WE've had terrible baseline noise problems transferring a customers RP method (for a Peptide)from a Gilson/Anachem LC (High Pressure pumping, specific motorised mixing module)to our HP1100 Quaternary Pump LC (low pressure pumping, no specific mixing module). Phase A is 0.1% TFA in Water, Phase B is 0.1% TFA in Acetonitrile. The gradient is a typically 15 - 50 %B over 30 mins.

In investigating the problem, I know the TFA is the problem because when I pump the same gradient (or any gradient) without TFA in Phase A or B (water & just MeCN), the baseline noise is fine and acceptable. I've spoken to a few service engineers & they confirm the problems of TFA Low pressure mixing and advise making Phase A a bit more Organic or B, a bit less organic (polar non polar I get confused). The problem is this is a validated method so I can't go changing the parameters too much. I've already increased the wavelength from 204 to 220nm (204nm ? mmmmm ?), improving baseline noise with a very bearable reduction in Target signal. We've also istalled an inline Static Mixer after the purge valve (from Suppelco) which basically has a torturous flow path. Things have improved, but I obviousley cannot report low level impurities with any degree of confidence due to the influence of the baseline noise on their reproducibilty of response.

I'm no Chemist (a biochemist) but I guess what's happening TFA likes water and MeCN to varying amounts and so jumps between the 2 - a bit of partioning so to speak eventually picked up by the detector. A good mix would probably help install a better equilibrium. Can someone please expand on what cause this phenomenom and why High Pressure systems lead to better baselines with TFA Buffers than Low Pressures systems do ? (unless the problem's to due with different detector's sensitivity ???) Possibly with a bit of GCSE Chemistry explanation on TFA's behaviour in Water and MeCN thrown in - please don't go too Chemistry on me with big equations and constants and stuff. Another vested interest is I'm friends with the analyst who developed the method and whenever I try to explain the problem, the defensive shutters come up - naturally. So I would like to expand his and some of my colleagues knowledge on this situation.

Many thanks whoever replies.

I had the exact experiance about three years ago. Same mobile phase and gradient. The final solution was to install a Dynamic Mixer (esentially a tiny chamber with a stirr bar in it) on our HPLC. We tried the static mixers, pulse dampners, back pressure regulators etc. Eventually, we found the static mixer was the only way to go. Our method was developed on a older HPLC system with a large dwell volume. We were transferring the method onto a new system with very little dead volume. Some how, the larger dwell volume allowed the mobile phase to mix better.
I don't remember the vendor for the dynamic mixer, but it was our only solution.

This is not a high pressure system /low pressure system issue. This is a mixing issue, pure and simple. Simply what happens: the abosorbance of TFA in water and TFA in ACN is slighty different. As you form the gradient, you get various mixtures of TFA/Water/ACN, all having different absorbances (this is what gives the baseline noise of which you speak, commonly called compositional ripple). The better the mixing in your system, the less ripple you will see. You have already added a mixer (what volume?) but you may need more. Of course, adding mixing will increase the system volume, making gradient equlilibration times longer. If you really think that high pressue mixing will work better, try the experiment on a 1100 Binary (high pressure mixing) system, it generally produces higher ripple that the quart.

As you can see, the problem is known. You need to improve/increase the mixing. Can you steel the "motorized mixing module" from the Gilson or buy a thing like this from somebody?

Also, try 214 nm. The absorbance spectrum of TFA changes as a function of ACN concentration (that's why you see the noise even with TFA added to both A and B). The spectrum has an "isosbestic point" (a wavelength at which the absorbance stays constant even as the shape of the spectrum changes) at 214nm.

Buffer A should be 15%ACN in 0.1% TFA, buffer B 50%ACN in 0.1% TFA.
You will find this will mix much better!

If instead of just having
A= H2O + 0.1% TFA and
B= ACN + 0.1% TFA,
you used (like Anonymous suggests above)
A= 85:15 H2O/ACN + 0.1% TFA and
B= 50:50 H2O/ACN + 0.1% TFA
with a 0-100 %B gradient over 30 mins

would this make a significant difference to the case above?.. and what are the general benefits of applying premixed eluents in a low-pressure mixing system? Personally I've never had the time away from QC testing [to documented methods] to experiment in this way :-(

It should make a significant difference. Suppose, for the sake of argument, that the system can pump and mix with a relative precision of 0.25%. The "absolute" precision of an A to B gradient is the relative precision times the composition difference between the A and B solvents.

In other words, that 0.25% relative precision equates to an absolute precision of 0.25% if A=water and B=organic, but only 0.09% if A=85% water and B=15% water.

Baseline noise and retention effects depend on the absolute variation in composition.

Thanks Anon, so using these premixed modified gradients should lead to a more precise and reproducible eluent composition gradient formed from pumps A & B, thus giving greater consistency in retention times and less baseline noise.

Would another variable such as temperature be more significant in affecting RT's and baseline stability/noise compared to pump/mixing precisions?

Lastly, from the point of view that preparing the 2 premixed mobile phases above is no more difficult (or laborious) than simply having H2O/TFA and ACN/TFA, is it worth premixing them for the benefits and is it already commonly done?

Views appreciated.

1. Yes, for the reasons mentioned above.

2. Temperature can certainly have an effect. Whether it is more/less significant depends on how big a temp change and on the design of your hardware.

3. I strongly recommend pre-mixing the mobile phases as discussed above.

I did a quick search through the "LC Troubleshooting Bible" CD-ROM compilation of all the John Dolan "LC Troubleshooting" columns in LC-GC magazine. At least a dozen of the articles deal with various causes and cures of gradient baseline problems. (I'll throw in a shameless plug here: the CD is available from LC Resources:
http://www.lcresources.com/resources/restscd.html