Sample Diluent Effects in HPLC

Most of us will know that the solvent (diluent) used to prepare HPLC samples can have an effect on HPLC peak shapes. The following discussion highlights some facts, figures, tips and tricks that can help in a practical situation.

Ask yourself this…..’why do we need to have a buffer in HPLC mobile phase systems?’. Well, buffers, help to resist small changes in eluent pH which may in turn effect the relative extent of ionisation of our analytes and hence their relative elution times – potentially affecting both resolution and selectivity of a separation. Then ask yourself……’why would the mobile phase eluent pH change’? Are there lab elves (we’ll call them colleagues here…..) who come to adjust the pH at night?  Clearly not. Apart from the ingress of CO₂ to the eluent reservoir which causes a gradual (but small) lowering of pH on standing, the only place within the HPLC system where pH may change dramatically (and hence require a buffer) is at the head of the HPLC column where the sample diluent and eluent mix. In this microcosmic situation, the effect of a good buffer can be invaluable in maintaining the quality and reproducibility of a separation.

So – this region at the head of the analytical column is an important region for our separation as the sample diluent and eluent are truly ‘mixed’ for the first time. This may seem anti-sense as the plug of sample is introduced into the eluent flow within the autosampler injection, however the sample is introduced as just that, a plug, which may mix partially at the head and tail of the plug, but there will be a substantial proportion of the analyte molecule immersed in 100% diluent at it arrives at the head of the column (analyte in the ‘green’ band of figure 1)

Figure 1 Sample plug flow through connective tubing. Note that some analyte molecules (red spheres)
remain totally in the ‘green’ band signifying no mixing with the eluent has taken place.
Blue- eluent / Green – sample diluent solution

As the sample plug reaches the head of the column, the plug is spread out over the entire column diameter using a device called a spreader.  As the diluent and eluent mix at the column head, the analyte dissolved in the sample diluent may behave differently to that already associated with the eluent system.

If the diluent is more highly eluotropic (contains a higher percentage of organic modifier, different buffer strength, different solvents, additives etc.) then the analytes will behave differently as they enter the column.  Figure 2 demonstrates what happens the when the sample diluent is more highly eluting (contains more organic for reversed phase separations) than the eluent.  One can easily see that some of the analyte molecules elute through the packed bed until the sample solvent is diluted out by the eluent, whereupon the analytes partition more like their counterparts who were associated with the eluent rather than the diluent when they entered the column.  One can imagine that this would cause fronting or splitting issues when the analytes elute from the end of the column.

Figure 2  Top:travel through the after injection in a sample solvent containing more organic than the eluent Bottom:

resulting peak shape

Of course, one can imagine a very similar behaviour on injecting a high concentration of analyte, which saturates the packing material at the head of the column.  Analytes not occupying a surface site will need to ‘flood’ forward to find surface with which to interact.  This would be called column overload and would lead to similar peak fronting as described in the Figure 2. 

Figures 3 and 4 show some examples of problems with solvent strength and analyte (volume) overload

Figure 3  dissolved in (A) 33% acetonitrile (B) 66% acetonitrile and (C) 100% acetonitrile. Caffeine at 0.75 mg ml−1, 4 μl injection volume. Column: Luna 3 μm C18(2) 50×2.0 mm. Mobile phase: gradient 5 to 95% acetonitrile (with 0.1% formic acid) in 7.5 min at a flow-rate of 1 ml min⁻¹. (Adapted from Reference 1 with permission

Figure 4  Effect of sample injection volume for a given injection solvent composition. Caffeine at 0. 75 mg ml−1 in 100% acetonitrile. Column: Luna 3 μm C18(2) 50x2.0 mm. Mobile phase: gradient 5 to 95% acetonitrile (with 0.1% formic acid) in 7.5 min at a flow-rate of 1 ml min⁻¹. (Adapted from Reference 1 with permission)

So – now that we know what problems can be created with poor sample solvent choice – let’s consider some facts to help us overcome these issues.

Solvent effects are typically seen through peak fronting or splitting and occur when the sample diluent is more strongly eluting than the eluent – this might include circumstances when the diluent contains more organic solvent, has lower ionic strength or contains a different solvent

Solvent effects typically result in constant retention times and peak areas but reduced peak heights.

To avoid volume overload effects, the sample volume injected should be less than 15% of peak volume of the first eluting peak (from reference 2) (this can be easily calculated from peak width at the base multiplied by the flow rate).  If a loss of resolution can be tolerated then up to 40% of the peak volume may be tolerated – however as always the peak symmetry and resolution values should be checked in practice to ensure no deleterious effects are occurring.  As an example – the peak in Figure 3 has a volume of around 100µl (0.1 min. peak width x 1 mL/min) – giving a maximum indicated injection volume of 15µl.

In terms of sample solvent strength the following guidelines are useful in terms of solvent strength v’s sample volume:

Often sample solubility dictates a high percentage of organic solvent within the sample diluent.  However it is usually possible that, once the sample has been initially dissolved, the degree of organic solvent within subsequent diluent solutions can be reduced in order to more closely match that of the initial HPLC eluent solution.