Bad Baseline

[bisnettrj2]

There isn't a difference between running the wash at the end of the method or not. I've had injections with bad baselines following both isocratic runs and runs with the wash at the end. Besides that, I've got the detector shut off for that part of the run, so you're only seeing the isocratic part of the baseline.

Some additional info - I'm running a binary pump with the solvent selection valve installed, using channels A1 (water) and B1 (methanol). I have isopropanol on channels A2 and B2.

[HW Mueller]

The solution is systematic checking, instead of wild guessing (lets try a new lamp , etc., lets clean the cell . . . .).
What does the baseline look like when you just run 52:48 H2O:MeOH for lets say 3 hours, without changing or doing anything? What do these weird actions (induced severe pressure changes, lamp shut-off, I wouldn´t exclude the washing act yet either) do to the basline, one at a time?

[danko]

Hi Bisnettrj2,
Here is what happens: When you â€

[bisnettrj2]

First off, I don't know what HWMueller means by "weird" actions - I'm running extracted soils from military firing ranges, and there are a lot of organic non-target co-extracts that will carry-over into the next run if I don't wash my column at the end of my run.

Also, danko, I'm not sure I understand what you mean. If my methanol channel was blocked, or I had air in the channel, I would have erratic retention times in my analysis. My retention times are good, even when this baseline issue happens. Just because I go from pumping 48% to 100% doesn't mean a blockage in the methanol channel wouldn't manifest itself during the isocratic portion of the run. I'll examine my method again, just to be sure.


Also:

1. I wasn't clear earlier - the detector/lamp isn't turned off, I'm just not collecting data during this part of the run.

2. As I've posted before, I run this same type of column on a substantially similar system (only difference is DAD vs MWD).

3. I run a methanol wash at the end of all my analytical runs on all my systems.

4. The pressure changes aren't "severe", and are related to the wash of the column - Normal pressure is about 190, going to 90 bar (100% methanol) over about 5 minutes, holding at 90 bar for about 5 minutes, and going back up to 190 over another 5 minutes, with about 10 minutes of re-quilibration time.

To describe what happens when I perform the column wash in regard to the baseline - at the start of the wash, the baseline rises substantially, coming to a relative plateau of junk peaks that I'm washing off the column. As I transition back to 52:48, the baseline drops at relatively the same slope as it rose, goes past zero slightly, then rebounds to approximately zero. Like I posted previously, this is almost exactly the same method I use on another system, and I don't have the baseline drops and rises duing the isocratic portion of the run that have been plaguing me on this system.

When I go back to the lab in a couple hours, I'll program in two instrument blanks to run for four hours each, with no column washing in-between. One method will be without reference wavelength, and the other with a reference wavelength of 360 with a 100 nm bandwidth for both wavelengths I monitor. I'll check on those runs when I go back in this afternoon/evening.

[bisnettrj2]

First run this morning, of a low standard. Peaks are where I would expect them, even with the baseline craziness. Used a reference wavelength, performed column wash at end of run.




Another low standard. Again, RTs are stable. Used reference wavelength, performed column wash at end of run.




Third run of the day. An instrument blank. Used reference wavelength, performed column wash at end of run.




Fourth run of the day. No reference wavelength, no column wash.




Fifth run of the day. Reference wavelength on, no column wash. 4 hour runtime.




Sixth run of the day. No reference wavelength, no column wash. Approx. 3 hour runtime.




Finally, the corresponding pressure profiles. Keep in mind, the bottom two profiles are for runtimes of 3-4x longer than the profiles above them.




Let the discussion commence. I don't see anything in the pressure profile that explains this rhythmic, regular weirdness. Teach me, proverbial Obi-Wans.

[danko]

OK, I see 2 problems:

1. You have an air bubble in the flow cell. I don’t know if it’s been there all the time or it’s been introduced under the recent tests. Maybe it’s been there all the time, but because of the reference signal correction it’s been “hiddenâ€

[HW Mueller]

There is no statement that says that a washstep is weird, I essentially said that this washing step should be watched for some weird action in connection with it. And weird it is as far as the results are concerned.

So, I wonder why the pressure profile starts with a rise in the last run, was the pump stopped and started again? Why?
Do all runs have these basline repetitions, continuously?

[grzesiek]

It seems to me that now you have problem with the pump
cleaning the valves, priming the pump with high organic can get thing back to where they were hopefully

do you degass solvents? is your degasser working properly? What is going on when you premix mobile phase and use only one channel? - i read about your degasser but it does look satrange and checking it is what I would have done

[Hollow]

[quote="danko"]OK, I see 2 problems:

1. You have an air bubble in the flow cell. I don’t know if it’s been there all the time or it’s been introduced under the recent tests. Maybe it’s been there all the time, but because of the reference signal correction it’s been “hiddenâ€

[bisnettrj2]

Hello all...

The outlet capillary was a short stainless steel capillary to a ZDV union with a waste line (came with the system). This morning I replaced the outlet capillary of the flow cell with, well, an old guard column I had lying around. Increased back pressure about 25 bar. I'm also flushing out the flow cell with the isopropanol I have on the instrument. I will resume my 4-hour isocratic runs, alternating no-reference and with-reference wavelengths after I purge the flow cell with IPA, then methanol, then mobile phase. Will report results tomorrow.

In response to the responses:

1. I heard air bubble in the flow cell twice, hence the change in outlet capillary, and the purging of the cell.

2. Danko - The runs I posted showed baseline noise with and without a reference signal, and with and without a column wash at the end of the run. Also, I already stated that I do not collect data points after the end of my analytical run of 62 minutes, and I posted that my column wash starts at 59 minutes. Put them together, and you have a pressure drop before the end of the analytical run, as it is easier to pump 100% methanol then 52:48 H2O:MeOH.

3. HWMueller - "...the pressure profile starts with a rise in the last run"? Look at the scaling of the pressure profiles - are any of those changes significant? And the pressure always jumps a few bar after the injection valve switches from bypass to inject, on every system I run.

4. I don't know that the degasser isn't the issue. I can say this - when Agilent degassers can't hit a vacuum level specified in their on-board programming, they go into a warning state (yellow LED) for about 8 minutes, then they display a red LED and shut off. Mine doesn't do that, and I hear the vacuum pump click on and off, so it seems to be working. I will try degassing manually and bypassing the degasser if the change in outlet capillary does not help.

Thanks for the help, I do appreciate it.

[grzesiek]

"I will try degassing manually and bypassing the degasser if the change in outlet capillary does not help. " - I would try that even if changing outlet capillary helped

[bisnettrj2]

Why? If changing the outlet capillary fixes the issue, why do the extra, unnecessary step of degassing manually and bypassing the degasser?

[grzesiek]

to learn
I would install old outlet capillary and check if I can get good results with it, but hey that's just the way I am, I'm not saying you should do this, I like to know my equipment
I wish you good luck and hope you'll post results you get

[HW Mueller]

bisnettrj2, you didn´t say whether run 1, etc., also have this periodicity.
In any case such a baseline is a severe problem. Dancho seems to be closest to an answer in regard to air bubbles. It seems to me that you have a periodic air leak somewhere, to me it is weird as I have never seen this. I have gotten dissimetric sawteeth from a continuous air leak in the plumbing. This caused a periodic builtup of bubbles which left the cell (apruptly, of course) when they had become large enough.

Regarding ignoring such a problem if detector exit pressure masks it: Even dissolved air causes scattering and/or refractive index changes which may wreck your baseline (maybe as seen initially).

Also, at this point I don´t see how Dancho´s second possibility (pump) can be ruled out. If the periodicity is caused by the pump in a continuous fashion your rt may still be repeatable.

[bisnettrj2]

But wouldn't you see a pump problem in the pressure profile (large, rhythmic pressure rises and drops, in concert with the rising and dropping baseline), or in the RTs of peaks (such as in a delivery problem from one of the two pumps within the binary pumping system)?