A curious case of high pressure!

[AnaChem1983]

Hi all,

Hope everyone is well.

I've encountered a curious pressure issue with my Waters Acquity Arc Bio HPLC. These HPLC's have the ability to switch between HPLC and UHPLC capabilities.

On Friday last week (8th April) I finished some work on a method I regularly use in the lab to monitor micro-fungal fermentation. Samples are prepped for analysis by centrifugation at 11000 rpm and then the supernatant is filtered through a 0.2um filter. This filtrate is what is injected into the system. The method in question uses an isocratic 5mM sulphuric acid mobile phase, flow rate 0.7 ml/min, USP designation L17 column (Waters Ion-Pak in this case) 300mm x 7.8mm, 7um particle, column oven 40degC, RID detection 55degC, inj vol 20ul. The operating pressure for this is normally 1150 psi and this is what was recorded at the end of operations on Friday. This is all done through the HPLC path of the instrument.

Skip forward to Monday (11th April) and I switched the system to the UHPLC side for a method I'm using to monitor mycotoxins in the same micro-fungal fermentations. Samples are prepped in the same way as above. The method uses a C18, 150mm x 3mm, 2.5um particle column. Mobile phase is a HPLC water:MeCN gradient, flow rate 0.6 ml/min. It starts at 17% B rises to 45% B over 20 minutes. There is then a 5 minute clean at 95% B before returning to 17%B for 5 minutes equilibration. So 30 minutes in total. Column oven 30degC, detection DAD 192nm to 800nm inj vol 50ul. The operating pressure for this is normally around 6500 psi (with the starting mobile phase) which isn't particularly unusual for a UHPLC application.

However after I'd set everything up on Monday I recorded a pressure of nearly 8500 psi! I did a few basic checks but nothing seemed untoward and so, perhaps foolishly given I wanted to get some data that morning, I ran the samples I had given the pressure, whilst high, was safely within a) the pump limit and b) the column limit. After concluding the analysis I allowed the LC to tick over at 0.1ml/min overnight with 100% MeCN (the storage solvent for the column is 100% MeCN in line with what it was shipped in).

The next morning I removed this column and set up my system for the aforementioned ion exclusion method. The pressure recorded was 1750 psi, a huge increase on what I'd seen only a few days ago. I ran a SST sample and the chromatography was horrendous. Preparation of fresh mobile phase and then SST (sequentially) did not improve the situation.

I then tried switching back to column 2. On this occasion the pump started to trip out due to the pressure being above the limit of the pump (9500 psi). I therefore ran the system at 100% MeCN with all other parameters the same. This gave me a pressure of 3149 psi. I left the instrument to tick over at 0.1 ml/min 100% MeCN again overnight. When I came back in the morning the pressure was nearly 3000 psi!

I've used a small 5cm C18 column in between switching my main two columns and a consistent pressure of 880 psi is recorded each time. I've ruled out there being any sort of blockage before you get to the column by running the system without a column and pressures are negligible for the HPLC component and what you'd expect for the UHPLC. Therefore it would appear that the issue lies with both columns given the system pressure for both seems to be increasing with time.

However I'm keen to try and find out what has caused this, particularly given column 1 and my routine HPLC method were fine on Friday. This issue appears to have started on Monday (or at least has been picked up on Monday). I've got some new columns on the way but, of course, without understanding why I'm possibly doomed to experience the same fate again further down the line.

Possible information overload here folks, but then the opening post from Tom does say provide as much as possible, lol. Any help would be much appreciated as I venture forward.

Cheers, James

[wss]

Have you tried backflushing your columns? You might be getting a buildup of stuff precipitating out at the head of the column over the course of your gradient. Possibly if you have salty/buffered samples, and an injection volume of 50 uL is relatively large...

I've had to deal with a similar issue recently on an LC-QQQ, not quite sure where the precipitate is coming (buffer components should be soluble up close to 100% organic phase) from but have recently considered it might even be the packing breaking down. I would try backflushing before getting new columns generally if I have the time.

[AnaChem1983]

Cheers wss.

I have given that a go and it didn't seem to alleviate the issue unfortunately. I'm just giving the system a good flush today and making sure there's no background contamination without the column in place.

I'll get there

[TylerSmith123]

Hello James,

I've read this thread every day for the past three days trying to figure out what could be the cause of these issues. I'm not too familiar with your sample matrix, but I have a feeling it may be causing some of these issues. Can you explain more about your sample prep? The filtrate that you get from your samples is it acidic or acidified prior to injection onto the system? I know this is relatively basic information, but with your acidified mobile phase you may see some precipitation of your sample onto the system especially on higher sample loads.

I think wss has the right idea where your method may, inadvertently, precipitate something onto your system. Your sample (again, I'm not familiar) may contain some salts that could be precipitated at higher ACN concentrations (personally, I would change the % ACN of your wash phase to something a little lower), or even something else precipitating onto the system/column. From your gradient it seems like the majority of your compounds are relatively polar and some may be polar enough to precipitate?

In all honesty, I'm grabbing at straws here as a lot of your diagnostic tests indicate pretty clean equipment other than the two columns. Rarer even that both of them fail at the exact same time, however they are both running the same sample (potentially even lot of samples), and depending on their work-ups something may have failed or been missed. Typically the pressure would build slowly in these scenarios, but perhaps a filter of yours failed during the preparation or another step didn't not preclude certain compounds etc. It also seems like your columns are gaining pressure when they are off-line, suggesting that something is precipitating in your storage solutions (which should be relatively impossible as it looks like they were only stored for ~3 days).

One paragraph that I believe deserves emphasis is that of the column 2 wash procedure. The pressure was 9500 psi at a 0.6ml/min flow-rate, which (under 100% organic conditions) provided you with a pressure of 3149 psi at that same flow-rate. When you left the instrument going overnight at 0.1 ml/min and checked in the morning, it was roughly the same at the pressure as the 0.6ml/min wash you did the previous day. This is very interesting and suggests a few things. The first is that this contaminant cannot be solved with ACN, and thusly, the source of this contamination/pressure build-up cannot be cleaned properly with your method's conditions. It potentially suggests that either ACN is precipitating out this impurity, or that the impurity lies within this mobile-phase constituent.

All in all, it's a very interesting issue and I'd recommend trying a smaller, stand-in column to try your same wash method on and see if you get a similar pressure increase from the overnight ACN. I'd also review the sample-prep for your samples to ensure that our injecting in your mobile phase so you're able to visibly tell if there will be issues with your sample upon injection. I'm sorry that I could not be more of help, but update us on any experiments you've performed in the meantime.

[AnaChem1983]

Hi Tyler,

Your apparent persistence in reading about my issue is very impressive, haha!

Yes I'm thinking that the sample matrix is probably involved in this problem. The sample matrix is overwhelmingly aqueous in nature with various mineral salts and glucose dissolved into it (these are the feedstocks for the fermentation). As described previously this is obtained from centrifugation of the broth at 11000 rpm for 3 minutes. The supernatant is decanted off and then passed through a 0.2um syringe filter. The filtrate is what is then injected onto the system.

I don't think the ion exclusion method I'm running is the issue. I've ran this analysis for a long time now with two companies and not noticed any issues beyond the column reaching it's natural end point. Plus the mobile phase is entirely aqueous and everything in the matrix should dissolve without any issue. However it's not beyond the realms of possibility that something in the matrix is crashing out at 95% MeCN. My intention is to prepare some of the supernatant and then add it to beakers of different strength MeCN and see if anything crashes out.

I've given the LC a good clean down since my last post and pressures without a column and with the sample manager connected directly to the detector seem fine with no apparent detector issues. Onwards and upwards!

[TylerSmith123]

Hi James,

I'm glad to hear that you are coming along with the investigation. Please update us if there are any breakthroughs or important discoveries.

Unfortunately, this problem seems more spontaneously than consistent. Your sample prep sounds good to me, but your sample matrix is, admittedly, quite messy. Which isn't a problem of course, but you may have columns dying sooner just due to the nature of the sample. This should not be the explanation as to your issues right now. It seems like you were taking great care of your columns before this and the instrument in general. I've still got a few ideas on what could be the cause, but they are really out there.

Could your injection method be causing this phenomenon you're seeing here? Obviously, all of the precipitable solids should be cleaned off by the start of your method, like those mineral salts your mentioned, so they should not even be in the system by the time the 95% ACN hits and, potentially, precipitates them. It sounds like you guys keep good logs of user use so it would be hard to imagine someone left a sample on the system, but who knows. Additionally, depending on if you're using manual injections or automated, there could be an issue with that technique. I've had a prep sample precipitate in my needle due to the incompatibilities of my wash solvent past a certain concentration threshold, maybe one of the samples had an unusual amount of dissolved solid or something. Like I said, these ideas are pretty out there haha. If it's a manual injection, it could be something that was stuck in your port if it wasn't cleaned lately. Your samples seem pretty messy and I imagine they could contaminate a manual injector pretty quickly.

My last idea revolves around some sort of mechanical error on the instrument, and as long as your have regular PM or do it yourself, I think, for the most part, this idea is avoidable. It may even be unprovable unless you get a tech out to look at the columns and instruments for you. Perhaps you had some sort of o-ring failure, or some particle found it's way to the head of your column and blocked it. However, this still does not explain the second column's issues and why it has an increased pressure while you were running literally nothing on it until after you had cleaned the first column. Perhaps when you changed the pressure it gave enough force to dislodge the rest of whatever built-up/broke-off and lodged itself into the other column following solvation/decay using ACN?

Anyway, I'm sorry I couldn't give you a clear-cut answer and can only help troubleshoot. Continue backwards up the flow-path and confirm each of the pieces as within normal pressures. Try the system with a union or an old column and see if the pressure has changed or not, I'd only place those other columns in-line once I know that the issue is not coming from the instrument as it seems the problem is only being exacerbated while they're in-line. I like your idea of trying the supernatant at different strengths of acetonitrile, I bet you'll see some of those mineral salts precipitate. Unforunately, this isn't a great explanation as your mobile phase does not contain buffers and it's only your injected samples so any precipitation would have to come right after injection (barring an issue where sample is injected AFTER the set time, or a second injection that could interact with the high organic gradient time).

Good luck to you!
Tyler

[AnaChem1983]

Hi Tyler,

You're bang on in that it all seemed so sudden and this definitely muddies the waters a bit.

You're point about the mineral salts not being present by the time 95% MeCN hits the column is a great one. I considered this myself. If I did it at the starting MeCN % and observed I suppose that might rule out the precipitation right after injection.

Of course the other problem is that whenever I've thought of something evidence from elsewhere contradicts this. Like the example you mentioned with the o ring.

I'll keep you posted when I have updates. Thanks for your help and advice.

Kind regards,

James

[AnaChem1983]

Hi all,

I've installed a new column into the instrument and the pressure is back to what it was prior to what we shall call "the issues"!

As I'm in quite a fastidious frame of mind regarding HPLC at the moment I just wanted to run a question by you all.

With regards to the pressure limit that is specified for a given column.......is this pressure exceeded when the pressure reading on your software goes above this value OR is the pressure exceeded when the pressure reading on your software MINUS the system pressure without a column is exceeded?

The only reason I ask is that I'll go over the limit at 0.7 ml/min (the column isn't the same manufacturer as the previous one as I needed a column quickly and I couldn't get it quick enough from Waters). However I have seen technical notes from Agilent (I've attached the link) that states you've only surpassed pmax for the column when you go over the latter description. This is the asterisk at the end of the tables on page 3 of the attachment.

Any advice always appreciated. Cheers folks.

https://www.agilent.com/cs/library/datasheets/Public/5991-6743EN.pdf

James

[TylerSmith123]

Hi James,

So this is an interesting question, and by no means am I an expert but this is what I think. The pressure measured on the software is coming from the pumps and they're measuring the pressure that is mostly contributed by the first resistor (ie the column), as well as any other parts (mostly minute) from tubing, your detector, etc. The pumps are programmed to pump at a constant rate, and that's what they are made to do, and not stay within pressure. So the pump is generating all of this pressure by trying to push the solvent (and thusly your samples) through the stationary phase, your tubing, your detector, all the way out into your waste collection. Now, what is measured is the ultimate pressure IN SERIES of all those components together. So lets say you had a column with some tubing that gave a pressure of 1000 psi, lets say that you want to put on some sort of restriction, like a guard column (and let's make it dirty so it gives a back-pressure of 100 psi). At the front of that guard column, it is experiencing 1100 psi of pressure, but at the front of the column, the pressure is experienced is only 1000 psi. It all effects the pump the same as it is providing the force to move the liquid with your sample, but the parts are under different pressures. So while your pump may be experiencing 1300 psi on a new column, only the pressure from the components following your column contribute to it's effective pressure, ie tubing, detector, fraction collector, etc. So the only pressure you would need to account for is that (which is generally extremely small compared to the column, but it's all somewhat relative to what chromatography you're doing). This is my opinion on things, based off of what I've read and heard.

The total system pressure without a column includes all of the facets, so it may be harder identifying the back pressure after the column exclusively, but I would say the total system pressure is somewhat misleading when applied to a column. However, it never hurts to be safe than sorry! I'd also love to be corrected, so if anyone else has anything to add, head their credence.

[hplcuserx]

I think this page can help you with this problem.

https://whatishplc.com/hplc-troubleshooting