50 years of HPLC

For some chromatographers, their first inkling that high-performance liquid chromatography was going to be a big deal came in 1969. That year, they traveled to Las Vegas to attend the fifth in a series of meetings organized by University of Houston chemist Albert Zlatkis.

Earlier meetings in the “Advances in Chromatography” series had focused on gas chromatography, which separates volatile compounds in a mixture as it passes through a gas-filled column. So, many of the attendees at that meeting in 1969 were GC experts who hadn’t heard much about the technique that was to become HPLC. They were in for an awakening. “That meeting was pivotal in terms of really giving a jump start to HPLC,” says Ronald Majors, who was in attendance. After hearing about the still-nascent technique, “everybody got really excited.”

GC was still the dominant separation technology at the time. But GC had its drawbacks. Most important, GC requires that compounds be volatile or be derivatized to become volatile. “But it’s such a pain doing derivatization,” Majors says. Therefore, many large or polar molecules couldn’t be analyzed by GC.

But they could be by HPLC. In HPLC, a liquid mobile phase carries mixtures of compounds through a column packed with particles. As a pump forces the mobile phase through the column, the compounds interact with the stationary phase—the particles’ surface—to different degrees and are separated in the process.

HPLC was poised to catch on with scientists at that fateful Las Vegas meeting because there were applications—drug discovery and biological separations, among others—where leading separation methods such as GC just weren’t getting the job done. Since those days, HPLC has become a workhorse of the analytical chemistry lab. And applications—from chiral separations to proteomics—continue to drive advances in HPLC today.

In 1969, though, HPLC was a relative unknown. Lloyd Snyder, then a scientist at Union Oil, remembers the Zlatkis-organized meeting as a turning point for HPLC. But even more, he remembers what happened when he came home. Together with other people who had attended the meeting, Snyder organized a half-day meeting at California State University, Fullerton, to update people who hadn’t been able to go.

“We planned for maybe 15 or 20 people showing up,” Snyder says. “They announced the meeting and scheduled it for an ordinary classroom. As the response came in, it progressively moved to larger rooms and finally the auditorium. About 150 people showed up. I could see this was hot stuff.”

According to some estimates, this year marks the 50th anniversary of HPLC. Those estimates typically point to Csaba Horváth and Seymour (Sandy) Lipsky of Yale University publishing a paper in 1966 on ion-exchange separation of organic compounds (Nature 1966, DOI: 10.1038/211748a0). They followed up one year later with a paper on fast liquid chromatography in which they used a pump operating at higher pressures than in their previous work (Anal. Chem. 1967, DOI: 10.1021/ac60256a003). They were among the pioneers at the landmark Las Vegas meeting who introduced the rest of the chromatography community to what would come to be called HPLC.

But pinning down an exact birthday for HPLC isn’t easy. It was predicted long before it was reduced to practice. In 1941, future Nobel Laureates Archer J. P. Martin and Richard L. M. Synge wrote that, for a method called liquid-liquid partition chromatography, the ability of a column to separate compounds was dependent on the liquid-mobile-phase flow rate and the square of the diameter of the particles packed inside the column. They talked in terms of theoretical plates, something chromatographers use to describe how well a column can resolve mixtures. The thinner the theoretical plate, the better. “The smallest [theoretical plate] should be obtainable by using very small particles and a large pressure difference across the length of the column,” they wrote (Biochem. J. 1941, DOI: 10.1042/bj0351369).

So people knew as early as 1941 that efficient liquid chromatography separations would require a combination of high pressure and small-particle packing materials. But it took nearly 30 years to move from theory to practice because there was still much to be learned.

Horváth is generally credited with building the first HPLC instrument. And the Nature and Analytical Chemistry papers he published with Lipsky may indeed have been the first published reports of modern HPLC. But if you ask chromatographers about the beginnings of modern HPLC, they are more likely to mention 1964. That year, Joseph Jack Kirkland visited Eindhoven University of Technology.

“In one of the labs I visited, there was this fellow, Josef F. K. Huber, who was doing what I now call HPLC,” Kirkland says. “He had a very crude instrument with a UV detector. He had packed a column with particles. They were actually GC particles, but he had coated them with liquids.”

Kirkland worked for DuPont at the time. He had tried to use liquid chromatography a few years earlier, but the available equipment and knowledge weren’t yet up to the task. His meeting with Huber encouraged him to try again.

When Kirkland came home, he convinced managers at DuPont to let him work on HPLC. “My work largely was in the pesticide area,” Kirkland says. “We had a lot of compounds that did not lend themselves to GC. Liquid chromatography was something I really had my eye on.”

Even before HPLC, chromatographers were already doing liquid-liquid chromatography. The first “liquid” in this method refers to the mobile phase, and the second one refers to the liquid coated onto the packing material. The two liquids need to be immiscible.

To make the technique work, Kirkland says, you have to make sure the ratio of coating to mobile phase is just right so the mobile phase doesn’t wash away the coating. He adds: “It’s difficult to operate, but it’s a very powerful technique.”

While Huber and Kirkland were working on their instruments, Horváth was working on an instrument to separate biological molecules. In fact, he called his instrument a nucleic acid analyzer.

Meanwhile, chromatography firm Waters Associates, now known as Waters Corp., was pushing to roll out the first commercial HPLC. Depending on which instrument you want to count, it was in the HPLC market even before Huber built his version.

In the early 1960s, Waters licensed technology for gel permeation chromatography from Dow Chemical. That method separates large molecules such as polymers on the basis of their size by passing them through a column packed with cross-linked polymer beads. Waters introduced its GPC-100 in 1963. That instrument included a pump that operated at 500 psi. There’s some disagreement about whether that GPC instrument should be considered the first commercial HPLC or merely a precursor to it.

Not long after that, the company introduced another instrument for liquid-liquid chromatography, incorporating technology licensed from Shell Development. That instrument was less than successful. In fact “it was a disaster,” company founder James L. Waters remembers. “The liquid that was supposed to be stationary on the packed particles wouldn’t stay on all the time. We built three instruments, sold one, took it back, and that was the end of that.”

But Waters didn’t give up on HPLC. The company’s next foray into the field was more successful. That instrument, called the ALC-100, used a higher pressure pump than its GPC predecessor and was intended as a general-purpose analytical instrument. The ALC-100, which was introduced in 1967, is widely considered to be the first successful commercial HPLC.

The market for HPLC opened up considerably when Waters won over world-renowned organic chemist Robert B. Woodward, a Chemistry Nobel Laureate and professor at Harvard University. One Friday in 1972, Woodward’s postdoc Helmut Hamberger came to Waters with a problem. He was trying to separate isomers of an intermediate in the synthesis of vitamin B-12.

At the time, “I didn’t know who Woodward was. He was just some chemist at Harvard,” James Waters says.

But Waters’s colleague James Little knew. Waters remembers Little telling him: “If we can solve a problem for him, we can sell a lot of instruments.”

Waters had been thinking that organic synthesis might become a major market for HPLC, so he took the ALC-100 to Woodward’s lab himself. “I had the problem pretty well solved by the end of the second day and separated small amounts of material for him,” Waters says. To separate larger amounts, Waters injected more material and cycled the output back through the column. “I had to cycle it something like seven times to get the other isomers separated so I had the one Woodward wanted,” Waters remembers.

Waters parlayed that success into a new market. He used the American Chemical Society Directory of Graduate Research to assemble a list of organic chemists. He mailed each one a brochure with a photo of Woodward and a description of what they’d done. In the following year, “we probably sold 40 instruments,” Waters says. “After that, we owned the organic synthesis market because we had the reputation and the experience.”

Since those early days, the development of HPLC has been one of evolution rather than revolution, which is fitting, given that even HPLC’s beginnings involved incremental changes. Although there have been developments in the hardware—pumps, detectors, and especially the introduction of electrospray, an ionization technique that allowed HPLC to be interfaced with mass spectrometers—much of the history of HPLC is the story of new column-packing materials.

Horváth packed his early columns with material he called pellicular particles, which were impermeable spheres coated with a porous ion-exchange resin. But Horváth’s columns were inefficient because compounds couldn’t diffuse deeply into them.

Kirkland, collaborating with Ralph Iler at DuPont, improved efficiency by making spherical particles in which the solid silica core was coated with a porous outer shell of tiny beads. The resulting core-shell particles had more surface area and capacity for interacting with and separating compounds. Kirkland and Iler followed these superficially porous particles with fully porous particles of different controlled sizes and porosities.

Then came bonded phases. One of the biggest developments in the history of HPLC, the bonded phase consists of molecules covalently linked to silica particles. Octadecylsilane, also called C18, was the most common of these molecules, which act as the stationary phase. Majors, who wrote about columns and sample prep for the magazine LCGC for more than 30 years, estimates that 800 different C18 columns have been introduced over the years.

One might think that with so many C18 stationary phases, they’re just duplicating each other. But that’s not the case, Majors says. The C18 phases can be monomeric or polymeric. They can be endcapped, meaning that any silanol groups remaining on the particle surface not attached to C18 are chemically blocked. They can be mixed modes, meaning that something other than C18 is also bonded to the silica. “There are so many variations of how you can make a C18,” Majors says.

And the various C18 columns don’t provide identical separations. The underlying silica, the surface area, and the pore sizes can all be different, creating different interactions with the compounds being separated.

In all, there are about 1,400 different stationary phases, Majors estimates. But many of them are for niche applications. “There are a lot of different stationary phases, but they’re not used that much. I would guess that probably the top five or six stationary phases would separate 90 or 95% of all samples that are done now,” Kirkland says.

Over the years, the particles have become progressively smaller and more uniform. Before the advent of HPLC, chromatography columns were packed with irregularly shaped particles larger than 100 µm. Horvath’s pellicular particles were about 40 µm. Throughout the 1970s and 1980s, the particle size decreased to 10 µm, then 7 µm, then 5 µm.

By the late 1980s, the field had settled on particle diameters of about 3 µm and upper pressures of about 6,000 psi, which were needed to push the mobile phase through the tightly packed columns. As the particles got smaller, the columns became shorter and the separations became faster, partially as a way to avoid operating at the pressure limit.

“They were trying to keep the columns operating around 2,000 or 3,000 psi at most,” says James W. Jorgenson, a chromatographer at the University of North Carolina, Chapel Hill. “Even though the equipment can go to 6,000 psi, you don’t want to run a column at the pressure limit. What kind of lifetime is it going to have before it starts plugging up and exceeds the pressure limit?”

In the 1990s, Jorgenson started using even smaller particles (less than 2 μm) and higher pressures (as high as 100,000 psi) to carry out separations. “We were after very high efficiency separations,” he says.

Waters Corp. began its own work with higher pressures and smaller particles, but its goal was speed. “That was a wise choice on their part,” Jorgenson says, “because a large chunk of the market—people who do pharmaceutical separations—needs high throughput.” If Waters could promise highly resolved compounds and get it done in 10 minutes instead of 30 minutes, Jorgenson adds, it was a huge selling point.

The move to sub-2-μm particles and higher pressures—eventually dubbed ultra-high-performance liquid chromatography, or UHPLC—required a complete redesign of the equipment around the column. “You didn’t just change pumps and everybody was happy,” says Peter Carr, a chromatographer at the University of Minnesota, Twin Cities. “You had to reengineer every part of the system. The pumps, the injector, the connections, the column itself, and the detector—all brand-new.”

Most important, the volumes of all those parts needed to be reduced. “The more efficient columns challenge the equipment in terms of injector volumes, connecting tubing volumes, and detector volumes,” Jorgenson says.

Even today, researchers continue decreasing the particle size, leading to the inevitable pressure increases. But some, such as Kirkland, question the need to continue pushing higher.

Pressures of about 15,000 psi are sufficient for most applications, Kirkland says. “People are beginning to understand better the compromise between pressure and resolution. If you’ve got a difficult sample with a lot of material in it you want to separate, you actually have to use a very long column with large particles. You use the pressure needed for larger particles to be able to operate that long column.”

And industrial chromatographers are likely to view pressures much higher than the current 15,000 to 20,000 psi with suspicion. “In industry, there’s going to be an evaluation of high pressure from a safety perspective,” says Mary Ellen McNally, a chromatographer at DuPont. Most companies would probably require instruments operating at exceptionally high pressure to be isolated in special labs, she says. That skepticism of high pressures has led to a resurgence of core-shell particles (see page 35).

Improvements in separation efficiency are still needed as chromatographers turn their attention to complicated biological mixtures. One way they are achieving such separations is by using two—or more—columns in series. In the simplest form of this multidimensional chromatography, each column has different selectivity, so the combination can tease apart more components than any of the columns on its own.

“When you have 10,000 or 100,000 compounds, you can’t separate them with one column,” Majors says. “You’ve got to go beyond that.”

In two-dimensional HPLC, the output from one column is injected onto a second column. When two columns are used in this way, the method effectively multiplies, rather than adds, their capability for resolving components.

Carr, who works in multidimensional chromatography, is cautiously optimistic about the method’s potential. He thinks it could eventually account for as much as 40% of liquid chromatography separations.

“If 2-D LC can change in a big way how methods are developed, it will be really important,” Carr says. “If the way methods are developed stays the same, then 2-D will be important to certain fields, such as metabolomics, proteomics, and lipidomics, but it won’t be a routine way to do business. The next five years will tell that story.”

And so the evolution of HPLC continues. And much as in the early days, the applications scientists want to use them for will push the developments.