HPLC Products

When solvents mix they occupy less volume than the sum of their volumes before mixing. In addition, the mixing can result in a temperature change, and also dissolved air may be eliminated.

<---1 litre MeOH + 1 litre water

If 1 litre of methanol is added to 1 litre of water, each measured separately, the final volume is 1945ml at 20oC. A similar but reduced effect is observed when mixing acetonitrile and water.

Measuring is made more complex by the fact that mixing methanol and water is exothermic. A test mixture increased from 16 to 25oC. Acetonitrile and water however is endothermic (it gets cold!) As a consequence the mixing effects are reduced in the case of methanol, because the expansion due to heating compensates for the contraction due to mixing, and are increased in the case of acetonitrile, where both mixing and cooling cause a reduction in volume.

Degassing is totally another issue, but the mixture should be thoroughly shaken, and time allowed for the released air bubbles to escape before attempting to measure the volume.

The effect of all this on HPLC? Retention times change with %B and with flow rate, and both can be affected by solvent mixing. So the method of mixing is important.  Consider an eluent comprising 60:40 methanol:water, and consider the following five possible methods of mixing, and their consequences:

1. You measure solvent B, then add solvent A on top and make up to the line, pumping isocratically. Because of the volume reduction, this will result in an excess of methanol, giving a stronger eluent than intended, and hence shorter retention times than one would expect.  This method is reproducible, but risky because of the possibility of someone adding solvent A first, giving the opposite effect and very different retention times, possibly leading to incorrect peak assignment, or poor integration.

2. Solvents A & B are measured independently and mixed, pumping isocratically. This is by far the best, as not only are the volume effects eliminated, but temperature effects on volume do not effect the volumetric mixing. Unless there is a good reason to do otherwise, this method of mixing is to be recommended.

3. Solvents mixed on-line by a Low Pressure Mixing gradient pump. The proportioning valve draws in slugs of each solvent in the ratio 6:4 and then they enter the pump. Degassing is critical here, because air released by mixing will be trapped inside the pump. The proportional accuracy is down to the pump manufacturer, but if it is good, the retention times should be the same as for situation 2 above.

Please note that unless the pump has a built in mixer, solvent will hit the column as tiny slugs of each solvent, like trucks in a goods train. This does not usually affect retention times, but it usually does give an unacceptably sinusoidal baseline if used with refractive index detection. The solution is to pre-mix using method 2.

4. Solvents mixed on-line by a High Pressure Mixing binary pump system. Here pump A pumps methanol at 0.6 ml/min and pump B pumps water at 0.4ml/min, and they mix after the pump. This is robust and reproducible, and air is less of a problem (until it gets to the flow cell!) but because the mixing occurs after the pumps, there is a volume reduction and the net flow rate is less than 1 ml/min. As a consequence the retention times will be a little slower than with any of the above methods 1-3.

Please note here that a dynamic or static mixer at the T where the two flows meet is essential to prevent laminar flow (the methanol and water flow side by side instead of mixing).

5. Measure out solvent A first and make up to the line with solvent B. As we have established, this results in an excess of water being added, giving a weaker mobile phase than specified in the method, and hence this method gives longer retention times than any of the methods 1-4 above. As with method 1, this is reproducible, but open to the risk that someone will add solvent B first, getting very different results.