Part 3: Effect of Flow Rate and Gradient Time
Hey everybody, welcome back to my third blog article on cannabis HPLC method development. So far, we have done some preliminary gradient screening to narrow in an effective “gradient window” (75-100% acetonitrile in 10 min) and found that a range of different acidic modifiers did not have a significant effect on the chromatography of our main cannabinoids. However, this information may become useful later when we try to manipulate matrix interferences. So, for this article, let’s look at changing the flow rate and doing some fine-tuning on our gradient profile.
By the way – thank you so much for the great responses that we have received so far. It sounds like we are on the right track, so let’s keep pressing forward.
Effect of Flow Rate
Changing flow rates can be a powerful tool in HPLC method development. Your total analysis time is related to the flow rate – faster flow rates will lead to shorter analysis times. But, flow rates can also affect analyte resolution since different chromatographic medias have different optimal flow rates depending upon their particle size and morphology. Of course, increasing the flow rate will also increase the pressure in a proportional fashion, which needs to be considered since staying at a relatively low pressure would be ideal. In general, the smaller the particle size, the higher the optimal flow rate will be. For instance, the optimal flow rate for a column packed with fully porous 5 µm particle is about 1 mL/min, but if we move to a fully porous 3 µm, the theoretical optimum flow rate is closer to 1.5 mL/min (I should note that these values are for 4.6 mm I.D. columns). If the flow rates go above or below the optimal amount, a loss of efficiency can theoretically be seen, which would translate into broader peaks and possibly worse resolution between closely-eluting analytes.
Now for core-shell based media, such as the Kinetex® 2.6 µm particles in the 100 x 4.6 mm column, the optimal flow rate is high at around 1.8 mL/min. So essentially, the flow rate should be close to that for the best possible performance. HOWEVER, it’s forgiving and you can generally operate above or below those values without sacrificing too much performance, especially if you are running a gradient. So let’s see what it looks like at the different flow rates.
You can see the resulting chromatograms in Figure 1, and the tabulated results in Table 1.
Figure 1. Effect of different flow rate on cannabinoids separation.
From the Tabulated results (Table 1), increasing the flow rate while it is decreasing total analysis time, causes loss of resolution between peaks 4 and 5. Considering one peak is missing already, losing resolution is not a good idea. An idea would be to re-visit this flow rate later, once more optimization has been done, but for now, I am sticking with 1 mL/min. As a side thought, it might make sense from a purely method development point of view to use 1.5 or 2 mL/min since it will dramatically reduce the time between runs as we make various iterations to the method. However, once the conditions are finalized, the flow rate should be returned to ~1 mL/min.
Effect of Gradient Time
The last factor to look at is the effect of changing the gradient time. Up to now, the gradient going from 75-100% B over 10 minutes, yielded the best results, allowing us to partially resolve 11 of the 12 target cannabinoids. Now it’s time to play with the total gradient time, while holding the starting organic at 75%, and seeing if there can be more resolution.
Holding at % B at 75% to start, and then running it to 100% B over periods of 7 min, 10 min (what we have been using), 15 min, and 20 min. Check out the chromatogram results below in Figure 2. The data reveals that there was very little change in chromatographic performance when making dramatic changes in the gradient time (about a three-fold change in gradient rate overall). This is VERY unusual, and I honestly cannot determine a good explanation. Perhaps it has to do with the extremely high hydrophobicity of these molecules and the gradient already starting in a high percentage of organic solvent. Most of my experience is with polar analytes, so I would have expected a much greater change in retention times if I made analogous changes to a gradient using a more aqueous mobile phase. However, there is an increase in resolution as we increase the total gradient time, so this is something to keep in mind as we do further method development.
Figure 2. Effect of changing the gradient time.
Even at the shallowest gradient evaluated, we still were only able to resolve 11 peaks. Our mystery peak still evades us! I think it’s time to play around with the column itself – maybe try a longer column to get more resolving power, and perhaps try some different stationary phases to see if the right selectivity can be found in order to uncover the elusive 12th peak.