![]()
Part 1
Fast Liquid Chromatography Advances
![]()
Chapter 1
UHPLC Separations Using Sub-2 μm Particle Size Columns
Julie Schappler, Jean-Luc Veuthey, and Davy Guillarme
School of Pharmaceutical Sciences, University of Geneva, University of Lausanne
1.1. General Introduction to Ultrahigh Performance Liquid Chromatography (UHPLC)
Since the early days of chromatography, various authors have demonstrated the utility of reducing particle size in liquid chromatography;1 indeed, it is well known that the chromatographic efficiency (N) is proportional to particle diameter (dp), according to the following equation:
where L is the column length and h is the reduced plate height. The first particles (100–200 μm) were developed in the 1950s for liquid chromatography (LC), prior to smaller porous particles (in the range of 10 μm) in the early 1970s, although packing reproducibility was an issue at that time. Irregular micro-porous particles were used throughout the 1970s, until spherical material was obtained. In the 1980s, 5 μm became the standard particle diameter and in the early 1990s, 3–3.5 μm particle diameters became commercially available; the latter demonstrated 30–50% faster analysis times and higher efficiencies compared to 5 μm particles. In 2004, the breakthrough came with the introduction of porous silica of very small particle size (1.7 μm), which enabled better resolution compared to the current 5 μm or 3.5 μm materials.2 Several column suppliers currently offer columns packed with particles in the range of 1.5–2 μm.3
The interest of decreasing particle size in HPLC is illustrated in Fig. 1.1, which shows the Van Deemter curves obtained for columns packed with 5 μm, 3.5 μm, and 1.7 μm particles. As the particle diameter decreases, the optimal linear velocity (uopt) is shifted to higher values: uopt = 0.7 mm/s for 5 μm, 0.9 mm/s for 3.5 μm, and 2.1 mm/s for 1.7 μm. It is then possible to work at high flow rates with small particles without any loss in efficiency. In agreement with Fig. 1.1, the corresponding H value decreases with the particle size: Hopt = 12.3 μm for 5 μm, Hopt = 8.8 μm for 3.5 μm, and Hopt = 3.9 μm for 1.7 μm. Thus, for a same column length, efficiency can be improved by a factor of three with a column packed with 1.7 μm particles instead of 5 μm. It is also possible to work at flow rates higher than the optimal value without major efficiency loss since the mass transfer is improved using 1.7 μm instead of 5 μm, as shown by the flatter Van Deemter curve.
Figure 1.1. Impact of particle size reduction on the Van Deemter curves. Columns: XTerra, RP18, 4.6 mm × 150 mm, 5 μm; XTerra, RP18, 4.6 mm × 50 mm, 3.5 μm; Acquity BEH Shield, RP18, 2.1 mm × 50 mm, 1.7 μm. Adapted with permission from Nguyen, D.T.T., Guillarme, D., Rudaz, S., Veuthey, J.L. (2006). Chromatographic behaviour and comparison of column packed with sub-2 μm stationary phases in liquid chromatography, J. Chromatogr. A, 1128, 105–113. Copyright (2006) Elsevier.
However, the reduction of particle size induces a significant increase in column pressure drop (ΔP), proportional to the square of the particle size, according to Darcy’s law (and even to the cube of the particle size when working at the optimal linear velocity, the latter being also inversely proportional to particle size):
where η is the mobile phase viscosity, L the column length, and Φ the flow resistance. Considering this constraint, it is required to employ columns packed with sub-2 μm particles exclusively on a new generation of instruments compatible with ultrahigh pressures, as stated by John Knox back in 1977.4 He mentioned that ultrafast LC (i.e. short analysis time but low resolution) would require a new generation of particles and instrumentations. Particles of 1 μm or 2 μm and column lengths between 20 mm and 40 mm should be used to obtain t0 = 10 s with reasonable pressures. Due to the strong reduction of the retention volume and the high applied-flow rate, the primary instrumental limitations would be the injector and detector performance (i.e. the injected quantity, the detector time constant, and the cell volume), as well as the system upper pressure limit. For this reason, 20 to 30 years have been spent to develop sub-2 μm particles and short columns. Today, such columns are available from several providers and instruments compatible with pressures in the range 1,000–1,300 bar are also accessible.
1.2. Proof-of-Concept of Ultrahigh-Pressure Liquid Chromatography (UHPLC) During the 1990s
Before the commercialization of UHPLC technology in 2004, the proof-of-concept was demonstrated by the groups of Jorgenson and Lee during the 1990s with a few impressive chromatographic separations.
The first outstanding separations performed at a ΔPmax of 4,100 bar were described in 1997 by Jorgenson et al. using fused-silica capillary (30 μm I.D., 52 cm length, packed with 1.5 μm non-porous particles). With this setup, a number of plates between 140,000 and 190,000 was reached for small molecules, together with analysis times lower than 10 min.5 In 2003, Jorgenson’s group increased the upper pressure limit and further reduced the particle size, to obtain exceptional chromatographic performance. A separation of five compounds under isocratic conditions was obtained (50 μm I.D., 43 cm length, packed with 1.0 μm non-porous particles) at a pressure of more than 7,000 bar. As illustrated in Fig. 1.2, analysis time was reduced to less than 4 min and plate count ranged between 190,000 and 300,000.6
In the meantime, Lee et al. also investigated UHPLC with capillary columns (29 μm I.D., 12.5 cm length, packed with 1.5 μm non-porous particles).7 They constructed a homemade chromatographic apparatus compatible with pressure up to 3,600 bar. Separations of several benzodiazepines and herbicides were reached in less than 60 s and 100 s, respectively, with efficiency ranging from 20,000 to 30,000 plates.8 Lee et al. also studied the combination of ultrahigh pressures with elevated temperatures.9,10 Indeed, because high temperature induces a reduction of the mobile phase viscosity, system pressure drop remains reasonable. For this series of experiments, a capillary column (50 μm I.D., 14.5 cm length, packed with 1.0 μm non-porous particles) packed with a zirconia phase was employed due to its chemical stability at elevated temperatures. A separation of benzodiazepines was performed in less than 1.2 min at 100° C at a pressure of only 1,480 bar, while five herbicides were resolved with excellent efficiency in 60 s at 90° C and 1,800 bar.
Figure 1.2. Chromatograms obtained with a column packed with 1.0 μm particles at run pressures of about 7,000 bar. Adapted with permission from Jerkovich, A.D., Mellors, J.S., Jorgenson, J.W. (2003). The use of micron-sized particles in ultrahigh-pressure liquid chromatography, LCGC Eur., 16, 20–30.
As illustrated with these few examples, in the early days of UHPLC Jorgenson et al. demonstrated that UHPLC was a viable strategy for attaining high plate count (N > 100,000), ideal for the separation of complex mixtures of compounds. Conversely, Lee et al. mainly used UHPLC for performing rapid analysis of small molecules (in few minutes) with limited efficiency (N < 30,000).
1.3. Benefits of UHPLC Technique: Speed, Resolution, Solvent Consumption and Sensitivity
By adequately selecting the column length in UHPLC, it is possible, from a theoretical point of view, to increase the throughput by a factor of 9 compared to conventional HPLC while maintaining similar chromatographic efficiency. For example, if the original HPLC separation is carried out on a 150 mm, 5 μm column, then a 50 mm, 1.7 μm stationary phase should be selected in UHPLC, to achieve equivalent performance. The L/dp ratio is similar between these two columns (equal to 30,000 and 29,400 respectively), generating equal plate number, in agreement with Eq. (1.1). The analysis time is reduced nine-fold, due to the three times shorter column and three times higher linear velocity. Such ultrafast separations have been experimentally obtained both in isocratic and gradient modes, and analysis times in the range 1–5 min can thus be expected.11,12 This enhanced throughput is illustrated by the HPLC and UHPLC separations of a standardized extract of a widely used phytomedicine, Ginkgo biloba (Fig. 1.3). In agreement with the theory, a nine-fold reduction of the analysis time was obtained by transferring the 60 min HPLC gradient (on a 150 × 4.6 mm, 5 μm HPLC column, Fig. 1.3A) to a short UHPLC gradient (on a 50 mm × 2.1 mm, 1.7 μm UHPLC column, Fig. 1.3C).3
Figure 1.3. Comparison of chromatograms of a standardized Gingko biloba extract using method transfer. (A) Classical HPLC analysis carried out on a 150 mm × 4.6 mm, 5 μm column, with gradient of 5–40% ACN in 60 min at 1 mL/min. (B) HPLC method transferred on a 150 × 2.1 mm, 1.7 μm UHPLC column, with gradient of 5–40% ACN in 60 min at 0.350 mL/min. (C) Geometric method transfer to a 50 × 2.1 mm, 1.7 μm UHPLC column (with the same phase chemistry), with ...
