Random Baseline Noise in Gradient Elution

[asicley]

Hello,
We are having a problem during a gradient injection on Waters 2695 with 2487 UV detector. During the gradient elution we are observing signifigant noise, in the form of negative peaks. The buffer is 0.01M Phospate Buffer, the organic modifier is ACN.
MPA is 100% Buffer, MPB is 25:75 Buffer:ACN.

The gradient is:

0 min A=50% B=50%
25 min A=45% B=55%
30 min A=40% B=60%
70 min A=28% B=72%
75 min A=28% B=72%
85 min A=50% B=50%
90 min A=50% B=50%

The peaks are random and appear starting as early as 30minutes, and will sometime continue to 80 minutes. The detector was rinsed with 50:50 ACN:water in reverse flow, and we moved to another instrument, and the problem was resolved for a few days, however we are now seeing the problem on the new system. We have run this method on multiple systems, and have not seen the problem on all systems. We are trying to understand the problem, so that we can control it.
Thank you very much for your input.

Thanks, Allen

[DR]

New batches of mobile phase every time you move to another instrument?

[asicley]

The same MP, column, and samples have been moved to other systems, and the problem has not been seen on the new system. The problem did occur on one of the other sytems, after a couple of days of running. The problem does not seem to be mobile phase or column specific.

[HW Mueller]

Maybe a degassing problem?

[tom jupille]

I'd say that HW's suggestion about degassing seems the most likely at this point, but a few other suggestions/possibilities:

1. Does the problem only occur when you have injected samples or standards, or does it also occur with a "dummy" (no injection) gradient?

2. Once it starts to occur, does the size of the peaks change as a function of the pre-gradient equilibration time?

Both of those questions address the possibility that there is some garbage building up on the column; I'd expect the result to be positive peaks rather than negative, but it might be worth checking

3. Is the size (or direction!) of the problem peaks wavelength dependent?

If they are more pronounced at longer wavelength, that might suggest a temperature problem, especially if the instrument is near a heater/air conditioner vent.

[Mark Tracy]

If it is a problem of outgassing in the detector, a backpressure device should suppress it. 50-100 psi at the outlet of the detector will keep bubbles in solution.

[asicley]

Hi everyone,
We are presently looking into degassing as the source of the baseline noise in this gradient. The baseline noise is not in every injection, so we are placing a pressure regulator after the detector, to increase the pressure to the mentioned 50-100 psi in the detector. We will inject multiple samples overnight, and see if the problem persists.
Thank you very much for your input on this issue, I look forward to replying with positive results in the morning!
Thanks,
Allen