The World of Separation Science: Advanced Column Technology Forces Next-Generation Instrumentation at HPLC 2011

Throughout the development of HPLC, advances in column performance, primarily efficiency, have exposed deficiencies in instrumentation performance, leading to another generation. The most recent case in point became evident at HPLC 2011, which attracted almost 1400 scientists and vendors to the Budapest Convention Center (Hungary), June 19–23, 2011. About a week before the meeting, this cynical correspondent received leaks from the two leading vendors about introduction of new instruments with "ultralow dispersion." My ontology of HPLC translates this to "low extracolumn band broadening," which was common usage in the 1970s to ~2000. So something was up, but what?

Prof. Georges Guichon (University of Tennessee, Knoxville) connected the dots in his opening lecture, "Recent Progress in Column Technology Begets Progress in our Understanding of Column Efficiency." He compared core/shell and porous particles with nominal particle diameters of 2.6 and 1.7 μm. The performance of columns packed with porous particles showed the expected trend of improved efficiency as measured by height equivalent to a theoretical plate (HETP) with smaller particle size. However, the efficiency of some of the columns packed with sub-2 core/shell particles (particularly Kinetex from Phenomenex [Torrance, CA]) was dramatically better, with a reduced HETP of 0.9–1.5. Previously, the lowest values were 2.0 or slightly less. Plus, the axial and eddy diffusion terms were much smaller. This has practical importance, since one might then run the columns at much higher mobile phase flow without losing chromatographic resolution.

Connecting the dots, the difference between the kinetics of the porous and core/shell particles was finally attributed to differences in heat conduction. As the mobile phase is forced past the particles in the packed bed, frictional heating occurs. The interior of the porous particles has very low heat conduction. Thus, the mobile phase heats up, especially in the center of the bed. Increasing temperature reduces viscosity, giving an increase in the parabolic flow and possibly fingering. Both contribute to band broadening.

With core/shell particles, the core is a better heat conductor, which reduces radial thermal inhomogeneity. Chromatographic bands are tighter and better shaped. Prof. Guichon speculated that core/shell particles could be improved by increasing the thermal conductivity of the cores. Using a metallic core such as gold is an obvious opportunity, but diamond should be even better. He pointed out that diamond dust is remarkably cheap. I recall that diamond dust for a stationary phase had been tried 40 years ago, but the results were unimpressive.

Making these measurements required revising the plumbing of the instruments to reduce extracolumn dispersion. Peak parking and pore blocking protocols were essential in unraveling the mystery of mass transfer kinetics. Waters (Milford, MA) and Agilent (Santa Clara, CA) became aware of the column improvements and rushed ahead with the announcement of reduced dispersion (RD) instruments. Reduced dispersion instruments are probably not for general use, however. One must pay a significant price in terms of pressure needed to force the mobile phase through the capillaries. In the Agilent 1290, this is 1500 psi at 1 mL without any column. At 5 mL/min, the instrument’s flow resistance is about 7000 psi, which leaves only about 10,000 psi for the column. Plus, in gradient elution, the RD instruments show little advantage, since the peak widths are already compressed by the gradient. This is discussed in more detail in the report on the exhibition, which will appear in a future issue of American Laboratory.

What can we expect in future core/shell columns? Sub-2-μm particles will be built around high-conductivity cores. The optimum particle size will probably go down as the molecular weight increases. The optimum is probably about 1.5 μm for small molecules and 1.0 μm for 100-kDa polymers. The diameter of the columns will probably decrease as well from 2.1 mm.

Several others also made presentations comparing porous and core/shell particles. A poster by Matthew R. Linford et al. (Brigham Young University, Provo, UT) described core/shell particles built around 4-μm carbon particles, but the large particle size would probably not show significant frictional heating. Others studied the effect of pore volume and particle morphology. For example, Tivadar Farkas of Phenomenex (Torrance, CA) described scaling down the Kinetex 2.6-μm particles to 1.7 μm diam. For scaleability, the ratio of shell to core remained constant. Holding this ratio allowed the same mobile phase composition to be used while maintaining the separation, and the column efficiency increased to 318,000 plates/m. With the new understanding, the next step will be to reduce the particle size to about 1.1 μm, which will increase the pressure required by 2.3 times. Thus, an ultrahigh-performance liquid chromatography (UHPLC) separation run at 12,000 psi would require 28,000 psi. A new generation of instruments would be required to take advantage of the predicted improvement in column performance. The limit of diminishing returns comes to mind.

Another column technology was described by Prof. Barry Karger (Barnett Institute, Northeastern University, Boston, MA). He used multimeter lengths of wall-coated, 10-μm porous layer open tubular (PLOT) capillaries for separation of proteins from 1 × 104 cells. The stationary phase is a copolymer of polystyrene and divinylbenzene. This column was used as the second dimension in 2-D LC-MS. The first dimension was a strong cation exchange trap with step elution. Prof. Karger’s laboratory identified and quantitated more than 2000 proteins from 10,000 cells from diseased tissue, selected by laser capture microdissection, using this protocol. The technology is being applied to triple-negative breast cancer cells using gene set enrichment analysis (GSEA), which is a weighted running sum protocol to group-enriched proteins in breast cancer tumorigenesis. The researchers found overexpression of minichromosome maintenance proteins (MCM) 2–7 in the tumor cells. These proteins play an essential role in cell replication.

In the future, Prof. Karger expects to improve detection sensitivity to reduce the number of cells required by the sample, ultimately to hundreds or even tens of cells. PLOT columns present an interesting challenge for detection. The volume of peaks is in the low-nanoliter range, which will be a challenge for optical detectors. MS appears to be the best bet.

Organic monolithic stationary phases continue to improve and are beginning to deliver on their promise of providing complex separations at very high speed. Dionex (since May, part of Thermo Fisher Scientific, Waltham, MA) reported expansion of the IonSwift product line with the IonSwift Max 100 column, which is available in 1 mm and 250 μm i.d. × 250 mm long. A combination of UV grafting and polymerizable anchors is used to keep the monolith attached to the column wall during synthesis. With the IonSwift, 19 anions can be separated in half the time (30 min) of the Dionex IonPac AS11-HC, as shown in a side-by-side comparison.

A comparison with the IonPac AS19 shows a similar reduction in run time for 11 common anions. Still another example demonstrated that mobile phase velocity can be increased dramatically without loss of resolution. The maximum flow rate appears to be restricted only by the pressure limit of the PEEK column tubing (~5000 psi). Profs. Frantisek Svec (Lawrence Berkeley Laboratory, CA) and Ziad El Rassi (Oklahoma State University, Stillwater) both presented lectures showing even longer columns that can be used for very complex samples.

Two-dimensional LC also makes sense for complex samples. Several lectures and posters described 2-D separations. One used a combination of hydrophilic interaction chromatography (HILIC) and reversed-phase liquid chromatography (RPLC) to analyze grape tannins for high-resolution separation. On-line 2-D was faster than off-line (stop-flow) protocols, but lost about half the peak capacity. Another lecture presented a novel design for multidimensional high-performance capillary electrophoresis (HPCE) with eight parallel columns in the second dimension. This design permits operating both dimensions under near-optimum conditions.

Green considerations attract global interest. Supercritical fluid chromatography (SFC), which recycles CO2 from industrial sources, was a major topic at the meeting. However, I was impressed with a detailed analysis of the lifecycle of HPLC instruments prepared by Dr. Laura Schneider (Technische Universität, Berlin, Germany). According to her analysis, the biggest environmental challenge is 3510 kg of CO2e, where the "e" refers to equivalents; 59% of this is due to electricity consumed in six years of operation of a Platinum Blue HPLC (Knauer, Berlin, Germany). This HPLC was designed to use less electricity than competitive UHPLCs—it requires only 700 W, while other instruments consume 900–1700 W. Interestingly, HPLC mobile phase accounts for only 12% of the total CO2e burden. Finally, Dr. Schneider noted that due to the high productivity of UHPLC, the CO2e burden per injection is only about 4.6 g. I wonder what the CO2e of a GC is.

Attendance and future meetings

HPLC 2011 set a modern attendance record, with 1332 scientists and vendors. The Hungarian delegation was the largest, with 207. The U.S. delegation was a credible 162, including large groups from Thermo Fisher Scientific and Waters. Chairman Prof. Attila Felinger (University of Pécs, Hungary) reported that the proportion of principal investigators, industrial, and students was about equal. I was impressed with the large number of attendees under the age of 35.

The 37th HPLC meeting is scheduled for October 8–11, 2011 in Dalian, China, followed by the 38th meeting in Anaheim, CA, June 16–21, 2012. Please check for details on future meetings.

Dr. Stevenson is a Consultant and Editor for American Laboratory/Labcompare; e-mail: