A generic liquid chromatographic method development workflow was developed and successfully applied to the analysis of phytocannabinoids and Cannabis sativaextracts. Our method development procedure consists in four steps:

i)The screening of primary parameters (i.e. stationary phase nature, organic modifier nature and approximate mobile phase pH) was carried out with a generic gradient on a short narrow bore column, using a system able to accommodate numerous solvents/buffers and columns. Instead of complete peak tracking, the number of peaks which can be separated was considered as a response at this level, to save time.

ii)The optimization of secondary parameters (i.e. gradient conditions, mobile phase temperature and pH within a narrow range) requires only 12 initial experiments and the use of HPLC modeling software for data treatment. It allows to find out the best retention and selectivity for the selected compounds. Peak tracking was performed with a single quadruple mass detector in single ion recording mode, and UV detection (in a broad wavelength range).

iii)The refinement step allows to further adjust column efficiency, by tuning column length and mobile phase flow rate. This can also be done virtually using HPLC modeling software.

iv)The robustness testing step was also evaluated from a virtual experimental design. Success rate and regression coefficients were estimated in about 1 min, without the need to perform any real experiment.

At the end, this method development workflow was performed in less than 4 days and minimizes the costs of the method development in liquid chromatography.

Chromatographic methods are progressing continuously. Increasing sample complexity and safety expectations lead to higher regulatory demands, hence challenges in liquid chromatography analysis are rising, even today, when faster and faster chromatographic systems are extensively employed and become widely accessible for successful method development.

The goal of this study was to investigate the impact of mobile phase influences as important factors of selectivity tuning in method development. This would mitigate mobile phase-related robustness issues throughout the method's lifecycle.

To discover and understand these effects, a new module of chromatographic modeling software DryLab (ver. 4.3.4. beta) was introduced and a special experimental design (DoE) was tested, allowing the simultaneous optimization of solvent-dependent parameters, such as gradient time (t_G), ternary eluent composition (t_C) and pH, requiring 18 input experiments (2 × 3 × 3 = 18).

Additionally, the model creation, using a UPLC system and a narrow bore column (50 × 2.1 mm), the entire experimental work could be finished in 2-3 hours. To demonstrate the applicability of this new design, amlodipine and its related pharmacopoeia impurities (A-H) were subjected to be used in a case study. Predicted vs. Experimental (or Verification) runs showed excellent agreement, average retention time deviations were typically less than 1 s. Modelled robustness testing was also performed, elucidating all important mobile phase and instrument parameters that could influence a method's lifetime performance. Furthermore, as the in silico robustness testing is the least time consuming part of the method development process, it can be used extensively to evaluate robustness even at the very early part in stage 1 of the Method Life Cycle (MLC).

Highlights:

• Proper correlation between logD and relative retention in LC.
• logD calculation is a good basis for starting method screening.
• Drylab software modeling tool to comply analytical QbD requirements in regards to lifecycle principles.
• In silico generation of extended knowledge space possible.
• Extreme high robustness of resulting HPLC method with improved reliability and understanding.

Highlights:

• A wide-pore silica-based SPP material with a high coverage phenyl bonding was evaluated.
• Kinetic performance, recovery and selectivity were systematically studied.
• Optimal conditions were determined using Drylab 4 software.
• The possibility to replace TFA with FA was studied.
• The new column was applied for the separation of mAb sub-units and ADC species.