Large Pressure Ripples with Decreased Retention Time

[LukeFrenn]

I am using an Agilent 1260 Infinity II with the Quaternary Pump. I was setting up to run a method I had previously used after someone else had used the machine. This involved a column change. After the column change I was flushing out the column which gave some larger pressure ripples from bubbles as expected which resolved itself. However shortly after, the large pressure ripples returned (ripples of 4 to 10 bar) and then continued for a total of approximatley 16 hours of run time. During that time we purged multiple times, swapped the column back out, removed the column altogether, flushed the whole system through with various ratios of acetonitrile, water, and methanol; and changed the guard column. Nothing seemed to work. I initially thought it was a trapped air bubble but due to its persistance I began to doubt that. Also at times besides the large ripples, the pressure would fluctuate erractically or in larger patterns that were always changing. I ran a standard for my method and it had its retention time shifted 6 minutes earlier than usual. I run isocratic at 60% 10 mM potassium phosphate, 33.3% Acetonitrile, and 6.7% Methanol prepared in one bottle and pH'd to 6.5. After my compound of interest elutes I switch to a bottle with 20% 10 mM potassium phosphate and 80% organics. The column is an agilent XDB-C18, 5 micron, 4.6x250mm RP with a guard column. At last it spontaneously improved and the retention time went back to its usual. The pressure ripples are about 1.5 bar (1% of running pressure) and it is less erratic. However I feel that the pressure changes during a fast gradient are softer than usual still and during the switch to 80% organic the ripples increase to 4 bar (I accept some increase is expected, however it still seems excessive. The problem has as far as I can tell greatly resolved itself (I am not sure thought that is gone altogether), however I am curious as to what was the cause and is it a sign of something failing in the instrument. For degassing I rely on the inbuilt degasser and I never saw bubbles after the degasser during normal operation. I am a graduate student doing HPLC so any insight I could learn from is greatly appreciated.
Luke

[Multidimensional]

Thank you Luke for providing a lot of initial info, but please provide some critical missing bits ...
What is your flow rate? This is really important to know.
You stated that your method is isocratic. What is the NORMAL system back-pressure you observe with this method (in psi or bars)?
Which detector are you using (and what are the exact settings) ?
Can you show us a chromatogram with the noise showing the scales?

There are many reasons for pressure to fluctuate and we really would need to be on-site to properly troubleshoot it. However, it could be something simple as you need to replace the PTFE frit inside the Prime/purge valve (this should be replaced every 1-month. Yes, every month). As it starts to clog up, all kinds of pressure problems start to happen. You also may have a mixing problem too (you mentioned that peaks change retention time which happens when either the flow rate is NOT what you think it is OR the mobile phase composition changes). Failure to properly flush and prime the system before use can also cause these problems too (and it should never take 16-hours to do this! More like 10-20 minutes at most). Yes, the degasser may be broken (and no, you probably will not see any error messages for this as they usually appear YEARS after the degasser is broken) as they last at most 3-years in the 1260 Quat and Binary pumps (very short lifetimes). When the integrated degasser is broken, the pump's performance will be poor in some cases, but it will also be adding debris into the HPLC system flow path resulting in pump instability, ghost peaks, and all kinds of other problems. One or more clogged solvent pickup filters can also lead to "pump problems" too so be sure and install new glass filters on the solvent bottle lines. If your pump is from Agilent's "value line" (IOW: No AIV), then it may have a contaminated passive inlet valve. When these valves start to 'stick', the pump will cavitate and become unstable with different solvent mixtures (clean the valve. NOTE: If your pump has the active inlet valve, leave it alone OR change the internal filter in the valve, but only have this done by an experienced tech).

BTW: Which Agilent Quat pump do you have? Please share the model number (found on the front label, it starts with 'G ####).

Here is a related article from our professional training classes. You may find it helpful, but please contact a professional chromatographer for assistance.

"Diagnosing & Troubleshooting HPLC Pressure Fluctuation Problems (Unstable Baseline)"; https://hplctips.blogspot.com/2014/01/d ... -hplc.html

[LukeFrenn]

Thank you for the info and resources.
My flowrate for the isocratic portion is 1.5 mL/min with a normal backpressure of about 149 Bar. I am using the 1260 Infinity II DAD WR detector (G7115A) at 246 nm with a bandwidth of 10 nm, reference wavelength of 600 nm, response time at 5 Hz and slit set to 16 nm.
I believe the pump has the active inlet valve (has a solenoid on the inlet line) and the model number of the pump is G7111B.
After running some standards curves I've noticed that the issue with the eractic pressure fluctuations and large ripples comes and goes which reflects itself on having variable retention times (varying 0.5 to 1.5 minutes) within a sequence of runs and with the same batch of mobile phase.
I'll work on attaching some graphs with the pressure trace.
Luke

[Multidimensional]

Appreciate the extra info. A few quick comments in no particular order:
(1) Ripple of 1% overall is considered very good and does not indicate a problem.
(2) A flow rate of 1.5 mL./min for a 4.6 x 250mm column with a 5u particle is not within the acceptable range. Please use a linear flow rate of 1.00 mL/min and follow good chromatography practices.
(3) Immediately turn OFF the Reference Wavelength feature that you turned ON. It could be causing so many problems with your method... Never use this specific software feature, esp w/o professional HPLC training. If anyone there has suggested you use this feature, then please correct them and explain that it must NOT be turned on unless you want ALL data collected on your HPLC system to be declared Invalid (as it will be if you use it). This is one of the most common mistakes we see inexperienced chromatographers make (NOTE: as a student, not a professional chromatographer, you would not be expected to know this so do not feel bad). If you would like to understand why this is, then please read this article; "REFERENCE WAVELENGTHS (as used in HPLC UV/VIS)"; https://hplctips.blogspot.com/2011/03/r ... -hplc.html
(4) The previous article I linked, "Diagnosing & Troubleshooting HPLC Pressure Fluctuation Problems (Unstable Baseline)" will address 99% of your problems and issues. Please work with an experienced chromatographer to resolve them.

[lmh]

I'd respectfully disagree with the idea that 1.5mL/min is too much for a 5u 4.6mm diameter column; the back-pressure isn't excessive, and depending on the exact design of column, while it may not be the perfect point on the van Deemter curve, many methods use flow-rates optimised with speed of analysis in mind, not just best resolution.

On reference wavelengths, I'd agree that it's generally not a good idea to include it in the method, because you can't see what the data looked like before you removed the background. If you subtract a reference during data-handling rather than acquisition, you can see the raw data and control what you do with it. For example, if the ripple is worse in your background channel than at the wavelength you are using for detection, then subtraction is making things worse. But for non-specific effects, such as those caused by an inadequate mixer, subtracting a background at a "safe" wavelength where you expect no chemical absorbance can be a very useful technique; just do it post-acquisition! BUT, and if you work in a regulatory environment this is a really important "but", it's not true that your data are invalid if you collect them with the reference turned on. In fact, if you're in the sort of environment where methods are validated, you should use the method exactly as stated in the SOP, because that's the method that's been through all the validation procedures. If you change it, you may have to revalidate.

On what caused the problem, given that it's transitory and largely gone, but happened after someone else's method, there are an awful lot of possibilities. Maybe something precipitated out in an inlet filter while you were changing solvent? (For example, if someone had a salty aqueous buffer, and dropped the inlet filter into what was supposed to be a safe intermediate solvent like 100% isopropanol, and some salt precipitated). Frit in the purge valve was a sensible thing to check. If it ever happens again, it's worth checking whether the pump is delivering the correct flow, and whether there is zero flow in channels that you're not using. If you've changed methods, and precipitate or incompatible solvents have been let loose in the system because of a misunderstanding during change-over, check-valves and gradient proportioning valves may do all sorts of silly things.

[Multidimensional]

"LMH" wrote regarding using software 'Reference Wavelengths" in a method: "BUT, and if you work in a regulatory environment this is a really important "but", it's not true that your data are invalid if you collect them with the reference turned on. In fact, if you're in the sort of environment where methods are validated, you should use the method exactly as stated in the SOP, because that's the method that's been through all the validation procedures. If you change it, you may have to revalidate." -

It IS in fact true that your data is invalid... And that all data collected using such a method in the past and present is completely invalid and unscientific. Revalidate it? I sure hope so as you have just discovered that the method is invalid and must not be used as written. Don't stick your head in the sand and pretend you did not see anything. Report it. The use of any destructive based "Reference Wavelength" software feature which destroys the original data and replaces it with new data is 100% INVLAID in ALL cases, no exceptions at all. IF an untrained user develops and "Validates" a method with the "Reference Wavelength" feature mentioned turned 'ON' then the method is now 100% INVALID too (as it was developed and "Validated" by someone who did not have training in operating the HPLC system). For validation to be valid, the user who validates the method should have the required training and experience to do so. If they make such an error, then they clearly are not qualified to develop methods (and certainly can not "validate' them). This is just another example of why you should never just follow a SOP or procedure because 'someone before you did' or someone told you, 'that is how it is done'. Following procedures that are by definition known to be wrong and invalid demonstrate a lack of training and knowledge. People who ignore these facts mislead and perhaps harm others by providing false results.

[Multidimensional]

This article from our professional training classes may help to educate HPLC users at all levels with the dangers of using instruments or software features without proper training. Never stop learning. Keep asking questions. Keep learning... Please share the article link with others so we can help to educate more users.

"REFERENCE WAVELENGTHS (as used in HPLC UV/VIS)"; https://hplctips.blogspot.com/2011/03/r ... -hplc.html

[LuccasLN]

Hey Luke and everyone,

I think all the discussions here are valid but we are all discussing the solution for the consequence and not the actual cause. Please, let's revisit the beginning of the first post: "I was setting up to run a method I had previously used after someone else had used the machine. This involved a column change. After the column change I was flushing out the column which gave some larger pressure ripples from bubbles as expected which resolved itself.". I did not see any discussion about this.

First point is that you must know what has been used by the other user, it might bring a hint on what to use to take it out or even if you must keep an eye on the instrument (idk if you are in a multiuser lab from a university, for example). you never know wether the other method had a whole lot of salt, an IP reagent, a somewhat different solvent for some reason...

Second, one must clean the system before puting in the column and not putting the column and flushing stuff! you see, there is a whole lot of "extra column volume" that is filled with the previous user's MP and it does have potential to harm your column, so just be sure your instrument is good and clean and it has solvent that is compatible to your column before you have your column in. you can have the capillary tube pointed to a becker or have it closed with an union if you must (might not be the best idea due to risk of carrying unknown dirt to frigile parts such as the detector flow cell).

i have had a fair share of all sorts of oddities from situations where the system/column was being equilibrated without proper cleaning previous to putting the next-method-column in the instrument. i have seen columns dying due to clogging, poor equilibration due to solvent heterogeneity, salt precipitation... and so on!

i personally think that there was something maybe "stuck" in one of the parts and pieces from your pumps like what multi has pointed out (maybe i would take a look at the check valves as well). but mobile phase substitution is a very serious matter and should be evaluated very carefully in order not to create unecessary misunderstending.

[LukeFrenn]

Hi All,
Thank you for the replies and discussion. I never fully understood what the reference wavelength did and honestly didn't have time or felt it was necessary to investigate. I was just told to use it, but I will revisit those decisions after the points raised.
It is a multiuser instrument and to the best of my knowledge, no salts were used previously, only acetonitrile and water with TFA added. I don't think the instrument itself was cleaned without the column. Since my post the problem did come back which resulted me in not performing my analysis. I then prepared the instrument for switching back to the previous user (column change) and since I use salts in my method I spent a lot of time purging the pump, mixing valve, and degasser with water before putting pure organic solvents back. After this process when I was flushing the column I noticed that the pressure went through some large inverse spikes (drops) in pressure that coincided with changes in the sound the pump was making. After some time the problem completely resolved itself and the user after me has had no issues. So I assume it was probably a dirty checkvalve or some salt precipitated somewhere that finally freed itself.
Thank you for the all the links, I will go through them and I appreciate the time taken to inform a novice user such as myself. I do have someone experienced in chromatography I can refer to but since they are not my supervisor, they don't have a lot of time to give me. I like learning and solving problems so I have taught myself a lot about chromtagraphy.
Luke