Noise and Drift

[alireza.shirkhany]

Dear All,

I have a real problem in obviating noises and drifts I see in the baseline chromatogram and even when no injection is running and lines are simply getting filled with solvents (No column, Union Attached).

I'm working with an Agilent 1100 equipped with a DAD detector. I have checked following:
Pump is working fine, the flow pressure when a phosphoric buffer is running off the column is at most at 100 bar. When purge is open, the pressure wouldn't do upper than 10bar. Also, with union attached and purge closed I haven't observed pressure more than 30 bar. Flow of the pump is true (I checked from 0.5mL/min to 5mL/min and each time the flow was correct)
PTFE filters behind the pump is tidy and unsaturated. Also, active inlet is fine.
The column is Siliachrome from Silicycle (25cm, 6 micron, 4.6mm) and the mobile phase has buffer in itself. I see too much noises when it is attached or when its not (no difference).
I have done self test in the instrument, holmium test and other tests separately and they have passed successfully.
Plus, I have passed following solvents in three days, each for 12 hours with no success (because I think flow cell must be contaminated which should be cleared with the following solvents but nothing good happened unfortunately).

- Isopropyl alcohol
- Acetonitrile
- Water
- Methanol

Plus, I increased the Peakwidth and Slit from 2s & 4nm to 16s & 16nm (in this case, noises were smoothed but I know it is still there. Also, the negative drift decreased from +-5mAU to +-1mAU).

The last try was to check with different wavelengths as follows (For Peak-width of 2s and Slit of 4nm):

210nm, BW: 16nm- Ref 360nm, BW: 100nm. Result: Negative drift +-5mAU, Noise of 5mAU
230nm, BW: 16nm- Ref 360nm, BW: 100nm. Result: Negative drift +-3mAU, Noise of 5mAU
260nm, BW: 16nm- Ref 360nm, BW: 100nm. Result: Negative drift +-3mAU, Noise of 3mAU
286nm, BW: 16nm- Ref 360nm, BW: 100nm. Result: Negative drift +-1mAU, Noise of 2mAU

Is it acceptable to have a 5mAU of noises and Negative drift of +-5mAU at 210nm when Peakwidth and Slit are 2s and 4nm, respectively? If not, what should I do more to decrese noises and drifts?

(Lab temperature alternates but it is mostly at 22 degree celcius. Plus, instruments' electrical plugs are attached to an adapter which compensates for alterations of electrical voltage. For the last thing, detector lamp has worked for roughly 1800 hours)

Thank you all for your helps.

[vmu]

More info is needed.
- The way of determining (calculating) the noise.
- "Negative drift of +-5mAU" per 1 min? per 1 h? per 10 h?
- Composition of the mobile phase; pH.
- Purity grade of reagents.
- The time for the column equilibration.
- Is flow cell test OK?
- Is the column a C18 column or a bare silica column?
- Use of the column thermostat and its temperature.

Generally, 5 mAU is a high noise, but it may be acceptable for some tasks.

[alireza.shirkhany]

Hello,

Sorry for the insufficient information.

I didn't calculate the noise, but when the peak height is 5.869, changes in noise is +- 5mAU.
Negative drift of +-5mAU per 25 minutes which is the run time of the method.
Mobile phase composition is as below:
A: Water+ Sodium Dihyrdogen Phosphate+ ortho-Phosphoric acid. pH=3.00
B: Acetonitrile+ A
Flow rate: 1mL/min, Gradient
(Consider that noise and drift are as above whether the method is running or it is not running)

Purity of reagents: Water is HPLC grade and double filtered with Cellulose Acetate 0.45 micron filter. Acetonitrile is HPLC grade as well. Sodium Dihydrogen Phosphate and ortho-phosphoric acid are reagent grade.

I did the flow cell test from the software (Chemstation) of the instrument and it successfully passed.
The column is C18.

No column temperature is controlled and it is at the room temperature.

[vmu]

A peak with the height of 5.869 mAU cannot be detected when the noise is of a similar magnitude (5 mAU). You confuse something here. Is the noise a high-frequency noise or is it just relatively long periodic (or non-periodic) waves on the baseline? What is the amplitude of the pressure pulsations?

Some baseline drift under gradient elution conditions generally is a normal situation caused by the change of the mobile phase composition. However, the baseline should remain stable when you pump the mobile phase of a constant composition through the column for a long time sufficient for equilibration.

What about working air conditioners or anything else blowing near the instrument?

[alireza.shirkhany]

There is a fan in a distance of 4 meters away of the instrument.
The composition is constant and it is only under the gradient form (No mobile phase percentage goes to zero).
I checked the noise again for the peak height of 5.869 and I found it is +-1mAU and sometimes +-2mAU.
The pressure pulsation is 5 bar at most.

[vmu]

Working fan in the room can cause baseline drift.
The noise cannot be "checked for the peak" since the baseline noise is a feature of the baseline, not that of the peak. For a particular peak you can only calculate the signal-to-noise ratio (i.e. the peak height divided by the noise magnitude).

For your instrument, typical pressure variations are about 0.4-0.5 bar at 100 bar pressure (according to the pump manual, maximum pulsations must be not higher than 1% of the pressure). Pulsations of 5 bar are too high.

[Multidimensional]

"Noise" and "Drift" are two different things which may also have two different causes. You need the help of an experienced chromatographer (local) to assist you troubleshoot the issues you are having. The changes you have made sound almost random so I think you would benefit from some local help. Perhaps there is someone at your school/company who can help you?

A few comments:

Running an HPLC method without a column oven, needed to control the temperature, may induce baseline drift (not noise). Some methods are very sensitive to small temperature changes (~ 0.5 C), some are not, but reproducibility is always impacted by temperature instability, so if possible, control it.

Yes, 5 bars of noise relative to a 100 bar average back pressure is way too high and indicates a problem with setup (e.g. inadequate degassing, poor priming, air bubbles, dirty PTFE, a leak, contamination, dirty solvent pickup filter(s)...). Pressure ripple of ~ 0.1 to 1% is more typical with any Agilent HPLC pump.

Running at such low UV wavelengths requires high purity buffer chemicals/additives too (reagent grade may not be high enough).

Running anything at 210 nm is likely to result in higher background noise and drift. This is to be expected and normal, if the DAD settings are appropriate and if the use of high purity mobile phase is used.

You have incorrectly setup your DAD with a bandwidth of 16 nm for all signals. When you set the bandwidth at 16nm for a signal at 210nm, then you are actually monitoring an averaged signal from 202 nm to 218 nm. This will pick up so much noise! A better bandwidth to use at 210 nm is 4 nm (208-212 nm). You also have incorrectly turned ON the special software feature "Reference Wavelength" at 360, 100 which collect all data between 310nm and 410nm, subtract it from your original data (i.e. 210nm), destroy the original data, then make a new signal with just the result and call that your new 210nm signal. For almost all sample types, never turn that feature ON. Please turn it OFF for all signals as default. Besides invalidating ALL data collected (because it destroys the original data and replaces it with new data leaving no record behind) it may hide any real peaks that show up in the subtracted region.

You mentioned your D2 lamp has > 1800 hours on it. Most lamps are good for ~ 1000 hours. Some are good to ~ 2000 hours. Proper testing must be done to determine if yours is within specification (esp at LOW UV LEVELS around 210nm which is the first area they fail at so your lamp may be bad). *Run the built-in lamp test with the flow cell filled with your mobile phase to get an accurate idea of what the output is.

I have included a few article links in basic troubleshooting to assist you too,

"Common Causes of Baseline Noise in HPLC, UHPLC"; https://hplctips.blogspot.com/2014/09/c ... noise.html

"HPLC Retention Time Drift, Change, Variability or Poor Reproducibility. Common Reasons for it"; https://hplctips.blogspot.com/2015/11/h ... hange.html

"Proper Wavelength Selection for HPLC Method Development (or Purity Determination)"; https://hplctips.blogspot.com/2013/08/w ... ethod.html

[alireza.shirkhany]

Thank you for your answers. I will do all your mentioned guides to see what will happen and will get back to you as soon as possible.

[lmh]

Try putting something in the line to add pressure (either a back-pressure regulator, a known clean column, or a very long bit of thin tubing). 30bar isn't a high pressure. Agilent 1100 pumps are fairly good, but even so, this is low. HPLC pumps aren't always guaranteed to pump the right mix of solvents unless they're pumping against a sufficiently high pressure, and when one of your solvents is UV absorbent (at least at some of the very short wavelengths you mentioned), while the other isn't, you could get strange UV effects.

It can also help to run the test just with water, or to use the same bottle of solvent for both A and B, so you can check whether you've got gradient issues, pump pressure issues, or PDA issues. Also, of course, if it's pump related then any periodic noise should vary with flow-rate and % B.