Retention modelling in hydrophilic interaction chromatography

A perspective on the use of deep deterministic policy gradient reinforcement learning for retention time modeling in reversed-phase liquid chromatography

Artificial intelligence and machine learning techniques are increasingly used for different tasks related to method development in liquid chromatography. In this study, the possibilities of a reinforcement learning algorithm, more specifically a deep deterministic policy gradient algorithm, are evaluated for the selection of scouting runs for retention time modeling. As a theoretical exercise, it is investigated whether such an algorithm can be trained to select scouting runs for any compound of interest allowing to retrieve its correct retention parameters for the three-parameter Neue-Kuss retention model. It is observed that three scouting runs are generally sufficient to retrieve the retention parameters with an accuracy (mean relative percentage error MRPE) of 1 % or less. When given the opportunity to select additional scouting runs, this does not lead to a significantly improved accuracy. It is also observed that the agent tends to give preference to isocratic scouting runs for retention time modeling, and is only motivated towards selecting gradient scouting runs when penalized (strongly) for large analysis/gradient times. This seems to reinforce the general power and usefulness of isocratic scouting runs for retention time modeling. Finally, the best results (lowest MRPE) are obtained when the agent manages to retrieve retention time data for % ACN at elution of the compound under consideration that spread the entire relevant range of ACN (5 % ACN to 95 % ACN) as well as possible, i.e., resulting in retention data at a low, intermediate and high % ACN. Based on the obtained results, we believe reinforcement learning holds great potential to automate and rationalize method development in liquid chromatography in the future.

Hydrophilic interaction liquid chromatography with methanol-water eluent on a zeolite

BACKGROUND

Hydrophilic interaction chromatography (HILIC) works with organic solvent-water mixtures as eluent and is based on the formation of a water enriched liquid phase on the surface of a hydrophilic stationary phase. Hydrophilic solutes are retained on that stagnant water-rich film depending on the difference of solvation compared to the mobile phase composition. However, the enhancement of selectivity by increasing the fraction of organic cosolvent is coupled with a limitation the analyte solubility, and the improvement of the HILIC principle by new hydrophilic stationary phases is the remaining option.

RESULTS

Y-zeolite (faujasite, FAU type) in the Na⁺-form with an average particle diameter of 5 μm was used as packing material in a 125 mm long HPLC column. The chromatographic response of the column was tested in methanol-water mixtures as eluent after injection of several aliphatic alcohols, polyols and monosaccharides with eluent conditions where no separation occurs on diol functionalized silica. On the zeolite the retention time increases according to ethylene glycol < glycerol < erythritol < sorbitol < inositol. The separation principle is explained to be superposed by two effects: firstly, a partition equilibrium between the water-rich phase in the zeolite micropores exists, and secondly, selective interactions with the inner crystalline pore surface and fixed-position Na⁺ ions, both serving to enhance the selectivity. Furthermore, arabinose and fructose monosaccharides could be separated into their tautomeric forms. Only upon increasing the temperature from 20 to 60 °C the tautomeric pattern merges into a single peak.

SIGNIFICANCE AND NOVELTY

Instead of the stagnant water rich surface layer, zeolite micropores now take over that function. As a result, the selectivity among polyols and between α/β-arabinopyranose and β-fructopyranose/β-fructofuranose tautomers is extraordinary superior towards conventional hydrophilic interaction liquid chromatography (HILIC).

Chromatographic separation of glycated peptide isomers derived from glucose and fructose

Amino groups in proteins can react with aldehyde groups in aldoses or keto groups in ketoses, e.g., D-glucose and D-fructose, yielding Schiff bases that rearrange to more stable Amadori and Heyns products, respectively. Analytical strategies to identify and quantify each glycation product in the presence of the corresponding isomer are challenged by similar physicochemical properties, impeding chromatographic separations, and by identical masses including very similar fragmentation patterns in tandem mass spectrometry. Thus, we studied the separation of seven peptide families, each consisting of unmodified, glucated, and fructated 15mer to 22mer peptides using reversed-phase (RP) and hydrophilic interaction chromatography (HILIC). In RP-HPLC using acidic acetonitrile gradients, unglycated peptides eluted ~ 0.1 to 0.8 min after the corresponding glycated peptides with four of seven peptides being baseline separated. Isomeric glucated and fructated peptides typically coeluted, although two late-eluting peptides were partially separated. Neutral eluents (pH 7.2) improved the chromatographic resolution (R_s), especially in the presence of phosphate, providing good and often even baseline separations for six of the seven isomeric glycated peptide pairs with fructated peptides eluting earlier (R_s = 0.7 to 1.5). Some glucated and unmodified peptides coeluted, but they can be distinguished by mass spectrometry. HILIC separated glycated and unmodified peptides well, whereas glucated and fructated peptides typically coeluted. In conclusion, HILIC efficiently separated unmodified and the corresponding glycated peptides, while isomeric Amadori and Heyns peptides were best separated by RP-HPLC using phosphate buffered eluents.

Comparison of the steric selectivity on hydrophilic interaction chromatography columns modified with poly(acrylamide) possessing different morphology

Poly(acrylamide) (PAAm)-modified hydrophilic interaction chromatography (HILIC) columns were prepared via surface-initiated atom transfer radical polymerization (SI-ATRP) and free radical polymerization (FRP) to generate brush-like and mushroom-like polymer chains on silica particles, respectively. The maltose homologues (MHs) and cyclodextrins (CDs) were chosen as analytes to evaluate steric selectivity by the different polymer morphologies in the ATRP-PAAm and the FRP-PAAm columns. The ATRP-PAAm exhibited superior retention than the FRP-PAAm and three commercial HILIC columns. The house-made PAAm columns provided significant hydrophilicity that enabled to analysis the oligosaccharides even in 60:40 mixture of acetonitrile-aqueous buffer. In the case of three ATRP-PAAm columns characterized by different polymer lengths and the density on the silica particles, those are different thickness of the water-enriched layer, and phase ratio φ, based on hydrophilicity of them columns. The logarithm of the retention factor (ln k) displayed a non-linear dependence on the inverse of the temperature (1/T, T = 278-333 K). Notably, a similar correlation was observed to exist between the logarithm of the phase ratio (ln φ), and 1/T. A van't Hoff plot was used to determine the thermodynamic parameters of the partition process for each MH. The values of the Gibbs free energy (ΔG°) for the analytes partition on the ATRP-PAAm columns were smaller than their counterparts measured for the FRP-PAAm columns; by contrast, the opposite trend was observed for the ΔG° values measured for CDs. The standard entropy ΔS° for MHs and CDs were comparable for the two types PAAm columns, while, the standard enthalpy, ΔH° displays significant difference between the ATRP and the FRP PAAm columns. These findings indicate that the differences between PAAm morphology and polymer densities on the stationary phase surface affect analyte differentiation on the basis of molecular steric factors. The higher selectivity for MHs and CDs displayed by ATRP-PAAm columns with respect to their FRP-PAAm and commercial amide columns will be useful for the fine separation of oligosaccharides.

Comparison of the Fitting Performance of Retention Models and Elution Strength Behaviour in Hydrophilic-Interaction and Reversed-Phase Liquid Chromatography

Hydrophilic interaction liquid chromatography (HILIC) is able to separate from polar to highly polar solutes, using similar eluents to those in the reversed-phase mode (RPLC) and a polar stationary phase, where water is adsorbed onto its surface. It is widely accepted that multiple modes of interaction take place in the HILIC environment, which can be far more complex than the interactions in an RPLC column. The behaviour in HILIC should be adequately modelled to predict the retention with optimisation purposes and improve the understanding on retention mechanisms, as is the case for RPLC. In this work, the prediction performance of several retention models is studied for seven HILIC columns (underivatised silica, and silica containing diol, amino and sulfobetaine functional groups, together with three columns recently manufactured with neutral, anionic, and cationic character), using uracil and six polar nucleosides (adenosine, cytidine, guanosine, thymidine, uridine, and xanthosine) as probe compounds. The results in HILIC are compared with those that were offered by the elution of several polar sulphonamides and diuretics analysed with two C18 columns (Chromolith Speed ROD and Zorbax Eclipse XDB). It is shown that eight retention models, which only consider partitioning or both partitioning and adsorption, give similar good accuracy in predictions for both HILIC and RPLC columns. However, the study on the elution strength behaviour, at varying mobile phase composition, reveals similarities (or differences) between RPLC and HILIC columns of diverse nature. The particular behaviour for the HILIC and RPLC columns was also revealed when the retention, in both modes, was fitted to a model that describes the change in the elution strength with the modifier concentration.

Performance evaluation of silica microspheres functionalized by different amine-ligands for hydrophilic interaction chromatography

In this work, for the first time five amine-ligands including mono-amine, di-amine, tri-amine, secondary and tertiary amine, were functionalized on mesoporous micro-silicas and developed as stationary phases for hydrophilic interaction liquid chromatography (HILIC). The investigations about the retention mechanisms, effects of different chromatographic conditions and stability were systematically conducted. Three kinds of polar and hydrophilic compounds (saccharides, sulfonamides, nucleosides and nucleobases) were selected as probe molecules to evaluate their separation performances. Among the five stationary phases, only aminopropyl-bonded silica has already gained wide developments and applications. Whereas, there are no related researches about the other four to be utilized as separation media. By a series of chromatographic evaluations, the results revealed the other four mesoporous micro-silica materials functionalized with di-amine, tri-amine, secondary and tertiary amine, had great potential to be explored as novel stationary phases of HILIC. Particularly, the two stationary phases functionalized with di-amine and tri-amine exhibited outstanding separation and retention abilities. This work offered some insights on the understanding of retention in HILIC mode and provided us possibility to explore other amine-based HILIC stationary phases.

Global retention models and their application to the prediction of chromatographic fingerprints

The resolution of samples containing unknown compounds of different nature, or without standards available, as is the case of chromatographic fingerprints, is still a challenge. Possibly, the most problematic aspect that prevents systematic method development is finding models that describe without bias the retention behaviour of the compounds in the samples. In this work, the use of global models (able to describe the whole sample) is proposed as an alternative to the use of individual models for each solute. Global models contain parameters that are specific for each solute, while other parameters ‒related to the column and solvent‒ are common for all solutes. A special regression procedure is presented for the construction of global models, which are applied to predict highly complex chromatograms, such as chromatographic fingerprints, for diverse experimental conditions in isocratic and gradient elution. Another interesting application is the prediction of molecular properties, such as log P_o/w, from the specific solute parameters of the global models. The examined adapted models are based on the equations proposed by Snyder, Schoenmakers, Neue and Kuss, Jandera, and Bosch Rosés to describe the retention. In all cases, the predictive capability was very satisfactory. Two cases of study were considered: chromatograms of camomile extracts analysed using acetonitrile gradients, and a set of 145 known compounds in a wide range of structures and functionalities, eluted isocratically with acetonitrile/water mobile phases.

Recent applications of retention modelling in liquid chromatography

Recent applications of retention modelling in liquid chromatography (2015–2020) are comprehensively reviewed. The fundamentals of the field, which date back much longer, are summarized. Retention modeling is used in retention‐mechanism studies, for determining physical parameters, such as lipophilicity, and for various more‐practical purposes, including method development and optimization, method transfer, and stationary‐phase characterization and comparison. The review focusses on the effects of mobile‐phase composition on retention, but other variables and novel models to describe their effects are also considered. The five most‐common models are addressed in detail, i.e. the log‐linear (linear‐solvent‐strength) model, the quadratic model, the log–log (adsorption) model, the mixed‐mode model, and the Neue–Kuss model. Isocratic and gradient‐elution methods are considered for determining model parameters and the evaluation and validation of fitted models is discussed. Strategies in which retention models are applied for developing and optimizing one‐ and two‐dimensional liquid chromatographic separations are discussed. The review culminates in some overall conclusions and several concrete recommendations.

Rapid HPLC Method for Determination of Isomaltulose in the Presence of Glucose, Sucrose, and Maltodextrins in Dietary Supplements

This paper presents a rapid HPLC method for the separation of isomaltulose (also known as Palatinose) from other common edible carbohydrates such as sucrose, glucose, and maltodextrins, which are commonly present in food and dietary supplements. This method was applied to determine isomaltulose in selected food supplements for special diets and athletic performance. Due to the selectivity of the separation system, this method can also be used for rapid profiling analysis of mono-, di-, and oligosaccharides in food.

Dual-retention mechanism of dopamine-related compounds on monolithic stationary phase with zwitterion functionality

Seven retention models have been selected to describe a dual-retention behavior of ten dopamine-related compounds on polymer-based monolithic stationary phase with zwitterion sulfobetaine functionality. Regression quality, as well as a statistical significance of individual regression parameters, have been evaluated. Better regression performance showed two four-parameter models when compared to three-parameter models. On the other hand, limited number of experimental points disqualified statistical robustness of four-parameter models. Among three-parameter models, retention description introduced by Horváth and Liang provided comparable quality of regression at significantly improved robustness. Multivariate analysis of the best three-parameter models provided the description of physicochemical properties of dopamine precursors and metabolites. Principal component analysis and logistic regression allowed structural characterization of dopamine-related compounds based solely on regression parameters extracted from an isocratic elution data. Both polarity and type of functional groups has been correctly assigned for 3-methoxytyramine that has not been part of an evaluation study. Among applied dual-retention models, Horváth´s model, initially developed to describe a retention of ionic compounds on nonpolar stationary phases, provided robust regression of experimental data and allowed an extraction of structural characteristics of dopamine-related compounds.

Chromatographic retention behaviour, modelling and optimization of a UHPLC-UV separation of the regioisomers of the Novel Psychoactive Substance (NPS) methoxphenidine (MXP)

A detailed investigation into the chromatographic retention behaviour and separation of the three regioisomers of the Novel Psychoactive Substance (NPS) methoxphenidine (i.e. 2-, 3- and 4-MXP isomers) has revealed the ionization state of the analyte and stationary phase, to be the controlling factor in dictating which retention mechanism is in operation. At low pH, poor separation and retention was observed. In contrast, at intermediate pH, enhanced retention and separation of the three MXP isomers was obtained; it appeared that there was a synergistic effect between the electrostatic and hydrophobic mechanisms. At high pH, the MXP isomers were retained by hydrophobic retention. Accurate retention time predictions (<0.5%) were achievable using non-linear retention models (3 × 3). This allowed the optimization of the gradient separation of the MXP isomers using a two-dimensional gradient and temperature design space. Prediction errors for peak width and resolution were, in most cases, lower than 5%. The use of linear models (2 × 2) still afforded retention time and resolution accuracies of <2.3 and 11% respectively. A rapid and highly sensitive LC-MS friendly method (i.e. R_smin > 5 within 4 min) was predicted and verified. The developed methodology should be highly suitable for the rapid, specific and sensitive detection and control of MXP regioisomers.

Chemometric-assisted method development in hydrophilic interaction liquid chromatography: A review

With an enormous growth in the application of hydrophilic interaction liquid chromatography (HILIC), there has also been significant progress in HILIC method development. HILIC is a chromatographic method that utilises hydro-organic mobile phases with a high organic content, and a hydrophilic stationary phase. It has been applied predominantly in the determination of small polar compounds. Theoretical studies in computer-aided modelling tools, most importantly the predictive, quantitative structure retention relationship (QSRR) modelling methods, have attracted the attention of researchers and these approaches greatly assist the method development process. This review focuses on the application of computer-aided modelling tools in understanding the retention mechanism, the classification of HILIC stationary phases, prediction of retention times in HILIC systems, optimisation of chromatographic conditions, and description of the interaction effects of the chromatographic factors in HILIC separations. Additionally, what has been achieved in the potential application of QSRR methodology in combination with experimental design philosophy in the optimisation of chromatographic separation conditions in the HILIC method development process is communicated. Developing robust predictive QSRR models will undoubtedly facilitate more application of this chromatographic mode in a broader variety of research areas, significantly minimising cost and time of the experimental work.