Column Pore Size—An Underutilized Variable for Optimizing

High performance liquid chromatogrpahy (HPLC) or liquid chromatography–mass spectrometry (LC–MS) methods are widely preferred for the analysis of both small and large compounds. As target molecules increase in size, maintaining optimum separation performance requires a change to larger column pore diameters. Giddings (1) cautions that if pores become too small for the sample, larger molecules lose freedom of movement and good access to stationary phase. When sample components approach the pore diameter too closely, poor equilibration and a slow diffusion rate within the pores may cause a sharp reduction in column retention and efficiency. Available data suggest that performance loss may become noticeable when target molecules are larger than about 10% of the mean particle pore diameter (a molecule-to-pore ratio of 0.1).

Discussion

Commercial silica particles in HPLC columns typically have broad pore distributions to accept and separate a useful range of solutes up to a size limit determined by analyte molecular weight (MW) and average particle pore diameter. Unfortunately, a complex relationship exists between MW and size. Molecules with the same MW can have different shapes and sizes in solution. Hydrodynamic volume-the solution space that a molecule occupies-determines whether a molecule can gain access to column pore structure. Since size-exclusion chromatography (SEC) directly measures hydrodynamic volume, it can be a very convenient way to estimate whether a good size match exists between samples and columns.

Figure 1 shows a calibration plot for 12 molecules with increasing size created with an SEC 150-Å column. A small-pore HPLC column could not possibly analyze this sample in reversed-phase mode because the MW and size range is much too large for good performance with larger components. It is clear from Figure 1 that only uracil (point 12) and other very small molecules have complete access to the total pore volume (V_M) of this 150-Å silica column. Fortunately, a certain amount of overlap between exclusion mode and retention modes can usually be tolerated, and good performance can still be achieved for molecules that are eluted in the lower MW half of the calibration curve (points 8–11). Figure 2 shows HPLC results in reversed-phase mode with 90-Å and 160-Å silica C18 columns for a sample containing bovine insulin and other small proteins. Bovine insulin (MW 5700) is about 40% excluded from the pore volume of 150-Å silica in the calibration plot. The 90-Å pores are clearly too small for the larger peptides, which would be greater than 50% excluded from the pores in Figure 1. Even the two insulin peaks experience some loss of performance because they become sharper with greater retention when the C18 column is changed to 160-Å pore silica. Note that the entire peptide sample shows more retention on 160-Å C18 even though that column has lower surface area and therefore less stationary phase. When larger molecules are partially excluded from the pore structure, column retention is determined by accessible stationary phase at a reduced level of column efficiency.

Figure 1: Calibration curve for SEC. Column: 30 cm × 4.6 mm, 3-µm Sepax Zenix SEC-150; mobile phase: 0.2 M potassium phosphate, pH 7.0; flow rate: 0.25 mL/min; pressure: 66 bar; column temperature: 25 °C; detection: UV absorbance at 215 and 280 nm; injection volume: 0.5 µL. (Data provided by AMT.)

Conclusions

The full paper to be presented at Pittcon 2018 in Orlando will review what is known about pores, how they compare in size to solutes, how to make early column selection decisions to maximize the chances for good separation, how to use SEC to recognize when a mismatch may exist between compound size and column pore size, and how to solve problems if poor performance is observed. Improved separations for very large molecules will be shown with new commercially available columns, and further improvements may be possible through a better understanding of the internal pore structure of HPLC and UHPLC columns.

Figure 2: Small protein and large peptide separations obtained using columns with (a) 90-Å and (b) 160-Å pore sizes. Peak widths in minutes are shown above each peak. Columns: (a) 100 mm × 4.6 mm, 2.7-µm, 90-Å Halo C18 (surface area: 135 m²/g), (b) 100 mm × 4.6 mm, 2.7-µm, 160-Å Halo ES-C18 (surface area: 90 m²/g); mobile-phase A: water with 0.1% trifluoroacetic acid; mobile-phase B: acetonitrile with 0.1% trifluoroacetic acid; gradient: 25–42% B in 10 min; flow rate: 1.5 mL/min; temperature: 30 °C; detection: UV absorbance at 215 nm. Peaks: 1 = ribonuclease A (13,700 g/mol), 2 = bovine insulin (5733 g/mol), 3 = human insulin (5808 g/mol), 4 = cytochrome c (12,400 g/mol), 5 = lysozyme (14,300 g/mol).

Reference

(1) J.C. Giddings, Unified Separation Science (Wiley, Hoboken, New Jersey, 1991), chap. 2.

Richard Henry

Richard Henry is an independent consultant in State College, Pennsylvania. Stephanie A. Schuster is Application and Quality Manager with Advanced Materials Technology, Inc., in Wilmington, Delaware.