Bad Baseline

[danko]

Hi Bisnettrj2,

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.

As I understand it (correct me if I’m wrong) the last thing you do in a run is washing the column with 100 % methanol and then you inject the next sample/blank or whatever. It means that you don’t equilibrate the column to start conditions prior to the next run/injection. If my assumption is correct, then the explanation of the “wildâ€

[Hollow]

@Danko, he does equilibrate
even it's a bit short for this column dimension (only about 2-3 column volumes if one takes a dwell time of 1 min into account).
But I think the reproducibility is not his problem, and if the RTs are ok, why not.
If the basline disturbance are due to bad equilibration, I would excpect the effect lot earlier than 15 min


I summarized the method informations from the different posts, please correct if I forgot something

HPLC: Agilent 1100 MWD Detector
Col Temp: 30°C
Column: Dionex Acclaim E2, 5µm, 4.6x250mm
Flow:1 ml/min
Det: 214 + 254 nm
Cylce Time: 82.8 min
Data collection: 62 min

Eluent A: H2O
Eluent B: MeOH

Pump program

Code: Select all

t	%A	%B
00.0	52	48
59.0	52	48
63.6	0	100
68.2	0	100
72.8	52	48
82.8	52	48

[bisnettrj2]

danko - see below. I have 10 minutes of built-in equilibration in the method, plus the time it takes for the autosampler to make the next injection (~3 minutes).

[quote="bisnettrj2"]Pressure goes down when the injection valve goes to bypass to fill the injection loop, goes back up to normal pressure once the injection is made. The pressure drops at the end because after my analytical run, I go from 52:48 H2O:MeOH to 100% MeOH to wash the column. My pump parameters look like:

Time - 0 - 52:48 H2O:MeOH
Time - 59 - 52:48 H2O:MeOH
Time - 63.6 - 0:100 H2O:MeOH
Time - 68.2 - 0:100 H2O:MeOH
Time - 72.8 - 52:48 H2O:MeOH
Total runtime - 82.8 minutes (for actual sample analysis - standards and blanks get run on an isocratic 62 minute run).quote]

[kerri]

Bubbles seem to be right. You have H2O in pump A, MeOH in pump B. It's possible that after the HPLC mixes, bubbles appear. They accumulate in the flow cell throughout the run, and are then skipped away when you wash your column in MeOH. The whole procedure happens again, and again, and again through runs.

You could try to premix your mobile phase, degass it for a quick bit, and then run it on pump A. you could still do the wash at the end by leaving MeOH in pump B.


Good Luck!! i really hope that you can have some fun time this long weekend!!!

Cheers
Kerri

[bisnettrj2]

Sad development - I think I blew my flow cell window when flushing it. Noticed small, consistent peaks (1 mAu, about 30 seconds apart) while flushing cell. I pulled it, and noticed one side was leaking. So, when I tried to pull the hex screw holding the window in place, the screw stripped. Now I have to devise a way to get a stripped hex screw out of this flow cell without hurting the rest of the cell. I'm going to blame Murphy's law on this one and go have a beer.

[kerri]

I'll have a beer for you too!

Buenos Suerte!

[bisnettrj2]

Thanks Kerri, I appreciate it.

I think someone much wiser than I needs to write a "Zen and the Art of HPLC Maintenance" book. Might help my wife understand why I spend so much time with my "machines".

[grzesiek]

after some maintance under your belt you'll be much calmer when next thing happens, trust me, it comes with time, we all have been there or if we did not we soon will be

[christina]

If your Agilent system is the same as ours (I don't know if they could be different), we often see problems of gas forming when 100% MeOH and 100% H2O are mixed after going through the degasser. So while the individual solutions are degassed, once they are mixed in the pump, gas bubbles form and create havoc. We avoid this by mixing the two off-line.

I doubt this is your issue, but thought I would bring it up.

[bisnettrj2]

OK, I think I have narrowed down the cause of my bad baseline....




This is my baseline with no flow cell installed in the detector, and the pump set at 0 mL/min. Previous tests with pre-mixed, pre-degassed mobile phase on either the A or B channel showed the same baseline.

Am I safe to assume it's the detector here? I've changed lamps and flow cells, and the problem persists.

[grzesiek]

hmm.. or enviromental, temperature change? Can you replace detectors between the systems?

[Uwe Neue]

Check for consistent changes in the lab environment (air conditioning on/off), power supply (heavy machinery next door). etc.
The only thing that does not fit is that in previous pictures, you had a change that happened consistently 30 minutes after your injection.

[bisnettrj2]

Basically, throughout yesterday, the exact same pattern kept occuring (this pattern is consistent, I just stopped the run after seeing this same pattern without the flow cell in the detector). I called Agilent, and the tech support thinks it might be a problem with the electronics in the detector. I don't have another MWD to swap in to check if it would do the same thing with the rest of this instrument stack. I did swap in a VWD last night, though, and am heading in soon to check how a calibration curve went.

Uwe - Our lab is a stand-alone building, and I have the HPLCs on the same bench, all plugged into power strips and plugged into the power supplies located on the bench (I think they're all separate fuses, though). This phenomenon happened overnight, during the morning, and the afternoon, so I don't think the A/C turing on and off was the issue.

Why this happened, I don't think I'll ever know. I think the only thing I learned in this situation is to try one thing at a time, and give that fix time to work or fail before moving on to the next idea.

Thanks, everybody, for your help, your ideas, and your suggestions. I'll be sending my detector off to Agilent on Monday. Hopefully they can identify which of the parts of the detector was the issue. If they tell me, I'll be sure to post it here.