A systematic approach to HPLC method development for small molecules: where do I start? Part -2

We were talking about how to develop HPLC analytical methods for multiple analytes, how to resolve issues involved in the gradient method, and how to get all the components separated, no matter how complex the sample composites are. We discussed how to determine whether the target analytical method is qualified for isocratic separation or whether gradient separation is mandatory irrespective of the number of analytes being separated based on the ‘25% rule’ corresponding to the total gradient time. We also discussed how to calculate the gradient time to be used for the initial gradient scouting step, which is the key to any gradient separation including for our imaginary drug substance "Khilgaon-X" and the associated impurities. Some information has been missing in the previous article, and in this write-up, I will include those missing points so that it becomes clear to obtain the whole picture of the gradient method development for any analytical method to be developed from ‘scratch’ or through a systematic method development approach.

We need to keep in mind the 'synchronization’ of any analytical method among the components involves, e.g., the HPLC column, column length, internal diameter, particle size, particle’s pore diameter, surface area of packing materials, flow rate being considered, void volume and void time, dwell volume, and the most important thing, the ratio of mobile phase B (organic phase) selected at the gradient scouting step ( for example 5% B to 95% B and the corresponding aqueous phase). We need to calculate the target gradient time (tG) using most of the components mentioned above. Once the expected separation is achieved at the initial gradient scouting step with 5% B to 95% B organic phase, and the corresponding aqueous phase, the next thing is to figure out the target initial composition of mobile phase A and mobile phase B based on the run time we want to select. Optimization of a reasonable run time is one of the major components of the target method development since it involves resources. After the estimation of initial gradient composition, now the important part to match the average capacity factor (K*) with the same average capacity factor obtained at the initial gradient scouting step. For both the cases, it must be same, otherwise, whatever achieved at the scouting step will not repeat at the subsequent optimization steps which we don’t want. At this step, we must calculate the target gradient time (tG) keeping the ratio of mobile phase B (for example, at the initial scouting step, percent B was 5% to 95%, at the optimization step, it can be 78% to 95%), associated tG and the retention factor K* consistent. This is the key thing to ensure upon which the chromatographic method is expected to maintain the internal column environment and a homogeneous separation mechanism. This is not the thing that the target gradient time can be changed intentionally; it solely depends on the ratio of the mean capacity factor (K*≈ 5) or any target set value. Without this synchronization, analytical methods can be developed, but that won’t work in the long run. Some random results may appear seems to be perfect; however, that’s not the true picture of the method. To obtain the true picture of the analytical method and consistent behavior of the analytical method irrespective of HPLC systems, the first and foremost thing is to calculate everything on a systematic approach and to ensure the so called ‘synchronization’ among components.

Most HPLC analytical methods have been found to lose the expected performance just because of either losing resolution during robustness testing with varied conditions, system to system switch-over, or losing target detection response, especially at the reporting threshold concentration or even lower concentration than the reporting threshold or other critical parameters to achieve, and all these happen just because of a lack of ‘synchronization’ of components which were incorrectly selected during the initial phase of method development.

The target analytes of interest separates from each other simply due to their respective polarity differences during reverse phase HPLC gradient separation, no matter how close they are to each other, structurally or based on physical or chemical properties; eventually they are considered as independent chemical entities, and gradient separation will allow them to be separated through partitioning between stationary and mobile phases based on the affinity. Since we are talking about the different polarity of analytes being separated, however, at our initial gradient scouting step the impact may not be significantly visible as the aqueous phase (purified water pH around 4-5) or organic phase pH has not been customized yet based on the analyte's acidic or basic nature, and that would be the next step of method development.

At the same time, we need to keep in mind the basic performance of any chromatographic method when analytes are either basic, acidic, or neutral in nature, as well as the requirements of the mobile phase pH, the target analyte’s retention time, and so on. Irrespective of gradient or isocratic separation, one basic chromatographic method performance goal is to keep analytes of interest in ‘un-ionized state’ which will ensure a longer retention time and a better peak shape. Method developers are aware about it. There is no impact of mobile phase pH on neutral analytes, however, acidic analytes are supposed to be in an ionized state if mobile phase pH is above the pKa value of the analyte, and the other way around for the basic analyte. So, it is important to adjust the mobile phase pH at the subsequent method development steps immediately after the initial gradient scouting step, depending on the separation requirements and the analyte’s nature.

In our previous discussion, we have shown how to qualify whether the target analytical method is qualified for isocratic or gradient separation since we don’t want to maintain an unnecessary burden of gradient if the method is truly qualified for isocratic separation, no matter how many analytes of interest are present. We need to apply the '25% rule'. For other situations, if the gradient method is mandatory, one key thing we need to keep in mind is to select the initial gradient composition based on the retention times of the first peak and the last peak (in our imaginary example, Khil-1 and Khil-6). The correct selection of this initial gradient composition of Mobile phase A and mobile phase B will allow us to optimize the subsequent step of method development and is considered the breakthrough of the method development. This initial composition will help us to reduce the total run time of the method, for example, from 50 minutes to 15 minutes while keeping the same separation profiles of all the target analytes of interest obtained during initial gradient scouting step. We can do simple arithmetic to calculate the initial gradient composition set up, or this composition can be selected from the linear curve.

Here is the two key takeaways from this discussion: First, the importance of adding a suitable column equilibration time (approximately 3–4 times duration of the void times) at the beginning, immediately before the gradient starts, and at the end of the gradient. Second and most important thing, re-calculating the total gradient time (tG) for the new initial gradient composition, which will be different than the initial gradient time already used at the gradient scouting step where the mobile phase composition was 5% B (acetonitrile) to 95% B (acetonitrile). With the new initial gradient composition, since we want to reduce the total run time from 50 minutes to 15 minutes for example, while keeping the separation profile exactly same as the initial gradient scouting step, we need to confirm the same mean capacity factor as it was in the scouting step (K*≈5) which we already mentioned above. This is the key. Otherwise, everything will be changed, our so called ‘synchronization’ will be lost, and the method will lose its consistency. We can play around at this step. We can change flow rate if there is any further room for improvement, considering column back pressure or possible viscosity of the mobile phase, including column temperature to add to optimize the column back pressure, and so on. The key thing is to keep average or mean retention factor, K* consistent to the initial scouting step.?We can make our own calculator to establish the new gradient time for the changed initial gradient composition.

This is all about the reverse phase gradient separation for multiple analytes in a single composite sample. As we mentioned, depending on the separation requirements, especially for the critical peak pairs, they still need to be resolved for achieving the target level resolution requirements (may be used USP valley to ratio as well for some cases), and depending on the analytes' acidic or basic nature, the mobile phase pH can be optimized to keep analytes of interest "un-ionized’ in the mobile phase, partitioning can be extended, and a better peak shape can be achieved at the subsequent method development steps.