Agilent 1100 Binary Pump Mixing Problem

[bisnettrj2]

Hello all...

I apologize in advance for my long-windedness, but it is my nature...

I am running an Agilent 1100 system my lab bought off of Dovebid. I have been working off and on trying to get it up and running for my primary analysis, but it has served as my 'oddball' analysis instrument since we got it. My oddball analyses only use one channel of the pump for the mobile phase, and the baselines are great. However, when I use the pump to mix two separate mobile phase constituents (methanol and water), the baseline is noisy and inconsistent. I finally had some time to do some troubleshooting last week, and I premixed two 50:50 portions of Methanol:H2O and pumped on channel A alone, then B alone, and then 50:50 A:B with the premixed mobile phase on either channel. The baseline looked great, very 'quiet'. I then switched back to 100% water on channel A and 100% Methanol on channel B, and used the binary pump to mix 50:50. Baseline is noisy again. So, I suspect the mixing system in the pump is not performing adequately. However, the pump diagnostics do not indicate any part of the pump as malfunctioning (leak test passed fine, no problems).

For reference, I'm pumping 50:50 water:methanol at 1.0 mL/min through an Agilent Zorbax Phenyl-Hexyl column controlled at 30 degrees C, monitoring 214 nm (necessary for two analytes in my analysis that do not have strong UV absorbance, and show virtually none past 225 nm).

So, I'm wondering if anyone has had this issue with the G1312A binary pump system from Agilent? If so, how did they fix it? Premixing mobile phase isn't really a great solution (my usual mobile phase on my primary analysis is 52:48 water:methanol, which is hard to mix reproducibly, IMHO, as I run alomst constantly and would be mixing mobile phase over and over and over....). Also, the other two binary pumps I use every day have no mixing issues using the same water and methanol at the same wavelengths. Any help is appreciated!

[lmh]

The behaviour you describe is absolutely not typical of a G1312A pump. They should mix water and methanol without any increase in baseline noise. This is also a high-pressure mixing binary pump, so there isn't really much of a mixing system to talk about; it's just a combination of two flows. Have you looked at the pressure, to check you don't have any saw-tooth patterns? Slight air bubbles are really bad news in this pump.

The only other thing I can think of is that if your pump's inlet valves are on the edge of failure, they may not work as well with pure methanol as 50% methanol (different viscosity). I wouldn't recommend just changing bits of pump at random as this is an expensive way to fix things.

Given that this is a good quality pump, there is a case for getting in serviced by Agilent. Otherwise, instead of just doing the system diagnostic tests, you could try doing a proper gradient test, in case that shows up some problem.

[DR]

I disagree with the above. The passive mixer supplied by Agilent with these is not very good (just a small column jacket full of beads). With good degassing of the constituents, and the addition of an active mixer, you can get a better baseline.

[HPLCCONSULT]

You bought a system at auction so the very first thing you should do is pay/hire someone to go over the complete HPLC system to find out what type of condition it is in. *We often hear about someone buying a system for ~$10K at auction and think that they no longer need to spend any more money on it, but have no technical knowledge how to maintain or repair it. My favorite comment from the auction house is "Previously running from a well known lab". Have you seen how people take care of their HPLC systems ?!? If you have not already done so, please have the system checked over 100% to find out what condition it is in and what maintenance and/or parts may be or are still needed plus training (cost $1-5K?). If you lack the training in operating and maintaining the 1100-series system, then please consider taking a class in the same or hiring someone to help you. Education might seem expensive at first, but it can provide you with the knowledge to SAVE tens to hundreds of thousands of dollars in the first year alone through not wasting time and materials on projects.

Things to consider:
(1) Are you degassing the two solvents ? You should be. Binary Pumps do not come with a degasser. These were ordered separately. Did your include one ?
(2) As mentioned, the Binary pump includes a column mixer (crude, but works fine in real life for flow rates of ~ 1 ml/min). If the previous user removed the mixer, then this could explain the problem. The mixer is often removed (or replaced) to fit a particular application such as running at 50 ul/min. It adds considerable delay volume to the system so sometimes is removed by the end user. *An ACTIVE mixer is not needed for this application.
(3) Have you checked the pump settings over ? Is it set for variable volume (auto) stroke or a fixed volume ? Did you clear the settings and load a fresh method? Are both pistons and seals new or in excellent condition? What is the condition of the active inlet valve ? If it uses the replaceable cartridge, has it been replaced ? Are the solvent lines connected correctly ? Did you replace the PTFE frit ?
(4) If you are using the electronic Helium degasser, then determine if it working correctly. Degassing a 50/50 mix of MeOH and Water is important. Did you flush the lines of the degasser from the bottles all the way down to the pump head proportioning valve ? Remember that on the standard degasser the lines and vacuum chamber have a hold up volume of 10-20 mls EACH !

[bisnettrj2]

HPLCCONSULT....

I'm currently running four Agilent HPLCs. I DO know what I'm doing when it comes to servicing and maintaining these instruments. I've run every Agilent Chemstation diagnostic test under the sun, from pump to detector. This is a complete system, so it did come with a degasser. If you'll re-read my original post, it's not like I got the system yesterday and screamed when it didn't work right. This problem is ongoing, and is not related to my degasser or mobile phase, and yes, the column mixer is in-place. I agree with lmh on this one, this is not typical of the G1312 binary pump system. I've had good success with these pumps in the past, so I am confused by this issue. I have considered sending it to Agilent, but I was hoping to get it solved for myself before it came to that (also, I don't have a lot of faith in Agilent's service techs, as my first experience with one took a system down for two months).

lmh - "..you could try doing a proper gradient test..." - what are the instrument conditions required for that?

[Consumer Products Guy]

lmh - "..you could try doing a proper gradient test..." - what are the instrument conditions required for that?

The check out tests are in the pump operating manual. The two pumps themselves are the same as in the Agilent quanternary or isocratic pumps, difference is in inlet solvent switching and the high pressure mixing. We have seven Agilent quat pumps here operating (four 1050 systems, three 1100) but just one binary system here, which works fine.

I'd run the gradient flow test, the step gradient with the acetone in the water. And load up stored bad chromatogram and one from a pre-mixed good run, and overlay the high pressure profile for those runs, it's stored, you just have to "request" such by clicking "More" before loading the data file, and look for pressure fluctuations. I'm assuming from your posts that the retention times are consistent when you use the high-pressure mixing, just more noise, correct? Also, change the solvent inlet frits.

[HPLCCONSULT]

"bisnettrj2" : No disrespect meant in my previous post. You did not include much information as to your experience, if the system had a degasser (as mentioned, they do NOT come with one so it was ordered in addition to the main system) or what you have tested. *Also, the built-in Pump tests are not worth much. Better to use old school tests to verfify function as even features like the flow rate measurement are not accurate with the Agilent pump under certain circumstances.

I see that you state you are running with a column in line, but it is not clear to me IF you have done any testing of the system w/o the column in line (place a restrictor in line) ? This is worthwhile to minimize variables. *Use the pressure monitor trace feature plus the baseline S/N features in the software to plot this with the premixed and online mixed composition so you can quantitate what the actual difference is. If the active inlet valve and pump head are working correctly, then the baseline should be relatively quiet at 214, but some noise would be normal.

Another question: Does your Binary Pump have the optional low pressure solvent selector valve installed ? This valve swaps solvent lines so you can pick either A1/A2 or B1/B2. If you have a leak on one side of the line it can aspirate solvent from the wrong bottle into the mix.

Do you have a spare active inlet valve to swap in ?
Have you checked the one check valve on top of the pump head to insure it is not sticking (very rare, but it could happen)?

[bisnettrj2]

I do have the solvent selection valve installed, and I am running a column in-line. I'll place a restriction capillary in-line in place of the column and repeat my tests. I'll also test it without the solvent selection valve installed. Thanks for the tips.

[tom jupille]

Another (slightly more elaborate) test is to put pure water in the "A" reservoir, and water spiked to give about 1AU of UV signal (e.g., about 0.1% acetone at 265 nm). Then, with the restrictor in place (no column!), run a typical gradient (e.g., 0-100% in 20 minutes at 1 mL/min). Print out the resulting trace and look for fluctuations or deviations from linearity. You can refine the test in "questionable" composition regions by running "stairsteps".

There is a longer description on the LC Resources web site:
http://www.lcresources.com/resources/TSWiz/hs450.htm and
http://www.lcresources.com/resources/TSWiz/hs410.htm

If the results of those tests look OK, then a degassing issue (or poor mixing due to viscosity effects) would be the most likely culprit.

[lmh]

the acetone step-test was exactly what I had in mind. I usually carry out the test using a back-pressure regulator set to provide pressure similar to the lowest I ever see in chromatography, but in general the binary pumps shouldn't be as pressure-dependent as the low-pressure mixing quaternary pumps.
Good luck, pump problems can be a nightmare.

[DR]

While the above is a good idea (make sure all is working properly), you may still find yourself in need of an active mixer to get a quiet baseline under your desired conditions. I'm just speaking from my experience.

Also - not all Agilent degassers are created equal.
Old ones - bullet proof but have a pretty high internal volume, modest degassing capacity.

New ones - much lower internal volume, higher capacity for degassing, very easy to break (by trying to vacuum solvent through them). If you have a newer one, check it thoroughly.

[mbicking]

One more thing.

Operating at 214 nm, you are less than 10 nm above the UV cutoff for methanol. This means your system will be more sensitive to minor fluctuations in methanol content, compared to, say, 220 nm. The question I have is, "how wide are the fluctuations?" Are they 0.01 mAU, or 0.1, or 1 mAU? Have you run this analysis on a different system, with a better baseline result?

IF this is indeed a problem with this unit, I would go with the above suggestions and replace the active inlet cartridge and the outlet check valves. However, I recently encountered similar problems with an RI detector, mixing MeOH/Water using a new 1200 quat pump, so it is clear that this system is a challenge even without pump problems.

[mbicking]

Sorry, one more question:

You said the "leak test" passed fine. Was this the actual Leak Test or the Pressure Test? The Pressure Test checks the ability of the system to hold pressure. The Leak Test will evaluate each piston and its check valves separately, and should be completed first before you start replacing components.

[bisnettrj2]

Of all the stupid things I'm waiting on, I need the adapter for the vacuum degasser tubing to the pump to finish my tests. I tightened every fitting in the binary pump, just for spec, and am eagerly awaiting the adapter (ordering dept decided to order from VWR instead of Agilent, even though VWR doesn't stock the part...).

mbicking: Leak Test, done with IPA, passed fine. I typically don't bother with the pressure test unless I suspect a bad injection valve seal. Since I was worried about the pump, I just checked it.

I believe the baseline fluctuations (not looking at data now, so trying to remember)) are on the order of 0.5 mAU, and random, at 214 nm. Noise specs by Agilent are all at 254, so they are of no use to me. I have two other systems running the same analysis with much better results, both running binary pumps, which is why I'm flummoxed. I do suspect the active inlet valves or cartridges, but would love an explanation for why they are the reason before I start randomly swapping parts.

Tom: That's what I thought lmh meant when he suggested the gradient test. When I get my adapter, I'll effort the test on both channels, starting with A @ 100% H2O and B 0.1% Acetone, and then vice versa.

I really appreciate all the feedback. Unfortunately, I'm self-taught in HPLC by a "cheap" lab and I've had to sink or swim by Agilent manuals and the internet when it comes to figuring our how to make do with the patchwork instrumentation I've beeen given. I think I've done a good job, but I love asking questions and getting a bunch of informative responses in a relatively fast timeframe by people wiser than I. Thanks everybody!

[tom jupille]

Unfortunately, I'm self-taught in HPLC by a "cheap" lab and I've had to sink or swim by Agilent manuals and the internet when it comes to figuring our how to make do with the patchwork instrumentation I've beeen given.

Actually, the "deep end of the pool school of swimming instruction" is a *very* good way to learn!