[fiftycal2004]

I am currently working with a Hewlett-Packard Series 1100 HPLC and trying to get it running properly. The method I have been using starts at 95:5 water:methanol and goes to 100% methanol over 5 minutes at the rate of 1ml/min. During this time, the pressure holds at approx 70 bar. However, once the method gets between 30-50% methanol, the pressure slowly rises to approx 100 bar and then spikes above 398 bar, shutting the system down (I have the max pressure set at 398 bar and the column is rated for 400 bar.) This is a very consistent result. I was wondering if anyone might have any information to help me.

The column I am using is an Atlantis dC18 5um 4.6x250mm purchased in January 2010. It has not been used since May 2010 and I do not believe it was used very much prior to then either. I am currently just running the solvents. Eventually, I will be using the method to separate out atrazine and its metabolites. This method was previously used 4-5 years ago on the same HPLC using an Adsorbosphere C18 5um 4.6x250mm column. Any help provided would be greatly appreciated.

p.s. There is more to the method, however, the 95:5 water:methanol to 100% methanol is the first part of the method and I cannot get past it.

[JGK]

what happens if you try to run the gradient without the column in place? If you get a similar effect, I would start checking the pipework between the column and the pump.

Running a methanol/water gradient with an increasing methanol concentration, I would expect a pressure drop rather than an increase.

For a 250 mm C18 column 70 -100 bar (100-1400 psi) seems about the correct pressure, 398 bar (5700 psi) is definitely not ususal and would imply a serious blockage in the system somewhere.

[fiftycal2004]

I tried to narrow down the problem by disconnecting different parts of the system. Here is what happened:

1. I disconnected the column at the back end, removing the detector from the system. Thus the flow was pump, injector, guard column, column, waste. The results were the same as before, pressure held steady at 60 bar, followed by a slow rise to 100 bar and a spike above the pressure limit in the 30% to 50% methanol range.

2. I disconnected the column at the front end, removing the column and detector from the system. Thus the flow was pump, injector, guard column, waste. This also produced a similar result. The pressure held steady at 30 bar, followed by a slow rise to 100 bar and a spike above the max pressure limit in the 30% to 50% methanol range.

3. Next, I disconnected the guard column. Thus the flow was pump, injector, waste. No spike in pressure occurred. The pressure held around 6 bar and I was able to run through the entire method with no problems.

4. Finally, I reconnected just the column. The flow in this test was pump, injector, column, waste. In this test, the pressure held at about 43 bar for approx 6 min before slowly rising to 100 bar and then spiking above the max pressure limit. To confirm this wasn't a one time occurrence, I ran the test 2 more times. Every time the spike would occur between 6.8-7.2 minutes or 89-91% methanol.

[JGK]

1. I disconnected the column at the back end, removing the detector from the system. Thus the flow was pump, injector, guard column, column, waste. The results were the same as before, pressure held steady at 60 bar, followed by a slow rise to 100 bar and a spike above the pressure limit in the 30% to 50% methanol range.

This will tell you virtually nothing as the pressure in HPLC is generated in the area between the pump and the column. The solvent pressure post column is negligible (pressures inexcess of 100 psi can crack detector flow cells)

2. I disconnected the column at the front end, removing the column and detector from the system. Thus the flow was pump, injector, guard column, waste. This also produced a similar result. The pressure held steady at 30 bar, followed by a slow rise to 100 bar and a spike above the max pressure limit in the 30% to 50% methanol range.

Was this repeatable? Were you using blank solvent injections or actual sample injections?

3. Next, I disconnected the guard column. Thus the flow was pump, injector, waste. No spike in pressure occurred. The pressure held around 6 bar and I was able to run through the entire method with no problems.

OK

4. Finally, I reconnected just the column. The flow in this test was pump, injector, column, waste. In this test, the pressure held at about 43 bar for approx 6 min before slowly rising to 100 bar and then spiking above the max pressure limit. To confirm this wasn't a one time occurrence, I ran the test 2 more times. Every time the spike would occur between 6.8-7.2 minutes or 89-91% methanol.

Was this repeatable? Were you using blank solvent injections or actual sample injections?


If the exercises performed were done using actual samples for the injections, I suspect that you have a precipitation issue. Something within the sample matrix is precipitating within either the guard or analytical column at high methanol concentrations. This precipitation is causing your blockage and pressure spike.

[fiftycal2004]

Everything tested so far has been repeatable multiple times on the same column and has been repeated using another C18 5um 4.6x250mm column from a different manufacturer and has been tested using only solvent, no injections have been made.

New information:
There was some concern that the water I was using was not entirely pure. I obtained water that was known to be HPLC grade and this did not solve the issue.

I watched the tubing for bubble formation. I did notice bubble formation throughout the gradient. The largest and most numerous bubble formation seemed to occur between 30-50% methanol. There were bubbles occuring between 82-85% methanol, but they were not bigger or more numerous than the previous bubbles, thus I am not convinced that they are the problem. (If you remember, the method would exceed max pressure at 90-94% methanol).

I then swapped the methanol and water bottles per a colleague's suggestion (I am 100% confident that the bottles were in the right position to begin with, he just wanted to see what would happen). Because the pressure previously was exceeded at 90-94% methanol, I expected the pressure to be exceeded right away. However, the max pressure was not exceeded until after 8 minutes, when 100% water was in the column.

I ran 100% methanol through the column and the pressure held at approx. 60 bar. I also ran 100% water through the column and the pressure held at approx. 105 bar.

I'm not sure if any of this is useful to you. Thanks for you help so far!

[JGK]

Bubble formation will ruin your chromatography quality. If you are using a low pressure mixing valve for your gradient you should ensure that the mobile phases are thoroughly degassed before use and, if possible, have a heliun sparge running while in use.

Does your method use constant flow throughout? or is the flow rate also changing during the run?

The only reason I can think for exceeding a pressure limit in a gradient where methanol is increasing and water decreasing, other than something precipitating in the column (see previous post), is if the flow is increased and the pressure limit exceeded.

Check your pump flow control file for the method. Also check the actual flow output during the chromatogram to make sure what you program into the instrument is actually what is happening.

[fiftycal2004]

This method uses a flow rate of 1ml/min throughout the entire run. I disconnected the column and ran the method. The entire method is 14.5 minutes long, but only 11.6 mls of solvents were collected. I attribute this loss of volume to the bubble formed in the mixing of the methonal and water, but I could be wrong. To check that the pump was working correctly, I ran 100% acetonitrile through the pump (with the column still disconnected) for 10 minutes and 10 mls of solvent was collected. So it appears that I am back at square one. Thanks for your help so far

[JGK]

OK, so far we have conluded that:

A pressure spike only occurs when there is a column (either guard, Analytical or Guard + Analytical) present.  You are using a methanol water gradient, starting at High Aqueous and increasing the methanol component (pressure limit exceeded at ~90% MeOH)

Flow is constant.

Your mobile phases are not sufficiently degassed (this is causing your bubble issue).

To me this literally shouts out a preciptation issue as, in a methanol water gradient system, pressure should descrease as MeOH increases.

What type (what is the composition ) of the samples are you are injecting?

Something in your sample injections is dropping out of solution as the % methanol rises, which is causing the pressure spike. Resetting the system with the initial mobile phase conditions redissolves this material during the column re-equilibration only for the same thing to happen again in the next attempt.

[fiftycal2004]

See, that is the problem, I am not running any samples. I am only running the solvents water and methanol (and eventually acetonitrile if I can ever get past the water:methanol gradient). So I don't know what would be in the methanol or water that would be precipitating out. 

I think the degassing issue is with the water. I can degass it under vacuum sonication for an hour, but if I swirl it, there are still some bubbles the appear. I've tried swirling it every 10 minutes or so but I guess that is not working. Do you have any tips on degassing the water? Thanks.

[JGK]

If you are not injecting anything and getting pressure spikes, I would remove all of the check valves in the pump and give them a thorough clean. It could be that working against a column pressure they are "sticking".

As for degassing, the only suggestion would be to set up a helium sparge on your mobile phase reservoirs.

Also do you actually need to go to 100% methanol in your method. When was at Uni studying analytical methodology it was recommended that we premix solvents to minimize outgassing when the gradient was being "mixed" by the HPLC instrument. For instance if you had the following:

Mobile Phase A: Water: Methanol (95:5) and Mobile Phase B: Water: Methanol (5:95). Then run the gradient from 100% A to 100% B over a similar time period to the original method it might reduce your bubble issue. It only gets you to 95% methanol at the end but it may improve things.

[Train]

I just came across the following as I was browsing the Acquity UPLC H-Class Sytem Guide from Waters:

. . . with gradient chromatography, the viscosity changes that occur as the solvents are mixed in different proportions can effect pressure changes during the run.  For example, a 1:1 water/methanol mixture produces twice the pressure of either water or methanol alone.

I wonder if it's an increase in viscosity created just by the mixing process or if a premixed solution would also give you higher pressure?

[fiftycal2004]

I tried mixing up two solvent phases as suggested (95:5 water:methanol and 5:95 water:methanol). It seemed to reduce the number of bubbles observed exiting the pump, but I still had the high pressure method. It is possible that the changing viscosity is the problem. I think now we are going to try changing the method. If we get satisfactory separation of the standards of compounds we are looking for, we can avoid using this method all together. Thanks again for all your help.

[fiftycal2004]

To bring some closure to this thread....

It turns out that the problem was my own stupidity.   There are two places in method program where I can set the max pressure. So even though I changed the max pressure in one spot, I forgot/didn't realize that there was another section that needed to be changed as well. After upping the max pressure, the method seems to work and does not exceed the max pressure anymore. Thanks again for all your help.

[JGK]

Aha! the most common cause of problems "the human factor"