Unusual gradient baseline distrubance

[Colette]

Hi Folks,
I am having issues with a new gradient RP-HPLC method, which I have been running now for the past three weeks.
I am getting a symmetrical ripple effect at approx. 33 to 36 minutes in all my blank and sample injections. Please take a look at an example of this interference at the website below.

http://tinypic.com/view/?pic=xgd5ps

The following information details my experimental and operating conditions:
Experimental:
The HPLC operating conditions used were as follows:
Injection volume: 20uL
Column temperature: 25 degC
UV detector wavelength: 225nm
Run time: 55 minutes

Flow rate of 1 mL/min using a gradient program
Time (mins) A % B %
Startup 80 20
1.0 80 20
45.0 20 80
46.0 20 80
47.0 80 20
55.0 80 20

Where A % = Buffer (0.1% phosphoric acid in water)
Where B % = Acetonitrile
Sample diluent: 30/70 (v/v) acetonitrile to purified water

The symmetrical ripple effect occurs at approx. 33 to 36 mins.
The pressure decreases steadily from 90 to 65 bars as the composition of the mobile phases changes from 80/20 (v/v) Buffer/ACN to 20/80 (v/v) Buffer/ACN over a period of 45 minutes.

At 33 mins, where the interference occurs, the composition of the lines A and B are 36% and 64% respectively. The interference ends at approx 36 mins, and the composition of lines A and B are 32% and 68% respectively.

The following was investigated:
• Column changed from zorbax SB C8 to an equivalent Zorbax Rx C8. (different lot number for silica packing)
• Changed the source of milliq water (different millipore systems, taken from different purified water loops)
• Prepared mobile phase A without the phosphoric acid to determine if the phosphoric acid is causing the interference. No improvement was seen.
• Performed 0uL injection of sample diluent to investigate if sample diluent is causing the interference. No improvement made.
• Changed to a second system (from 42 to 47). No improvement.
• Could the interference be air? Degassed with helium for 20 minutes but made no difference to chromatography.
• Premixed the mobile phases A and B to give compositions 80/20 (v/v) buffer to acetonitrile and 20/80 (v/v) buffer to acetonitrile respectively. Initial results saw an improvement in the chromatography but subsequent mobile phase preparations and system setups, the interference reoccurred.
• System washed with hot water to remove any traces of previous buffers.
• System needle was washed with 100% acetonitrile by performing 25 injections at 100uL. No improvement.
• Altered the gradient table, by holding the gradient at 33 mins at composition 36/64 (v/v) buffer to acetonitrile for an extra ten minutes to see the effect on the interference. The interference had the same profile but eluted over a longer period.
• Changed the needle and needle seat No difference made to chromatography.
• Bypassed the degasser and the interference disappeared but background noise increased after several injections. I attempted to repeat this scenario on a second system but failed. Bypassing the degasser on the second system, did not get rid of the interference.

If anyone can suggest what is causing this interference I would be most grateful to you.
Any information, tips or suggestions would be received with many thanks.
Regards,
Colette[/url]

[Bryan Evans]

Did you try running the gradient without the column?

[DR]

Does the total area for these peaks increase if the amount of time pumping your initial conditions increases? To check, let it run at your initial conditions for 10-15', then do two identical blank injections back-to-back. If the peaks are quite a bit larger in the first of the two injections, you need to pretreat the water you're making your A phase with (search board for the word Empore).

[Uwe Neue]

Looks like an oligomer separation that you can find in techniques designed to do this stuff. Unfortunately, at your wavelength, one will see a lot of things. Actually, some of the people that are doing this type of separations would be very proud to get such a nice pattern over at least 13 oligos.
Check for sources in your plastics that are in contact with the mobile phase, or the soaps that you are using to clean your classware, or related things...

[Colette]

Thank you all for your suggestions and ideas. I have tried your suggestions and these are my findings.

Bryan,
I did try running the gradient without a column and yes the interference disappeared. However, I'm not sure what this shows me. It is retained on the column?

DR,
I did try what you suggested and yes the total area count of the interference does increase when I hold the initial conditions. To demonstrate this I held my initial conditions for two hours, then ran two back to back blank injections. Please refer to the image below. THe total area count for the first injection was calculated as 768,666 vs. 154, 922 for the second. I will perform this a second time to confirm that this result wasn't a fluke. Can you suggest a suitable filter to use to pretreat the water? I've discussed this with a supplier for Whatman but he is unable to advice something for me.
http://tinypic.com/view/?pic=xlxnwk


Uwe,
I prepared my mobile phase in new glassware. As a result, the mobile phase has not been in contact with any detergents or plastics. I am still getting this interference despite the new glassware.

[Uwe Neue]

If DR's experiment repeats, we are getting closer. The stuff could still be in the acetonitrile or the methanol or the water or the instrument.

[DR]

I suspect you need to replace your water system's styrene based resin beds. I suspect resin monomers, dimers... are getting into your MP prep. In the mean time, get some Empore extraction disks (Varian). Rinse w/ MeOH and flush w/ water, then pull 1L water through, repeat rinsees, pull another 1L water through... until you have enough water to make a batch of your A phase. This should reduce the peaks to almost nothing.

If you have particularly large injection volumes (>100µL), you may want to consider pretreating the water for your sample diluent prep. too.

[Colette]

Just a quick note to let you know we finally got to the root cause of the problem.
Having derived to the conclusion that the interference was due to the mobile phase components (DR's experiment of holding initial conditions for longer and running two back to back blank injections etc), I decided to change from gradient HPLC acetonitrile to FAR UV.
Guess what? The interference went. I reverted back to the gradient HPLC grade ACN and the interference reappeared. This was repeated twice to ensure that this wasn't a once off occurance. I then began to investigate different lots of gradient HPLC grade ACN. The results showed inconsistency in the purity of the acetonitrile. Some lots did not give any interference while other lots did.
I decided to conclude my investigation by gathering some UV spectra of the various lots of acetonitrile (both gradient and FAR UV) over the range 400-190 nm. As you may recall, my HPLC wavelength is 225 nm. Take a look at the spectra gathered and the absorbance values measured at this wavelength (ref to link)

http://tinypic.com/view/?pic=11b1buv

Two lots of acetonitrile gave the most HPLC interference. From the UV spectra gathered, these lots have the highest absorbance values, which equates to approximately 61 and 74 % transmittance. This does not meet the vendors specification of greater than 99% transmittance. I am questioning the vendor's procedure for testing the quality of their ACN.?

Thank you all for your helpful suggestions!

[tom jupille]

Colette, many thanks for posting the solution to the problem!

[DR]

I'm glad you found out where it was from. 4/7 lots failing - that's impressive (but not in a good way). I suspect that the vendor owes you some freebies.