Hydrophilic interaction liquid chromatography (HILIC) has become a powerful technique for the separation of polar and ionizable compounds in the context of samples of biological interest, particularly in the field of omics sciences (metabolomics, genomics, proteomics, glycomics…), but also in pharmaceutical, environmental, and food analysis [1], [2], [3], [4], [5], [6], [7], [8]. HILIC stationary phases consist mainly of bare silica or silica functionalized with polar bonded phases, that can be neutral (amide, cyclofructan, cyclodextrin, diol, pentanediol, polyacrylamide, polyhydroxy…), anionic (aspartic acid, carboxylic acid, sulfoethyl…), cationic (amino, tertiary/quaternary amine, polyethyleneimine…), or zwitterionic (phosphorylcholine, sulfobetaine…) [9,10]. Mobile phases commonly used in HILIC consist of hydroorganic mixtures, such as in reversed-phase liquid chromatography (RPLC), but with a significantly higher content of organic solvent (usually acetonitrile, >70%).

In this paper, the effect of the mobile phase composition on the behavior of HILIC systems consisting of a bare silica column and acetonitrile/water mobile phases is investigated. Understanding the behavior of underivatized silica in HILIC mode is essential to proceed with further and more complex studies involving bonded phases. Surface residual silanol groups have shown to interact much more strongly with water than with acetonitrile resulting in the formation of a water layer adsorbed on the surface of the stationary phase [11], [12], [13], [14]. Additionally, molecular dynamics studies performed in silica with acetonitrile/water mixtures, pointed out the existence of three solvent regions of different composition inside the pores [15], [16], [17]: (i) a water layer strongly adsorbed on the silica surface with a very reduced mobility; (ii) adjacent enriched water transition layers containing a gradually decreasing amount of water and with a translational mobility that increases until the composition and the mobility of the flowing mobile phase is reached; (iii) the bulk hydroorganic mobile phase flowing through the column. However, since adsorbed water is in dynamic equilibrium between these three regions, there is not a clear separation between them [18,19].

Although the retention mechanism in HILIC is complex and still under study, it is usually assumed that it is based on the partitioning of solutes between the bulk mobile phase and the enriched water layers partially immobilized inside the pores, as suggested by Alpert who was the first to introduce the term HILIC [20]. Therefore, this solvent region enriched in water is considered as the effective stationary phase. However, the role of secondary interactions between the column material and the solute, such as hydrogen bonding and electrostatic interactions, should also be considered in the retention behavior. The extent of each contribution to retention depends on the characteristics of the solutes, the column material, and the mobile phase composition, especially the water content [21], [22], [23], [24], [25]. The chemistry of the functionalization and the chromatographic support have a significant effect on the water uptake capacity of the column: the more polar they are (particularly zwitterionic and aminopropyl), the greater the amount of water adsorbed inside the column [14,[26], [27], [28], [29]].

Depending on the mobile phase composition, the same HILIC column can also show an RPLC or even a mixed HILIC-RPLC retention behavior. At high concentrations of acetonitrile, a HILIC mechanism controls the retention. When gradually increasing the water content in the eluent, solute retention decreases since differences in polarity between the mobile phase and the enriched water layers acting as stationary phase become less pronounced. When further increasing the water concentration, the enriched water layers are reduced until solutes are able to interact directly with the silica. If the column exhibits a polar bonded phase, such interaction is extended to this phase as well. At this point, solute retention starts to increase again, indicating that an RPLC mechanism takes control of the chromatographic retention [29], [30], [31], [32].

Obviously, the exact composition of the liquid filling up the mesopores will influence the intra-particle diffusion Dpart, which is the factor determining the stationary zone contribution to band broadening [33]. Dpart to a large extent also determines the effective diffusion (= longitudinal diffusion responsible for the so-called B-term band broadening) [34]. As this effective diffusion is isotropic, it also largely determines the radial dispersion, which in turn determines the A-term band broadening [35]. Knowing how these factors are influenced by the exact composition of the liquid in the mesopores is hence very important.

The main objective of this work is the comprehensive analysis of the effective and intra-particle diffusion of very weakly retained compounds in a fully porous particle silica column operated in HILIC conditions. The weak retention conditions were selected out of pure scientific curiosity, to see how sensitive to small differences the intra-particle diffusion can be. For the sake of comparison, some stronger retained analytes (inosine, uridine) displaying pure HILIC-behavior were also added. To interpret the diffusion measurements, the different column volumes, mainly the void volume and the part of it only accessible to unretained solutes, and the particle porosities were first estimated. Additionally, the calculation of the enriched water layer thickness provided valuable insights into the relationship between the mobile and the stationary phases.