My LC Blog: Retention (k)

Today's article will elaborate on what I wrote in part 3 and 4 on separation modes and reversed phase HPLC. In method development, a first crucial step is to obtain adequate sample retention times with a selected column and mobile phase combination. But what does that mean - adequate retention? To answer this we have to look at the so-called retention factor (k), which is a measure of the degree to which a compound is retained by the column, relative to an unretained component. Any compound that is injected in the flow path must travel to the detector cell before a peak is displayed in the chromatogram. Even a sample that exhibits no affinity to the stationary phase will not elute at time 0, but after the time it takes to travel that distance. This "dead-time" or t0 depends on the column volume and flow rate of the mobile phase. To render k a value that is comparable and independent of column format and operating conditions in an isocratic run, it is calculated by taking these into account. So that means - more equations - but also a visual explanation, as shown in figure 1. For gradient separations it's a bit more complicated, we’ll leave that for later.

So, with this you can calculate the volume of your column and resulting "dead-time", which then allows to determine k of any compound peak. With a k value of all analytes between 2-10 there is sufficient retention to offer good resolution, and peaks are less likely to be affected by matrix effects, which elute near the void of the column. Also, there is no merit in a k > 10, as there is little gain in resolution above 10, but greater values will mean excessive elution times, therefore longer analysis time and greater solvent consumption, as well as increased band spreading and decreased peak height.

There are 2 main factors affecting retention in reversed phase HPLC, the sample's interaction with the stationary phase and the solvent strength of the mobile phase. A lower amount of organic solvent will result in an increase in retention. The so-called "Rule of Three", which states that a 10 % change in aqueous/organic ratio should change k values by a factor of about 3, allows for an approximate prediction of the effect of changes in the amount of organic solvent in the mobile phase. This and other shortcuts to estimating chromatographic parameters can be found in this helpful (free!) article written by John Dolan and published in LCGC Supplements: "My Favorite Shortcuts".

So, as can be seen in figure 2, if you were to plot log k versus the %organic modifier, you'd obtain a linear graph. If you did this for several compounds, and the result were parallel graphs, it would mean that all analytes in question exhibit the same retention mechanism and will not co-elute, even if you increase the amount of %B to speed up the analysis. However, if the result are crossing lines, than that means changes in %B will affect not only retention, but also selectivity and may result in co-elution. Here the graphs can give you an indication which aqueous/organic ratio would be a good choice with regards to resolution, but also elution order.

Now, coming to factor 2 that controls retention in HPLC, the stationary phase. I already used the term "retention mechanism" which explains the reason why, or in what way an analyte may be interacting with a bonded phase. I list a few examples in table 1.

With some experience and a knowledge on the type and characteristics of the target samples, a chromatographer can make an informed choice on which column to select to start method development. However, the sample compounds may not always be known, or there are a variety of different compounds in the mix, that make it difficult to judge which column would offer the best retention (and selectivity!) for the analytes of interest. In that case a column screening is the way to go: Select a set of several columns with different characteristics and start a list of experiments on all of them to see which column looks the most promising. Ideally, you'd also pick a variety of different mobile phases to evaluate ideal starting conditions for method optimization. However - this will be the topic of a future article: Next up: Selectivity (α).