The LCGC Blog: From Reversed Phase to HILIC and Back Again: Recent Evolutions in HPLC and UHPLC Stationary Phases
When I was an undergraduate student, I had the opportunity to work for a summer as a research assistant in Prof. Harold McNair’s group at Virginia Tech. I had no idea what I was going to be doing, besides something called “chromatography” — I actually only chose to work with McNair because he was the first to respond to my query about the availability of a position, and he offered to pay me the most. Little did I know how that decision would define my career trajectory. I remember when I joined the lab, the other members of the group were in a flurry preparing to deliver an American Chemical Society (ACS) high performance liquid chromatography (HPLC) short course to visiting industry professionals. I was to sit in on this course as a first introduction to chromatography and HPLC. I still have that course book and the notes I took in it. The course was about half lecture and half laboratory. In the laboratory, a series of experiments introduced beginners to concepts such as isocratic separations, gradient method development, and the effect of pH on separations. All of the experiments were performed on C18 reversed-phase columns. That made sense, because reversed-phase method development is for the most part straightforward. Assuming the analytes are in the right form in solution, and deleterious interactions with underlying solid supports are mitigated, separations based on hydrophobicity are fairly easy to manage. The course was immensely helpful for me as a beginner, and obviously this chromatography thing resonated well with me, so I returned to McNair’s lab to complete my PhD work after college. Funny enough, in the end, the topic of my dissertation was adduct ion formation in electrospray ionization, but that did not limit the exposure I received to the fundamentals of chromatography through my participation in and administration of the myriad different ACS short courses offered by the group.
The main point of that story is that when I began my independent academic career and considered doing some research on the fundamentals of HPLC separations, I found C18 reversed-phase separations to be quite boring. It would be quite difficult to publish a paper that described the elucidation of different retention behaviors for a set of analytes on a C18 column alone. That has been done. So, when I came across hydrophilic interaction liquid chromatography (HILIC) as an alternative separation mode that was gaining momentum, I became interested. HILIC provides a retention and separation mode for highly polar and hydrophilic analytes that are generally poorly retained on traditional C18 stationary phases. Popularity for the technique had mainly grown out of its use in pharmacodynamic ADMET (absorption, distribution, metabolism, excretion, and toxicity) studies and metabolomics. When a chemical compound is metabolized by the body, it is broken down into smaller pieces or appended with generally hydrophilic groups that allow the drug to be more easily processed and excreted by the body. This process usually results in compounds with decreased or no reversed-phase retention relative to the parent drug compound. HILIC can provide a means to analyze these important metabolic conversion products, but it is anything but straightforward when it comes to method development.
The mechanisms associated with HILIC separations can be widely varied, hard to predict, and quite complex; in fact, HILIC is very often referred to as a mixed-mode process. The fact that HILIC separations are most often performed in mobile phases having high acetonitrile–low aqueous content means that when the separations are coupled to electrospray ionization mass spectrometry detection, higher ionization efficiencies are achievable compared with those achieved with the high aqueous content in reversed-phase separations (1). Further, extracts eluted from a solid-phase extraction cartridge or tube in high acetonitrile content can be directly injected onto a HILIC column. If you asked a student about the mechanism of HILIC, they would most likely have trouble recounting it, but if they could then they would probably talk about the partitioning of analytes from the bulk polar mobile phase into a thin water layer associated with a polar stationary phase. In truth, while this mechanism may be a main contributor to retention and separation, there is also plenty of opportunity for direct adsorption by analytes onto the polar functional units on the HILIC phases. The fact there are so very many different analytes out there, as well as so very many different stationary phases (diol, amide, bare silica, cyclofructan, zwitterionic, and so forth) to choose from, means that it can be quite difficult to predict (or even interpret after the fact) how a particular selectivity was achieved. We found in some of our work that for a series of estrogen metabolites, their partitioning versus adsorption retention character varies considerably between different stationary phases and can even vary within a single stationary phase type (2). “Mixed mode” is certainly an apropos designation for HILIC.
As HILIC was drawing a lot of attention, so was the development and application of superficially porous particles as stationary phase supports. These supports, where the inner core of the particle is solid and the outer portion is porous, provide not only better efficiency HPLC separations, but also better stability under high pressures; they enabled the revolution referred to as ultrahigh-pressure liquid chromatography (UHPLC). Virtually every chromatography column manufacturer began manufacturing product lines associated with superficially porous particles. For our part, we were very happy with the performance provided by these materials. In one study we published in 2011, we used C18 and modified C18 phases arranged in series to achieve the baseline separation of four dansyl derivatized native estrogens (estrone, 17α-estradiol, 17β-estradiol, and estriol) (3). The run time was more than 50 min. In the years since, we were able to significantly improve the performance of this separation using a C18 phase attached to a superficially porous support. Run times were reduced to under 15 min to achieve baseline separation of the dansylated estrogens (4). Incidentally, this last piece of research also featured bulk derivatization (direct addition of the derivatization reagent to a sample without prior pretreatment) of the estrogens in cerebrospinal fluid and on-line sample preparation using restricted-access media to achieve low parts-per-trillion detection limits. The use of restricted-access media for on-line sample preparation is a blog topic I have written about previously (my first “LCGC Blog” entry ), and one for which we have also recently published a comprehensive review article (6).
Let’s get back to HILIC for a moment. Though its complexity makes it a rich subject for research, there are some significant shortcomings to its practicality. In reversed-phase separations, it is generally known that your sample should be dissolved in a mostly aqueous solution (comparable to a weak mobile phase), lest it might compromise peak shape; it is not a hard and fast rule, just a good thing to keep in mind. For HILIC, it can be disastrous if you do not match the sample solution to the initial weak mobile-phase composition. Injecting an aqueous solution (remember, water is the strong solvent in HILIC) onto a HILIC column can often generate a peak for an analyte that extends from the void volume through to the middle of the chromatogram. The analyte enters the column in a hydrated form, and simply does not achieve a consistent and tight band at the head of the column as a result. Additionally, re-equilibration times for HILIC can be unreasonably long between runs (10–15 min is not unheard of). For this reason, it is also preferable to avoid the use of extensive gradient mobile phase programs in a HILIC method; and then if a gradient method is needed, a good recommendation is to try to keep the ionic strength of the mobile phase consistent through the chromatographic run. This will shorten the re-equilibration times and will improve the reproducibility of the separations.
For these reasons, some manufacturers have branched out, expanded, and began to heavily promote some new reversed-phase stationary phase solutions that can reduce the need for HILIC to retain and separate more polar analytes. Instead, the phases rely on alternate functional units (for example, biphenyl groups) or functional groups embedded in an alkyl framework to improve selectivity. In some cases, a U-shaped retention profile can be observed for some analytes on these phases. A U-shaped profile means that as you vary the mobile-phase content from highly aqueous to highly organic (think of this on the x-axis), the retention of the analyte is found to be highest at the extremes (in other words, if the retention factor is on the y-axis, the plot of mobile-phase content versus retention factor is in the shape of a “U”). This basically means that such a phase can act in the reversed-phase mode at high levels of the aqueous mobile-phase component, and in the HILIC mode at high levels of the organic mobile-phase component. Thus, a broader range of selectivities can be achieved for a broader range of analytes; and if a few such phases are available, then perhaps a more universal solution to myriad separation problems can be achieved without resorting to HILIC. We are currently evaluating such a system in our lab and will report our findings in due course.
So, it seems my experience with different HPLC phases has, for the moment, come full circle. I do not think we are abandoning research on and use of HILIC phases, but some new alternatives sound promising. Moreover, when you begin to now combine advances in superficially porous media to achieve high efficiency with novel and predictable new stationary phases to achieve high selectivity, then it sounds like the evolution of HPLC and UHPLC phases over the past decade or more has brought some exciting new possibilities. I look forward to writing more about these developments in the future.
Previous blog entries from Kevin Schug: