Linear Gradient Accuracy Check

[chetan_m43]

I am working on large scale HPLC column, we are having problem with Elution. The Elution should happen at linear gradient step, however it happen much sooner at step gradient itself. We have crosscheck the buffers composition, buffer found to be OK. The problem seems to be with pumps, its not delivery accurate or precise gradient. So anyone have any idea, how I can check that pumps are delivering accurate gradient.

[Gerhard Kratz]

A flow meter at the outlet can do it, or a stop watch and a graduated cylinder.
Do you operate at elevated temperature? If so please make sure that your eluent has the same temperature than the column before it goes on the column.
Good luck

[DR]

While Gerhard's method will check flow accuracy, gradient formation accuracy is another matter.

What we do is pump water in one line and water with a little acetone in the other.
We program a linear gradient 0-100% B in 10 minutes, then do a linear regression on the slice areas of the linear potion of the plot using the raw areas of each slice vs. time.

So, to do this, you need to know how to export your raw data file into Excel, you need to know your sampling rate so you get your times right and you need to know how to do a linear regression.

You can also use this plot to calculate your dwell volume based on your flow rate and the difference between your actual midpoint and time where the midpoint should be.

[carlo.annaratone]

Assuming you are using a UV detector, all you have to do is to load on line A water and line B water + optically active additive. Someone uses acetone, but I found that evaporation will invalidate results, better 10 mg/l of a paraben as Waters suggests. Then replace the column by a capillary restrictor that gives an appropriate backpressure in the flow range selected, and check the gradient on the UV.

[ilee]

... Then replace the column by a capillary restrictor that gives an appropriate backpressure in the flow range selected, and check the gradient on the UV.

Perhaps a stupid question: Do you really need an "appropriate backpressure"? Does the linear gradient accuracy change at different backpressures?

[lmh]

depends on your pump. I haven't tested extensively with a lot of different systems.

Some years ago I had a low-pressure quaternary system from Thermo where the gradient proportioning system was physically mounted on the bottom of the chambers where the pistons moved. It pumped very inaccurate gradients if used with no back-pressure. I believe what was happening was that it relies on having reasonably high back-pressure to close the check-valves between the gradient proportioning chamber and the first piston chamber, and between the two piston chambers. If these valves aren't properly closed, then solvent can diffuse back from the piston chambers into the proportioning chamber, and quite whether the proportioning chamber takes up fresh solvent from appropriate proportioning valves, or mixed solvent from the pump, is rather random. With decent back-pressure (I used to use a regulator set to about 40-50 bar) it pumped very nice gradients and passed its tests fine.

At the same time, I had an Agilent 1100 binary system mixing at high-pressure, which pumped gradients very accurately whatever the pressure.

Since you really want to know whether the system is pumping accurately under working conditions, I would recommend using a pressure as close to your working pressure as you can. This may be impractical (you don't want a big dead-volume because it makes the gradient test very long by introducing a long delay into any change of solvent composition) but the only item you will have in the lab capable of producing a 300-bar back-pressure is probably your column! At least give it some back pressure if possible.

[ilee]

This is plausible. Thanks for your explanation !