our problem within the paramters of what we require. Any ripple still present is barely perceptible and only if you're looking really hard for it. The improvement is large and dramatic and is enough to make the instrument completely suitable for our purposes. It was a very satisfying answer with a satisfying explanation.

The static mixer for the Waters instrument costs in the ballpark of an expensive column (we paid ~$900 Canadian, we probably got gypped on the exchange rate now that the dollar is at par, like with everything else ). It's an avenue I highly recommend to anyone experiencing a similar problem with similar instrumentation.

[DCL11]

DCL11, what is your mechanism for ripples via checkvalves?

The check valve has a ruby ball inside a PEEK chamber. This ruby ball serves as a switch to turn on and off the flow (in one direction and the other). I am not sure what kind of reaction TFA has with ruby ball. But apparently it does happen. The ruby ball may get stuck sometimes, or at least, not moving as smoothly as without TFA. This might disturb the flow, and cause the baseline ripple.

I also want to take back what I said about the absorbance of TFA. It is possible that the absorbance of TFA amplifies the flow fluctuation. Without background absorbance, there might still be some baseline noise (due to pressure fluctuation and/or refractive index change), but probably not as bad as what we are discussing here.

Please correct me if I am wrong.

Regards,