Debating Volume-Based HPLC

An article by Monika Dittmann on volume-based HPLC (tas.txp.to/1013/volumeLC) triggered a debate on the practicality of the concept. Here, we present the questions that were raised and how they were answered.

Volume-Based HPLC Isn’t Practical

Posted online by “Chris” an R&D Director/Manager in the USA

In real life [volume-based HPLC] couldn’t work. Since a column invariably increases in pressure with use, the constant pressure mode will translate to drifting flow-rate mode.

Imagine for example, that you have a column operating at 7000 psi at the beginning of a gradient separation. In this example, let’s say that the column flow rate at the start of the separation was 1 ml per minute. Now, imagine that, after injecting real samples for a week, the flow rate at the beginning of the separation drops to 0.9 ml per minute when the pressure is set at 7000 psi, due to the accumulation of particulate at the head of the column from the samples injected. Surely there’s a problem when the flow rate has significantly changed from the conditions originally used the week prior? If one were to stick with constant pressure mode, the flow rate would be slowly dropping over time as long as the pressure is held constant. To put it another way, when we operate at constant flow rate we are used to seeing the pressure increase slowly over time on a given column, so if we choose to operate at constant pressure we can expect to see the flow rate slowly decrease over time. The difference is that nothing significant changes about the analytical method when we work at constant flow rate but the method will be significantly different in terms of speed as the flow rate slowly drops. This is the issue that concerns me about this proposed mode.

Even if there were good tools for converting an elution time plot into a volume-based plot, which would partially obscure the issue, it doesn’t address the underlying issue of gradually decreasing flow rate over time as the pressure on the head of the column increases with use. Because the actual analysis time is flow rate dependent in the constant pressure mode this would mean that the analysis time would be continuously increasing over the life of the column. This would have two significant impacts: (1) the analysis time would be significantly increased, decreasing sample throughput, and (2) the chromatographic efficiency of the separation would change over time, changing resolution and in some cases even elution order, since elution order can be mobile phase composition dependent.

Because gradient programs are time-based not volume-based (unless you change that to a volume-based program too) there would be issues with reproducing gradient chromatography over time. I guess it’s possible that instrument manufactures might choose to implement such a tool, but it’s hard to see why they would try to do so unless they had an online flow meter to enable accurate plotting of the chromatogram on a volumetric basis. While it is true that one could theoretically get away with a time-to-volume conversion without a flow rate measurement, the accuracy of such an approach is questionable. Frankly, I think it’s extremely unlikely that any instrument manufacturer would undertake such an effort for such a minimal benefit. Furthermore, it’s hard to imagine that anyone would consider this as a viable option in a real analytical lab.

Volume-based HPLC Works

Konstantin Choikhet, Research and Development Chemist at Agilent Technologies, Boeblingen, Germany, and one of the inventors of Volume-based HPLC, responds.

Chris’ comment gave me pause to stop and think about how volume-based chromatography is perceived in the field. Here, I discuss the concept, how it could look and feel for a user, and what constraints and benefits one could face using these approaches.

As discussed in a number of publications and conference presentations that have referenced “constant pressure” or “volume-based” chromatography (1-3), the essential parameter that governs retention in chromatography is the volume of mobile phase passed through the column. Time-based description of a chromatographic process (although widely used) is in fact a special case, which is only adequately applicable if the flow rate is constant. Although this is convenient for a number of reasons, it also causes significant limitations.

It has been shown that chromato-graphically-consistent results can also be obtained without strict control over the flow rate (2, 3, 5). If the eluent composition plotted against delivered volume (the gradient program) remains unchanged, the plot of the detector signal versus delivered volume (the chromatographic output) will also stay essentially unchanged, nearly independent of how the flow rate was changing during the separation.

First, a few definitions: “Volume-Based (VB) LC” means executing a gradient program in accordance to the delivered volume (rather than to elapsed time) and handling the chromatographic output over an X-axis representing eluent volume. “Constant Pressure mode” is a special implementation of the VB approach. In this mode the pump keeps system pressure constant by continuously adjusting the flow rate. As with any innovation that changes established processes, the VB approach faces certain reservations from potential users. I will assess how justified those reservations are by answering a number of frequently asked questions.

Do I need to change my methods?

VB mode is not a replacement, but is rather an add-on to the common operational modes. An instrument capable of VB operation can also be used in constant flow regime but gives the user the option to run an existing method in constant pressure mode.  Furthermore, VB-operation might not even be perceived as a method change by the majority of users. For example, your conventional method with its flow rate and composition time-table are displayed, but now an additional parameter called “execution pressure” and a checkbox “optimized throughput” are available. You set the execution pressure as you like, say 1150 bar, and tick the checkbox. That’s it! The instrument will take care of the rest.

What are the benefits?

The benefits of using constant pressure mode are the elimination of overpressure shutdowns (you have defined an execution pressure that will be actively maintained and thus never exceeded), throughput increase by 10-25 percent due to the more efficient use of the available power range, also the column stress is lowered by eliminating gradient pressure cycles.

What would my chromatogram look like?

You could let it be plotted with milliliters at the X-axis, if you like. But you might also want to make it look more familiar, in which case, the evaluation software would convert the volume X-axis to a time X-axis corresponding to the flow rate of your original method. The chromatogram would then match the one you got running your method conventionally. Not only is there similar reproducibility between modes, but also an excellent coincidence of retention volumes and peak areas (3).

Is the transition to VB operation entirely seamless?

Not exactly. There would still be some differences to the conventional constant flow method execution, which should be taken into consideration. These differences are:

• If any action outside the instrument needs to be synchronized to certain phases of a separation (a possible situation in multi-vendor systems with only rudimentary communication between system parts), the transition might be challenging, because the duration of every single separation and separation phase might vary depending, for example, on variations in system permeability.
• Due to flow rate changes, the efficiency for some peaks might be increased or decreased compared to the constant flow mode, depending on where your original separation was on the van-Deemter curve.
• As already understood from method transfer from HPLC to UHPLC, selectivity changes could occur for substances with strongly pressure-dependent retention coefficients (4).

And what about quantitation?

If the detector you use is composition-sensitive (for example, UV or fluorescence) the converted chromato-gram can be integrated exactly as the one run with the original method and the integration result will be essentially the same. Quantitation with mass-sensitive detectors (optimized LC-MS interfaces under certain conditions) is also straightforward. Detectors that are not purely concentration- or mass-sensitive are more challenging; that discussion exceeds the scope of this commentary but has been covered in the literature (5).

Will the approach be broadly accepted?

We will see if the VB approach finds its way into a broad user community, but it seems attractive enough to give it a try.