UHPLC’s Seven Year Itch: Quo Vadis?

At Pittcon® in 2004, Waters (Milford, MA) introduced the ACQUITY UPLC® chromatograph with a 15,000-psi Pmax, and the rest of the LC world had to play an intense game of catch-up. This included creating fear, uncertainty, and doubt (FUD) by the competition as they struggled to respond. But, by the spring of 2010, each of the top seven vendors had introduced competitive instruments, some with a Pmax greater than 5000 psi. Users responded that the benefits were real and often more valuable than expected. Many sales arguments were properly swept to the dust bin. So after seven years, let’s see what UPLC has brought and what might be coming in the future.

First let’s look at the benefits users realize from ultrahigh-performance liquid chromatography (UHPLC) technology. Speed of analysis and associated direct improvement in throughput are conspicuous and expected attributes, but there are more. Detection sensitivity is improved several-fold due to reduced peak width and reduction in the diameter of the standard column from 4 mm to 2 mm or even narrower. Most early adopters did not anticipate that improved speed would bring dramatic increases in productivity arising from improvements in work flow. Customers found that, in many situations, having results in near real time helps support decision-making, also in real time. This can be critical on the production floor. Waters was quick to follow ACQUITY® for the laboratory with ACQUITY PATROL™ for the process floor. Although adoption of new technology and work flows in process settings is much slower, PATROL chromatographs are rapidly attracting a following, particularly supporting production of pharmaceuticals. No doubt the FDA’s initiative on Process Analytical Technology (PAT) is helping.

A common concern is: Can legacy methods transistion to UHPLC chromatographs? The answer is “It depends.” If the method involves isocratic elution, the flow rate is simply reduced by the ratio of the column diameters. This keeps the linear velocity the same. Run times are reduced by the ratio of the column lengths. One should also check on the ability of the data system to handle the narrow peaks by comparing the resolution of critical pairs. In most cases, all of this is very simple. It may not require much more than constructing a new retention time table. If more confidence is needed, one can run both instruments in parallel for perhaps a month and compare the results. They should agree.

For gradient elution chromatography, the system dwell volumes are much different. For example, Agilent (Wilmington, DE) introduced the Jet Weaver mixer with dwell volume from 10 to 35 μL, depending on the mixing element. This dwell (mixing) volume is much smaller than in the 1100 series LCs, so gradient slopes are much steeper than with conventional (1100) mixers. Most vendors have worked out computerized conversion protocols that take care of converting the flow rate settings. DryLab® (Molnar Institute, Berlin, Germany) also extends this to include robustness evaluations.

However, if the existing method is poorly constructed, the changes may be so large that a new validation study may be required. Indeed, Agilent has talked about bridging packages for the 1260 and 1290 to step backwards to older instruments. For instance, if one has a validated method on an Agilent 1100 series instrument, the 1290 can be “detuned” to give equal performance to the 1100. One can then get comparable results (retention time and resolution of critical pairs) with the new and old instruments. With the bridging from 1100 to 1290 instrument platforms in place, it is easier to revise the methods to take advantage of the improved performance potential of the 1290—or, even easier, just use a gradient hold to decrease the steepness. This seems to work quite well, at least in some cases.

Another question is: Do all modes in conventional HPLC translate into corresponding UHPLC? Reversed-phase LC (RPLC) gets a resounding “yes.” Since RPLC is about two-thirds of the market, this is a big base to start with. Normal-phase LC, including hydrophilic interaction liquid chromatography (HILIC), also appears to translate. The answer is less clearly demonstrated for ion exchange and steric exclusion chromatography. Conventional ion chromatography should work, but there is more of a question about the reagent-free mode, where the membranes are subjected to high pressure. Mechanical integrity of the membrane is one issue; plus the pressure could induce osmotic flow to the high-concentration side, which would need to be controlled.

Sample compatibility with ultrahigh pressure is another issue. Ultrahigh pressure is not much concern for small molecules, but proteins and aggregates are known to change structure with very high pressure.1 Protein aggregates can be disrupted by hydrostatic pressure as low as 2000 bar. Disaggregation may even start at lower pressure for some. Further, protein folding is also known to be pressure sensitive. High hydrostatic pressure unfolds and thus exposes the hydrophobic patches that are responsible for tertiary structure. An RPLC packing, with avidity for hydrophobic regions, could shift the thermodynamics of the unfolding process. Worse, this could change as a protein transits down the column. The highest pressure is at the column inlet. If the pressure is high enough, the protein might start to unfold, exposing usually inaccessible hydrophobic patches for interaction with the stationary phase, but near the exit, the pressure is much less, which could allow the protein to refold, partially or totally. This transition of the form of the analyte along the column would certainly give peaks with a nonideal shape. In a quick search online, I could not find any report on the effect of high-pressure mobile phases on protein structure during chromatography. However, if high pressure turns out to be a real problem for separation of large proteins, perhaps field flow fractionation (FFF) can fill the void.

Other factors

Short column lifetime is often expressed as a concern. Columns for UHPLC are usually priced at 20% or so higher than the corresponding columns for a legacy instrument. Yes, this compares a 2.1-mm for UHPLC with a similar phase in a 4-mm-diameter column. Early war stories from UPLC detractors reported that UPLC columns did not last as long as conventional technology. Waters countered that this may be true occasionally, but during their useful life, UPLC columns ran many more samples due to the short run time. So, both can be true. My limited interviews with UHPLC users indicate that column lifetime is seldom more of a problem with UHPLC.

If short column lifetime should be an issue for a sample, several vendors suggest filtering the mobile phase and the sample through 0.2-μm filters. This will remove the particulates that are usually responsible for plugging the column. Most vendors have kits for the mobile phase as well as sample filtration.

Fittings are another consideration. Initially, the fittings were a very weak point in the UPLC chromatograph. The ferrules could slip, which let the tubing pull back or in extreme cases pull out of the fitting. Fortunately, the latter was rare, but ferrule creep did occur. The loss in column efficiency was easy to recognize and fix. Now, several firms have solved the problem with new designs. Dionex (Santa Clara, CA) introduced the Viper™ fitting, which provides a reliable connection by design, every time. Optimize Technologies (Oregon City, OR) has a novel double ferrule that bites into the tubing wall even with only finger-tightening.

Shimadzu Scientific Instruments (Columbia, MD) introduced a novel omega flow path in the HPLC injector of the Nexera. This keeps the sealing surfaces of the injection valve flat. In contrast, conventional designs have about one-third of the rotor seals at system pressure, but the remainder is at ambient. This could tilt the seals and increase wear, and ultimately leaks.