Assuming a linear gradient and reversed-phase chromatography, the steepness of the gradient controls the
average k' of the peaks (the average k' is symbolized by k*). In that sense, the steepness is analogous to solvent strength (%B). Peak width in a gradient will depend on plate number and k*, just as isocratic peak width depends on plate number and k'.
One major difference from isocratic is that gradient k* also depends on flow rate and column volume. Very roughly, the relationship looks like this:
k* ~ (gradient time / gradient range)*(column volume / flow)*(1/S)
where S is the slope of the relationship between log(k') and %B (typically somewhere around 500 for small molecules).
So, if you want narrower peaks, you work at lower k*:
- use a steeper gradient (analogous to a higher %B),
- a higher flow rate (assuming constant column dimensions) and
- a higher plate count for given column dimensions (i.e., smaller particles). If you want to increase sensitivity further, use a smaller (narrower) column.
Note that working at very low k* can run you into the same loss of resolution you would encounter at low k' in an isocratic separation. In principle, you will get the best
sensitivity if you use an infinitely steep gradient and an infinitesimally small column (i.e. inject directly with no column).
