HPLC Column Expert Predictions - Revisited

August 1, 2007

LCGC North America

LCGC North America, LCGC North America-08-01-2007, Volume 25, Issue 8
Page Number: 692–702

Columns | <b>Column: Column Watch</b>

In this installment of "Column Watch" columnist Ron Majors revisits expert predictions from a survey conducted in 1987. A cross section of column experts of the time were asked a series of questions on the future directions in high performance liquid chromatography (HPLC) column technology. Now, 20 years later, these predictions are contrasted against current column technology. In many cases, the experts were entirely correct, while in other cases, they were dead wrong. Some current trends were not even considered 20 years ago. The author backs up his analysis with current survey information.

While preparing my lecture for a session honoring Dr. Maurits Verzele, professor emeritus of Ghent University (Ghent, Belgium), at the recent HPLC 2007, I had the occasion to look back at Professor Verzele's illustrious career in gas chromatography (GC) and high performance liquid chromatography (HPLC). Verzele made many contributions to HPLC column technology, especially in the areas of stationary phases, column packing techniques, and preparative chromatography. Two decades ago, I published an article in LCGC Magazine entitled "The Future of HPLC Column Technology: A Survey of Experts" (1) for which I interviewed Verzele and a host of other column experts at the time. The experts, listed in Table I, were selected from a cross section of academia, industry and column manufacturers. Some of the experts are no longer with us, some are retired or winding down on their careers, some have left the field, and some I have no idea of their whereabouts. Nevertheless, at the time, these scientists were some of the leading thinkers in HPLC technology and expressed their opinions on where the technology was heading. A series of questions were asked of each expert and I summarized their inputs in my 1987 article (1). Some of the experts were right on with their predictions while others were completely wrong (at least so far). I thought it would be interesting to look back at their thoughts and contrast them with where HPLC column technology is today or where it might still be heading. I will use some of the information gathered from recent surveys to confirm today's status. I will not refer to an individual expert's predictions but will look at their collective wisdom (but you can always go to the original article if interested [1]).

The Status of HPLC Column Technology in 1987

Back in the 1980s, LCGC Magazine frequently surveyed readers to get a general profile on users and to note technology trends. In late 1987, an HPLC column usage survey of readers was conducted; the results were published in early 1988 (2). At that time, reversed-phase chromatography was firmly established as the dominant mode of HPLC with use of C18 (octadecylsilane) phases far ahead of everything else; C8 (octylsilane) was a distant second. Then, ion exchange–ion chromatography was the second most popular mode with separation and purification of biological compounds the driving force in HPLC applications. Silica-based columns were the predominant media but there appeared to be interest in the use of polymeric columns for reversed-phase, ion-exchange, and size separations. Their ruggedness and wide pH range were cited as the advantages over silica-based phases. Chiral separations were merely noise on the baseline. Spherical silica particles were favored over irregular particles 3:1. In particle sizes for analytical separations, the use of 5-μm particles was now in the majority but many users still preferred 10-μm particles, probably because many routine methods were based upon these larger particles. The 3-μm particles were preferred by less than 5% of the workers. In preparative work, 10–15 μm particles were favored but use of 5-μm particles was increasing. In terms of analytical dimensions, 25 cm × 0.46 cm was the most popular size but smaller internal diameter columns were gaining ground, especially 2-mm i.d. columns. The so-called "fast LC" columns (short 50-mm columns packed with 3-μm particles) were discussed at the chromatography meetings but had gained little support in the applications world.

Ronald E. Majors

An interesting observation was the interest displayed in the 3-μm particles and sometimes even lower particle sizes in the mid-1980s. At the chromatography meetings at the time, such as the HPLC series and the European chromatography series, small diameter microbore (1.0–2.0 mm i.d.), capillary (0.3–0.9 mm i.d.), and even microcolumns often were talked about, yet in practice, surveys showed that these smaller dimensions had not established a foothold. In the author's personal opinion, it takes at least a decade for the leading edge technology to get transferred to the practicing laboratory where it becomes a routine technique.

Table I: HPLC column experts surveyed in 1987

Where the Experts Agreed

Table II summarizes the predictions where the experts' opinions turned out to be right on the mark. Experts predictions of reversed-phase chromatography and C18 maintaining their stronghold did not take much of a crystal ball to predict nor did the prophecy that silica-based columns would be competitive long into the future. For example, our recent HPLC columns' survey (3) showed that reversed-phase chromatography still dominates the application of HPLC, with C18 being the most popular phase type, although C8 is creeping up into a strong second place. Figure 1 shows that at Pittcon 2007, silica-based packings still command the most attention by column manufacturers based upon demands from the marketplace. The experts' prediction that smaller particles in shorter conventional columns (less than 50 mm long) will see growth in fast LC was correct but it took nearly 20 years for them to become more commonplace. The needs for increased productivity and high-throughput analytical results in the new millennium drove users to look to shorter columns as one of the answers. Short columns with small particles down to sub-2-μm average dimensions are becoming in vogue (4).

Figure 1

An interesting side note: Professor Verzele and his colleague Chris DeWaele published a sub-1-min separation on a 2-μm bonded phase silica column way back in 1983 (5), which is reproduced alongside a recent separation on a 1.8-μm column in Figure 2. Although the sample and conditions are somewhat different in the two examples, fast, sub-1-min separations were clearly demonstrated nearly 25 years ago. However, instrumentation improvements, better particle stability, and improved column hardware designs have made such separations more easily accomplished with longer column life than in Verzele's day. Nevertheless, younger chromatographers sometimes do not appreciate the efforts of the "founding fathers" in establishing the groundwork for what we take for granted today.

Table II: Experts predictions that were accurate (1987)

Another prediction that came to pass was the disbelief that microbore columns would see a great growth in the years following the 1987 survey (1). There was a great deal of investigative work going on with microbore LC columns in the mid-1980s but, at the time, the user needs for smaller dimensions had not yet developed. The 2-mm columns only became popular when requirements of mass spectrometry (MS) dictated flow rates lower than 1 mL/min. Using these smaller internal diameter columns, users could get nearly the same separation as their 4.6-mm i.d. columns would give them but at flow rates that were reduced proportionally as the column inverse radius ratios squared. For example, a 2.1-mm i.d. column running at 0.4 mL/min gave separations equivalent to a 4.6-mm column running around 2 mL/min, a popular flow rate. Electrospray ionization sources could handle the lower flow rate. In addition, when lesser amounts of sample were available, as was encountered often in trace analysis of biological compounds, column internal diameters of less than 4.6 mm gave better sensitivity. Nowadays, microbore and capillary columns are more commonplace; even columns below 100-μm in diameter have found use in proteomics applications in which sample availability, analyte concentrations, and MS requirements dictate such column dimensions.

In the early days, many investigators thought that there would not be the proliferation of stationary phases that had plagued GC. The reason for this belief was that LC had an additional experimental parameter available that GC did not have — the mobile phase. In GC, there is little or no interaction between the carrier gas and the stationary phase. The carrier gas serves to move the analytes down the column as they volatilize at temperatures above their boiling points. Contrast this observation to that of LC in which mobile phase pH, ionic strength, solvent strength, and so on have dramatic effects on analyte retention and selectivity. Thus, liquid chromatographers felt that a few stationary phases would be sufficient to separate most samples and the addition of modifiers and other additives would be enough to separate nearly every analyte encountered. Well, needless to say, the experts were entirely correct. In the last two decades, there have been hundreds of columns introduced to the buying public. I have been tracking the new column introductions since Pittcon 1985. Figure 3 shows the cumulative introductions through Pittcon 2007, a total of 1884 columns. This number added to the hundreds of columns that had been introduced from the early days of HPLC in the late 1960s to 1984. No wonder the newly initiated liquid chromatographer has a difficult time choosing the appropriate column from this vast number on the market.

Figure 2

Along with the proliferation of common phases like C18 and strong anion exchange, there also has been the development of many specialty phases. Because many of the HPLC phases are silica-based, there has always been a concern about the separation of basic compounds without the presence of tailing. Tailing is a result of the interaction of unreacted surface silanols (sometimes called "residual silanols") with the basic functional groups on these analytes. Under the worst conditions, at pH values around the pKa of the silanols (~4.5–4.7), ionic interactions can occur with protonated analytes so that tailing can be quite significant. The experts predicted that specialty silica-based columns would be developed to handle these basic substances. Indeed, reversed-phase chromatography columns called "base-deactivated" or "inert" have appeared over the years to deal with basic compounds, many of which are pharmaceuticals. Not only these type of columns but hundreds of other specialty columns applied to a variety of compounds ranging from proteins to metal-chelating organics to environmental (for example, polynuclear aromatics) to buckyballs to enantiomers have been introduced, dwarfing the total number of GC phases introduced since the beginning of time.

In 1987, the advent of the science of biotechnology in the U.S. was bringing more attention to the chromatography of biomolecules. Although low-pressure polydextran and agarose phases were well accepted in the life science community, these materials could not withstand the high pressures required for HPLC so workers began the search for more biocompatible phases. The experts noted this in the survey and indeed in the remaining 1980s until today, column investigators have been developing phases that could provide rapid, efficient separations of proteins, peptides, and other biomolecules without nonspecific interactions causing low recovery. Indeed, the hydrophilic, deactivated silica-based packings and perfusion packings of the late 1980s were some of the first high-performance biocompatible materials, and now poroshells, polymeric monoliths, and gigaporous hydrophilic polymers have been introduced since then. giving the biochemists an arsenal of column choices.

Table III: Experts predictions that were dead wrong (at least so far) (1987)

Column life of silica-based packings has always been a concern. On the low-pH side, the hydrolysis of the siloxane-bonded phase and, on the high-pH side, the dissolution of the underlying silica, have plagued chromatographers. The experts predicted that alternate support materials would come about to solve these column-killing problems. At one time, due to the absence of silanol groups, polymeric packings were thought to solve this dilemma but their acceptance in reversed-phase applications has been hampered by their markedly lower efficiency, sometimes as low as a third of the silica-based materials. Since 1987, newer materials have surfaced that appeared to have some favorable properties relative to silica; some of these were accepted while others never received widespread attention. Graphitized carbon, phases bonded–coated onto zirconia and titania, polymer-coated silica, bidentate bonded silica phases, and organic-silica hybrids all have been brought to market to address the bonded silica deficiencies.

Finally, the experts predicted that LC would move into the preparative and the process LC world. Over the years, as the need for a larger amount of purified substances has emerged, LC has indeed found its way into the preparative laboratory and into the process environment, often as a last resort in the latter case. Preparative separations can be scaled up from analytical separations quite easily because most column manufacturers develop a family of products whose chemistries are quite linear. Using mathematical models, these results can be quite predictable. High-pressure preparative instrumentation with flow rate capabilities into the hundreds of milliliters per minute has come onto the marketplace. Along with the growth of preparative HPLC, the growth of preparative supercritical fluid chromatography (SFC) also has come, in which the mobile phase is often supercritical carbon dioxide, easily removed from collected fractions because it is a gas at atmospheric pressure. SFC in the preparative fractionation of single enantiomeric pharmaceuticals has been fairly successful. Removal of HPLC solvent from preparative fractions can sometimes present a challenge because evaporation of solvent can leave trace impurities in the pure compound. In process LC, specialty companies with an engineering flair have worked with manufacturing engineers to scale up production to kilogram quantities. Simulated moving bed chromatography (7) has found great use in the process environment. Still, nonchromatographic techniques are often a first choice for large-scale process work, provided the analyte purity and recovery requirements can be achieved.

Figure 3

Where the Experts Were Wrong

Table III provides a brief listing of the more notable areas where the experts' predictions failed to materialize. As with many technologies, chromatography has "fads" that are hot for a time then tend to fade into the sunset with few practical applications demonstrated. A few examples that come to mind: liquid–liquid chromatography, dry-packing of microparticulate particles, micellar chromatography, nonporous silica, and capillary electrochromatography (CEC). Our experts also had a few "bloopers." Because this survey involved the future of HPLC columns, most of the comments were directed to this part of the chromatography solution. One of the hot buttons at the time of the survey was the variety of incompatible column endfittings. There was no standard endfitting design and a column from one manufacturer would not always fit the nuts and ferrules of another manufacturer, resulting in a mismatch. One expert predicted that the endfitting dilemma would be taken care of by manufacturers agreeing to standardize. In retrospect, this standardization never happened, although the advent of PEEK ferrules that came more recently overcame some of the mismatch problems.

At the time of the survey, microbore columns were under widespread study and one of the experts predicted that 50 mm × 1 mm would become the standard column dimensions. From our recent study, this is far from the truth. The 4.6-mm columns have maintained their dominance and smaller internal diameter columns have come along as specialty columns for more dedicated applications such as LC–MS or proteomics. Radial flow columns were predicted by another expert to become as popular as standard packed columns. Radial flow columns consist of two concentric porous cylindrical frits between which the stationary phase is packed. Solvent flows radially across the columns, not longitudinally down the column. Because the column length is dictated by the gap between the two cylinders, such columns are only useful for high-flow rate preparative separations of easily separated substances but have never really been considered seriously for analytical use. Conventional packed microparticulate columns are still the mainstay but some radial flow monolith columns have been shown to be useful for the preparative separation of large biomolecules.

One expert predicted that open-tubular columns would see increased use. In GC, open-tubular columns are, by far, the most popular, and theoreticians have predicted that extremely small diameter, open-tubular columns (perhaps in the 5- or 10-μm i.d. range) could be equally as popular in HPLC. So far, the open-tubular LC columns have been only research curiosities with few practical papers published on their successful use. That might change with some recent work reported by Professor Barry Karger of Northeastern University, Boston, Massachusetts, at the HPLC 2007 meeting in Ghent, Belgium. During his plenary lecture (8), Karger described his work in the area of multidimensional LC–LC–MS-MS where, using a 10-μm i.d. polystyrene–divinylbenzene (PS-DVB) porous layer open tubular column as the last column in an on-line serial column configuration reversed phase → strong cation exchange → PS-DVB, with electron transfer dissociation MS, he was able to identify more peptides than earlier work using less than 5 μg of an in-gel digested protein. So time will tell if this expert prediction will become more commonplace, at least in the proteomics laboratory.

Several experts predicted that cartridge column systems would prevail over the traditional compression-fitting type of column hardware. The original idea was that once the user purchased the column housing, the individual column costs would be considerably lower. In addition, the column could be replaced without disconnecting the endfittings. In some of the later designs, the guard column was integrated. Although cartridge systems were popular in the late 1980s and 1990s, today, although still available, they are used to a lesser degree than the compression-fitting columns. Early cartridge columns were plagued with leaking problems and, even though later versions were more robust, users have a long memory. In addition, replacement cartridge columns were never sold at a significantly lower price than conventional columns. So, this prediction never really came to pass. New designs of compression-fitting columns that will withstand high pressure with minimal extracolumn effects have become available, solidifying the dominance of this column hardware approach.

In the late 1980s, small-particle nonporous media became available. The idea was that if the stationary phase on a solid, impervious particle was very thin and the particle size was very small, rapid separations could be achieved with excellent efficiency, particularly for large biomolecules that, by their very nature, had much lower diffusion coefficients in solution than small molecules. Although the stationary phase coverage was quite low, these particles could be used for analytical purposes if care was taken not to overload them. Their small particle diameter (approximately 1.5 μm) gave rise to large pressure drops but, because relatively short columns were used, 20 or 30 mm, pressure was not overwhelming. One big problem at the time was the need to have frit materials that would not allow these tiny particles to pass through yet would not be clogged by particulates and other debris that might happen to be in the mobile phase coming from the pumps and injector. Experts predicted that a majority of future biological applications would be performed on these media. The nonporous media never became mainstay separation products for either biological compounds or small molecules.

Biochromatography was just developing in the 1980s and much attention was devoted to better media. Affinity chromatography was an area of active development, especially by those experts with a biological background. Affinity was a popular purification technique in the days of "soft gels" and naturally high-pressure versions were being explored. Predictions were made by the experts that affinity chromatography would be a dominating HPLC mode for biomolecules. It was predicted by one expert that user-derivatized high-pressure affinity phases would be commonplace. Although high-pressure affinity chromatography is one of the modes for the separation of biological compounds, it never became the dominating method but is used in preparative work in conjunction with other HPLC modes.

At the time of the expert survey, polymeric columns were coming into their own, especially for reversed-phase chromatography applications. As stated earlier, silica-based columns were dominating HPLC applications. Polymeric columns were being pushed strongly by a number of companies who had polymer expertise. One of the experts even predicted that sales of polymeric columns would increase substantially faster that silica-based columns, mainly because of the need for greater chemical stability. As Figure 1 showed and the most recent column survey confirmed (3), polymeric columns have not become the dominant force in HPLC column technology but have established their niche, mainly in ion-exchange and size-exclusion applications.

Some Things Never Change

One area where the 1987 survey (1) and our most recent survey (3) agreed evolved around the question of the biggest current problem with HPLC columns. Both the 1987 experts and the recent survey showed that column-to-column reproducibility and column stability were the biggest challenges facing chromatography users. I believe that this feeling is more myth than reality. In the last two decades, column reproducibility has dramatically improved due to following drivers: ISO, GLP, six sigma and other quality initiatives; competitive pressures; a better understanding of manufacturing processes; and the development of more rugged, robust methods built around regulatory requirements (ICH, FDA, USP). As far as column stability, the recent survey (3) showed that column life has increased dramatically since the early 1980s for reasons stated in the reference. In the 2007 survey, 50% of the users had column lifetimes in excess of 10 months, 2.5 times the expectations of the experts in 1987 (1).

In the 1987 expert survey, inadequate sample preparation was a barrier to HPLC columns; not only in their expected life, but efficient sample preparation was expected to simplify the overall separation requirements. Today, sample preparation is still a strong prerequisite to successful chromatography, and work is going on to provide safer, more efficient, and faster sample preparation methods. Experts predicted that the use of multidimensional chromatography would be required to successfully handle complex sample mixtures, but multidimensional LC and comprehensive LC × LC has been seriously addressed only recently, mainly because of the very complex mixtures now being encountered in proteomics (thousands of proteins and tens of thousands of peptides), natural products (such as traditional Chinese medicines), and food safety.

Conclusions

The 1987 experts predicted some of the modern HPLC column trends that are commonplace today and overall did a decent job predicting where column technology eventually ended up. Some of today's hot topics were not even envisioned in 1987 such as monoliths, inorganic–organic hybrids, ultrahigh-pressure HPLC columns and ultrahigh-pressure instruments, and chip-based and nanoLC columns, along with nano instruments. Other topics that are of interest today such as 2-μm particles and microbore columns were addressed in the 1980s but didn't become popular until much later as the user needs arose and product quality was improved. Better sample preparation protocols were a need in 1987 and still are a need today, but new, efficient approaches to solid-phase extraction, solid-phase microextraction, stir-bar sorbent extraction, pressurized solvent extraction, and supercritical fluid extraction were not available to researchers in 1987. The experts were dead wrong in some of their predictions, which proves that not everybody has a crystal ball at their disposal.

References

(1) R.E. Majors, LCGC 5 (6), 454–462 (1987).

(2) R.E. Majors, LCGC 6(4), 298–302 (1988).

(3) R.E. Majors, LCGC 25(6), 532–544 (2007).

(4) R.E. Majors, LCGC 25(3), 248–266 (2007).

(5) C. Dewaele and M. Verzele, J. Chromatogr. 282, 341–350 (1983).

(6) N. Afeyan et al., Bio/Technology 8, 203 (1990).

(7) R.E. Majors and R.M. Nicoud, LCGC 18(7), 680–687 (2000).

(8) B. Karger, "Recent Advances in Ultratrace LC/MS Proteomic Analysis," PL.01, HPLC 2007, June 17, 2007, Ghent, Belgium.

Ronald E. Majors "Column Watch" Editor Ronald E. Majors is business development manager, Consumables and Accessories Business Unit, Agilent Technologies, Wilmington, Delaware, and is a member of LCGC's editorial advisory board. Direct correspondence about this column to "Sample Prep Perspectives," LCGC, Woodbridge Corporate Plaza, 485 Route 1 South, Building F, First Floor, Iselin, NJ 08830, e-mail lcgcedit@lcgc-mag.com.