HPLC 2012: New Mysteries in Column Efficiency

HPLC is the premier meeting series for liquid chromatographers to gather to learn about the latest advances in technology and applications. HPLC 2012—the 38th International Symposium on High Performance Liquid Phase Separations and Related Techniques—attracted 834 scientists and vendors to the Marriott Hotel in Anaheim, CA, June 16–21, 2012. Core/shell column technology, hydrophilic interaction liquid chromatography (HILIC) columns and separations, and supercritical fluid chromatography (SFC) drew the most attention at the lectures. The exhibition featured new pre-pump gradient systems from two vendors, plus a range of new column packings. A concise review of the exhibitors of new products can be found at http://goo.gl/WW7nu.

Column efficiency: An active topic and confusing picture

Until about three years ago, it seemed that column efficiency was a mature topic and technology. Core/shell particles were promoted as an alternative to ultrahigh-performance liquid chromatography (UHPLC) since they offered improved efficiency but at pressures compatible with existing (6000-psi) instruments. Core/shell columns entailed a small loss in capacity, but this was seldom an issue in analytical HPLC, since sample loads were usually submicrogram.

Then, last year at HPLC 2011 in Budapest, core/shell technology developers extended the particle size to the sub-2-μm range, expecting the smaller particles to provide greater efficiency due to the smaller size. This was observed, but the effect was twice the predicted improvement. To be sure, others did not see this effect, probably because of limitations in their instruments. However, both Waters (Milford, MA) and Agilent (Palo Alto, CA) introduced low-dispersion models of their UHPLCs (Waters ACQUITY UPLC® I-Class and Agilent 1290 Ultra Low Dispersion).

The improved efficiency was attributed to better radial heat conduction through the dense cores. Porous particles have much lower heat conduction than the solid spheres of the same size. Fast forward to 2012: Lectures and posters show that the improved efficiency, measured in reduced plate height, is not limited to sub-2s. Reduced plate heights, h, and even 10-μm particles could be as low as 1.5, even for particles as large as 10 μm. Columns packed with porous particles struggle to be as low as 2.0. Radial heat convention is probably a factor for the sub-2s, but Joule heating with 5 and especially 10 μm should be insignificant at normal flow velocity.

A lecture by Dr. Tivadar Farkas of Phenomenex (Torrance, CA) reported on the performance of experimental columns packed with 1.1-μm core/shell particles. As expected, the pressure drop increased 2.3 times compared to the company’s 1.7-μm commercial core/shells. However, height equivalent to a theoretical plate (HETP) vs linear velocity plots was flat as the linear velocity increased to above 2 mm/sec. At 4 mm/sec, the instrument reached the Pmax. This is not consistent with the expectations of strong frictional heating. Plus, the reduced plate height was nearly independent of column diameter (2.1, 3.2, and 4.6 mm), and 4.6 mm i.d. was marginally more efficient than 2.1 mm. Thus, frictional heating seems to play a small role.

Table 1 - Finalists selected from 444 posters presented at HPLC 2012*

So what is going on? The picture is not clear. Dr. Jack Kirkland (Advanced Materials Technologies, Wilmington, DE) attributed the efficiency gain to improved column packing technique. This is a difficult assertion to verify since the protocols for packing columns have many variables. These are optimized empirically for each column. However, a poster (poster 446, Table 1) and lecture from the laboratory of Prof. Ulrich Tallarek (University of Marburg, Germany) reported a computer-based comparative study of bed structure of poly- and monodisperse particle mixtures. They also used confocal microscopy to study the bed structure of capillary columns (this lecture won the Csaba Horváth Award).1 It was concluded that monodisperse particles might form beds with more uniformity, especially in the wall region, but that the exact structure depends on the packing protocol.

Pressure-driven flow

Pressure-driven flow through conventional (dp larger than 1 μm) packed beds for HPLC is described by Poiseuille flow. The flow profile is parabolic, with more rapid movement at the center, and no movement at the extreme of the column walls. This assumes that the column wall is not disrupted by channeling due to packing imperfections. The old standard was “a well-packed column has a reduced plate height of 2 times the particle diameter.”

Prof. Mary Wirth (Purdue University, West Lafayette, IN) described unexpected improvements in column efficiency with submicron-size particles due to slip flow (poster 87). Slip flow applies when the flow channel is so narrow that the flow velocity at the wall is not zero. Thus, the trailing segment at the wall is removed, which improves the flow profile in the column, giving more plates/meter. She claimed efficiencies of over 2 million plates/m for 2-cm-long columns in glass chips. Further, she reported a reduced plate height of 0.15, compared to 2.0 for Poiseuille flow. This reminds me of perfusion chromatography, which was a hot topic in the 1990s.

Improved column hardware after more than 40 years

Column hardware is possibly another contributor to low efficiency. Today’s column fittings all are derived from a 1960s design developed at Oak Ridge National Laboratory (Oak Ridge, TN). These used snubber unions from Swagelok® (Solon, OH), which included a sintered metal frit pressed into a union. The problem is that the flow path of the perimeter stream line is significantly longer than center-to-center. This difference produces tailing since the time to exit is longer on the perimeter. Poster 278 changes all this by introducing “curtain flow” whereby the inlet and outlet fittings are radially segmented to split the flow from the center and perimeter on the injection and collection end. The image in the outer flow forms a curtain around the center flow, which is the most efficient.

This seems similar to the infinite diameter effect introduced by Prof. John Knox (retired, University of Edinburgh, U.K.) in the 1980s, where improved efficiency was found for point injection on short, fat columns packed with small particles, which kept the analyte away from the dead zone near the column walls. The curtain flow columns will probably be introduced by Thermo Fisher Scientific (Runcorn, U.K.) since some of the coauthors are affiliated with the firm. Plus, selecting the center flow from the curtain flow columns improves compatibility with MS interfaces.

So why is column efficiency so important? The simple answer is that column performance drives instrument design. One might see radically different designs optimized for specific needs. Chip-based systems optimized for submicron might enable real-time assays with response times of less than 10 sec. Complex assays such as ’omics might benefit from long, very efficient columns optimized for ultrahigh peak capacity, speed, and, of course pressure. However, the current picture is not clear. I’m sure we will see more about improving column efficiency at HPLC 2013 in Amsterdam.

Hydrophilic interaction liquid chromatography

HILIC dominated the new product introduction for columns with several introductions. Dr. Andy Alpert, President of PolyLC (Columbia, MD) and the founder of HILIC and related ERLIC (electrostatic repulsion-hydrophilic interaction chromatography), presented a sunrise tutorial that filled the room with almost 250 people. HILIC attracts attention due to orthogonal selectivity to reversed-phase liquid chromatography (RPLC) and improved detection sensitivity in LC-MS. SFC-HILIC is also attracting attention since SFC is an extension of normal-phase separations mode. More about the new columns appears in the exhibition report (http://goo.gl/WW7nu).

Supercritical fluid chromatography

Applications of SFC are growing rapidly, particularly for chiral separations. However, chromatographers are finding that simply adding CO2 to the eluent decreases viscosity and improves separation efficiency (HETP). At high pressure, CO2 is a nonpolar liquid with very low viscosity. Since the operating temperature can be lower than the critical point, the question arises: Is this SFC? It may not be operating in the critical region, but the results speak for themselves. One does need a pump optimized for CO2 and a backpressure regulator to keep the mobile phase from forming bubbles before the detector.

The question of method validation for SFC often arises. Amindine Dispas used a quality-by-design approach to define the design space for an assay of eight neurotransmitters, including catechol amines (poster 345). The assay had previously been in the HILIC mode using a mobile phase containing more than 80% acetonitrile. The lab wanted a greener method. The initial screen involved four columns, three mobile phase modifiers, and three ion pairing additives. The most promising results were obtained with an ethylpyridine column eluted with CO2 + methanol spiked with 10 mM trifluoroacetic acid (TFA). A second set of runs delineated the design space. Test runs at the predicted optimum gave a 12-min run. Exploration of the limits of the design space showed coelution of at least two critical analytes at the periphery.2

With all the positive attention, one should be a bit cautious. CO2 is under high pressure in the column. The pressure drop is a function of its position in the column. Adiabatic expansion and associated cooling appear to produce some interesting chromatographic effects when operating near the critical point.3 Heating of the column inlet is reported to be necessary in at least some cases.

Poster awards

The active poster program with 444 contributions provided details and examples that complemented the research reports in the lectures. Agilent sponsored and organized a poster competition with $500 awards to each of 10 finalists. I was honored to be on the award jury. The finalists are listed in Table 1, along with a short review of their posters below.

Capillary columns

As mentioned above in the discussion on column efficiency, capillary columns packed with submicron-diameter colloidal silica have plate heights 10 times better than predicted (poster 87). This has been attributed to slip flow. Permeability is also increased despite a packing density of 0.28, which is very close to 0.26 for geometric close packing. Clearly, this new flow regime needs more attention.

Polystyrene/divinylbenzene (PS/DVB) stationary phases were prepared in 200-μm-i.d. capillaries. The columns were characterized with test probes ranging from 30-kDa proteins to 75-Da amino acids under UHPLC conditions. Short columns with high capacity were prepared with nanoglobules for rapid separations. Long columns with high porosity gave very high peak capacity (poster 240).

For several years, reports from the lab of Prof. Barry L. Karger (The Barnett Institute, Northeastern University, Boston, MA) have shown that porous layer open tubular (PLOT) columns are a powerful approach for ultratrace nanospray LC-MS. The columns are eluted at less than 20 nL/min, which improves ionization efficiency and transmission. Problems with ion suppression are also reduced. Use of a 3 m × 10 μm PLOT capillary and LTQ Orbitrap XL MS (Thermo Fisher Scientific) provided a 10-zmol detection limit and useful quantitative results over the range 50 zmol–100 amol (poster 190).

More efficient column hardware

As discussed in the section on improved column efficiency, typical band profiles with pumped mobile phase eluting packed bed columns are parabolic with the center moving faster than the walls. The authors described how new column hardware routes the center flow to a detector (such as an MS) while the slower mobile phase near the wall can be ignored (poster 278).

Optimizing peak capacity for small molecules

For several years now, the laboratory of Dr. Peter Carr (University of Minnesota, Minneapolis) has been exploring the theoretical limits of high-resolution chromatography in one and two dimensions for large and small molecules. The optimization approaches are different, as are the conclusions. The authors presented a protocol for optimizing peak capacity for i.d. gradient elution separations of small molecules, starting with the restricted set of commercially available column dimensions. A general strategy is to maximize the peak capacity for particular column dimensions since this minimizes the risk of coelution. A previous study for very large molecules (proteins) predicted that peak capacity increases monotonically with run time. However, for small molecules (<1000 Da) there is a maximum and then decrease with increasing run time (poster 437).

Biopharmaceutical analysis

Both the European Medicines Agency (EMEA) and the FDA require assay of topoisomers of plasmids for gene therapy and vaccines. The authors developed an on-line 2-D HPLC assay (poster 53). The first dimension removed extraneous material from the fermentation with size exclusion chromatography. The DNA fractions were selected by heartcutting. These fractions were separated on a laboratory-produced quinine carbamate anion exchanger. This system was also used to monitor topoisomers during fermentation. Equilibrium composition was achieved after about 13–15 hr. This new assay is expected to aid plasmid research and QC (poster 53).

Modeling leads to improved MS interface

Slip flow, described above, illustrates that the behavior of miniaturized microfluidic devices should not be viewed as a linear scaledown from the macroscale objects. Modeling is an essential research tool. Poster 189 described a novel miniaturized liquid junction interface for nanospray chromatography. Modeling played a key role in development.

Sample prep

The authors prepared two selective sorbents targeting phosphopeptides to improve MS detection. One immobilized iron nanoparticles via photografting to a methacrylate monolith. The second added hydroxyapatite nanoparticles to the methacrylate monomers prior to initiation of the polymerization. Both were effective in selective extraction of phosphopeptides, but showed different, often complementary, selectivity (poster 403).

Immobilized enzyme bioreactors

LC column technology can spill over into new applications. For example, poster 239 reported on an enzymatic reactor with immobilized lipase on a monolithic polymer support that was used to catalyze the transesterification of triacylglycerides into the fatty acid methyl esters suitable for biodiesel. The reactor maintained activity during processing of 1000 reactor volumes of substrate (soybean oil).

Martin Gold Medal and Silver Jubilee Medal

The Chromatographic Society of the U.K. awards the Martin Gold Medal annually to a scientist who has made an outstanding contribution to chromatography. Prof. Ed Yeung (retired) certainly deserves the medal for his work in developing detection technology in HPLC.

The Chromatographic Society presented the Silver Jubilee Medal to Dr. Monika Ditman (Agilent Technologies, Waldbronn, Germany) in recognition of her work in advancing separation science in Europe, particularly in capillary electrophoresis and capillary electrochromatography.

HPLC 2013

The next two meetings in the HPLC Series are scheduled to be held June 16–20, 2013 in Amsterdam, The Netherlands, and November 18–21, 2013 in Hobart, Australia. Please monitor the web sites for more information.

References

    1. Bruns, S.; Tallarek, U. Slurry Packing Parameters and Their Influence on Capillary Column Morphology; lecture-03-05, HPLC 2012.
    2. Dispas, A.; Lebrun, P. et al. Innovative Green Supercritical Fluid Chromatography Development for the Separation of Neurotransmitters Using Design of Experiments Design Space Strategy; poster P-345, HPLC 2012.
    3. Poe, D.; Ranger, M. et al. Pressure, Temperature and Density Drops Along Supercritical Fluid Chromatography Columns; poster 318, HPLC 2012.

Robert L. Stevenson, Ph.D., is a Consultant and Editor of Separation Science for American Laboratory/Labcompare; e-mail: [email protected].