Thirty-Two Years Connecting with Gas Chromatography

I have edited the “GC Connections” column in LCGC for over 32 years. In that time, gas chromatography (GC) has changed from a specialized separations technique to being more routine for standardized methods, while expanding outward to complex multidimensional separations and scaling downward towards miniaturization and portability. In my last “GC Connections” article before stepping away from the role of Editor, I give some perspective on how GC has changed and grown over the years, and where it might go in the future.

I had been working for a few years at PerkinElmer in Norwalk, Connecticut, USA, as a Gas Chromatography (GC) Applications Manager when I started to receive a new technical publication called LCGC Magazine. Its predecessor, LC Magazine, had started in 1983. After several years the publication added “GC” to its name, and in 1987 about sixteen GC-related articles appeared. About halfway through that year, in an effort to increase GC contributions, the editors had inquired around the chromatographic community for someone who could author a new regular GC-centric column in parallel to John Dolan’s “LC Troubleshooting” (1). Leslie Ettre, who worked at PerkinElmer from 1956 until his retirement in the late 1990s, was among those who were asked about it, and, with his recommendation, I embarked on this new venture in the September 1987 edition, with the first instalment entitled “A Systematic Approach”.

It was a great privilege to work directly with Ettre on various projects at PerkinElmer from that time and onward past his retirement. As a chromatography pioneer, he possessed insights and knowledge that guided all of us there. We published in chromatography journals about headspace, capillary (open-tubular) columns, and multidimensional separations, primarily. I was also able to work with Marcel Golay on his final capillary column project, which was an important lesson on how to think outside the box, or beyond the oven, perhaps.

At the time, PerkinElmer took customer application questions over the phone; e-mail was still in its adolescence, and not well adopted by large corporations, while postal mail was too slow. We also had a fax machine, which was useful for receiving chromatograms, and a Telex (remember that?), a type of end‑to-end telegram system. Incoming queries were passed on to one of the GC department personnel, myself included, in rotation. I felt confident that this incoming stream of troubleshooting opportunities would serve well, appropriately anonymized, as source material. I was also able to travel extensively to research laboratories, conferences, and seminars, which broadened my knowledge and experience, and supplied much material for the column.

The first few years of instalments took a tour through the GC system from inlets to detectors, and then delved into more details of specific techniques, columns, and optimizations. One early challenge involved the mechanics of writing, editing, and transmitting articles to the publisher. PerkinElmer was in transition from traditional secretarial writing assistance and had recently installed a newfangled “word processor”, which involved a computer the size of a two-drawer filing cabinet, and a trained operator who, in reality, was the effective word processor because all original material and edits still had to go through her. Around the same time, the GC Applications Lab had acquired two IBM PC-XT computers with a printer that had an interchangeable typeface ball. These included a more flexible word processing program that was chosen specifically because it could print the super- and subscripts that were needed to render chromatographic equations properly. The most significant feature of this system was that it displaced the human word processor and handed over full control to us authors. It was full speed ahead from there. I’m writing this article on much the same equipment. The main difference now is the printer doesn’t get used much anymore.

In 1993, the title of the “GC Troubleshooting” column was changed to “GC Connections”, to reflect a broadened scope that addressed ongoing developments in the field, including annual coverage of new GC-related products at the Pittsburgh Conference (Pittcon). Along the way I became one of those Pittcon fixtures, having attended every one (save two or three) from around 1986 through the present year, and I have the badges to prove it! That’s about thirty of them, each unique and, for me, always split between gathering material for “GC Connections” and playing whatever role my employer had in mind for that year. In 2019, as in most years before, the quantity and innovative nature of the new GC offerings found there was impressive. GC technologies continue to grow and adapt as much as ever.

This instalment is a personal retrospective about GC columns and hardware. It is impossible to mention all of the researchers, engineers, and others who made possible the tremendous advances that have propelled this branch of separation science forwards, and I hope the reader will be forgiving of my many omissions. I’d like to thank everyone with whom I have worked over the years, and especially all the editors and staff of LCGC for their patience and good advice.

Fused-Silica Columns

I started practising GC in graduate school. I worked in Charlie Lochmüller’s group at Duke University, where I synthesized chiral liquid‑crystalline and polysiloxane stationary phases, and then coated glass capillary columns with them as best I could. Before going to Duke, I graduated from college in 1973 with a B.S. in organic chemistry, and then worked at Aldrich Chemical in Milwaukee, Wisconsin, USA, for a short while. I left Aldrich to return to academia after discovering that, although I was very good at organic synthesis, producing commercial fine chemicals wasn’t really what I wanted to do.

At Duke, I applied the gonzo‑synthesis techniques I had learned to smaller challenges, and soon I had new columns ready for testing. My first experiences were with soda-glass columns pulled from a larger tube with the aid of a clunky Huppe–Busch drawing machine. After coating, the column ends had to be straightened by hand with a blowtorch before fitting to the inlet and detector. Some of the chiral phases worked well, and for a short while I could claim the largest published separation factor for a derivatized amino acid enantiomeric pair.

I knew I was a little too early starting to work with capillary columns when in 1979, towards the end of my graduate studies, fused-silica columns came onto the scene. We received two uncoated columns from a supplier, and, of course, I had the task of coating and testing them. I remember the first chromatogram looked very good, but after that, no peaks at all were eluted! Upon opening the GC oven, we discovered that the precious fused-silica column was now about fifty pieces of broken column, due to, at that time, the not well understood effects of thermal stress. So, in the end, I was grateful that I had published several papers using old-school glass columns.

In 1979, with my freshly minted Ph.D., I signed on to work in Varian’s GC engineering department, where my first assignment was to devise ways to make fused-silica columns at a commercial scale. One of the first pieces of equipment I obtained was an aquarium from the local pet store, to serve as a low-gradient temperature bath for column coating. This was the butt of many jokes, and I even dropped a few tropical fish in for fun. But it soon became clear that the fused-silica column business was better done by a dedicated independent operation, and not adjunct to another production facility. This is still the case today, although now all of the column suppliers that I’m aware of are part of a larger instrument company or other business. Essentially all raw fused‑silica column tubing is produced in just a few facilities worldwide.

For a while, glass columns were the only way to perform high‑temperature GC, in excess of 300 °C. Early in fused-silica column development, an aluminium outer coating became available, which made higher temperatures attainable. Polyimide outer coatings, at first not useful at higher temperatures, were improved eventually and made high temperature operation up to 400 °C possible. The stationary phase became the temperature‑limiting factor, as it still is today. The work of many academic and commercial researchers advanced stationary phase polymer chemistry for cross‑linking and bonding to the inner wall. The temperature limits, column bleed, and consistency improved to a point that much of the available high performance is now considered routine. Metal wall-coated open‑tubular capillary columns endure, only now they are made by coating steel tubing with a thin layer of silica, an inversion of the older techniques.

One of the shortcomings of wall‑coated open-tubular columns is the thin film and low volume of the stationary phase. It bestows high separation efficiency, but limits solute mass before overloading and peak distortion occur. A compromise solution is to increase the surface area on the inner column wall by etching or by depositing support material, such as in a porous-layer open-tubular (PLOT) column. A lot of research was performed with glass columns to improve their inertness while increasing sample capacity. We had experimented with some salt deposits at Duke, but never got it right-the resulting columns were not at all inert, and had terrible peak shapes. As stationary phase polymer chemistry advanced, PLOT columns with polymeric phases, such as polysiloxanes and polyethylenes, were largely supplanted by thick‑film columns. Particulate stationary phases, such as molecular sieves, alumina, and porous polymers, now predominate in PLOT column offerings.

Programmed Temperature Inlets

After concluding the fused-silica column work at Varian, I switched over to GC instrument development. On a trip to Europe, on the way to my first International Symposium on Capillary Chromatography (ISCC), I had visited a number of research laboratories and chromatography businesses. I recall that this was also the occasion of the first ISCC to be held in Riva del Garda. At the time, there was a lot of interest in GC inlet systems, specifically around ways to improve the accuracy of split and splitless injections as related to certain samples that were unstable under conventional injection conditions. Unlike in liquid chromatography (LC), GC solutes must enter a gaseous phase for elution to occur. Flash vaporization in a GC inlet prior to entering the column is the norm, and it works quite well in most cases. However, some sample compositions can be strongly affected by flash injection, namely high boiling or wide boiling‑range components, as well as solutes that are thermally labile or subject to catalytic decomposition. These effects were exhaustively documented in myriad publications, along with various methods to suppress them, such as specially configured inlet liners, syringe-loading techniques, injection timing, and so on.

The alternative direct on-column injection technique eliminated such problems, but other difficulties arose from depositing the entire sample directly onto the column. Nonvolatile material would build up and eventually contaminate the column head, and the sample volume was limited to a relatively narrow range. High concentrations had to be diluted, while not enough trace sample could be injected without flooding the column. An innovative solution, a programmed-temperature inlet, had been developed by several groups and was into early stages of commercialization at the time of the ISCC conference. This type of inlet system transfers sample into the inlet near room temperature, which suppresses many of the undesirable effects of hot injection. The inlet subsequently heats the sample to vaporize it into the column. Split, splitless, direct, and large-volume injections were all possible. While not suitable for all samples, programmed‑temperature inlets are found in a wide variety of applications today.

Electronic Gas Control

In the mid 1980s, microprocessors were supplanting discrete designs in analytical instruments across all disciplines. For GC, opportunities beyond improvements to thermal, valve, and detection control included firmware-enabled electronic control of gas flows to columns and detectors using electro-mechanical metering valves. This made it simpler to include flows and pressures along with temperature and detection in the GC method parameters that could be accessed via a front panel display and keyboard or attached computer. Beyond these basics, programming the column pressure was made more practical and accessible. Increasing the column pressure drop during a programmed-temperature run had been achieved much earlier with components such as a clock-motor driven pressure regulator. This helped overcome losses in column efficiency with constant pressure operation at higher temperatures: Column flow and linear velocity would drop below the optimum, due to increasing carrier gas viscosity. However, the early implementations were largely abandoned, due to their impracticality.

The mathematical relationships that describe the interaction of column pressure drop, flow, velocity, and viscosity are well known, and, with microprocessor control, it became much easier to achieve fine control of carrier gas parameters across the time domain of a separation. In the space of a few years in the early 1990s, all the major GC instrument companies released their versions of computer‑electronic carrier gas control under a few different names with associated three-letter acronyms. Looking back at the state of microprocessor capabilities 30 years ago, I’m amazed that at the time we could achieve such a high level of carrier gas control.

GC instruments of the past decade or so lack most of the knobs and pressure dials so familiar to me and others of my generation; they display on-screen readouts instead. But the accuracy of the readouts relies on the good practice of regular gas calibration. Instrument validation and suitability still require externally calibrated flow and pressure measurements.

Miniature, Portable, and Micro GC

Everything about GC has been shrinking for years now. Full‑size multipurpose laboratory GC instruments with large column ovens and dual inlets and detectors are still available, but newer GC instruments trend towards smaller ovens and single‑channel hardware. They occupy less bench space, and can be dedicated to one or a few related applications. They are the result of a stabilization of GC methodology for many routine sample types. The development of new or modified methods remains in the R&D laboratories, while everyday analyses are performed on systems designed more for increased throughput of repetitive samples.

The shrinking GC system also makes possible better integration of high-speed GC components. Smaller components have less thermal mass, which makes faster temperature programming rates easier to achieve, and also reduces thermal recovery times between analyses. Smaller GC instruments that consume less power are easier to move around, and portability becomes more practical. But, for GC, its not as easy as bringing the analyzer to the sample. There are few mixtures that can simply be injected directly into a GC system. Extensive preparation is required for most real-world samples in the form of isolation, concentration, filtration, extraction, derivitization, and so on.

The sampling aspect has held back a broader adoption of portability in GC, and also has impacted the wider deployment of microcolumn GC systems. I would define a microcolumn as one with inner diameter less than 100 µm, possibly etched or otherwise configured onto a chip-like substrate. A practical application was seen at the Pittcon a couple of years back for air monitoring with a mesh‑network configured array of micro GC instruments deployed on telephone poles and other urban structures. Microcolumns are lightweight and very portable, but their small diameters require corresponding reductions in inlet and detector dimensions, as well as enhanced detector sensitivity, to match their reduced sample capacities. Perhaps these challenges will be resolved in the future-who knows? The laboratory GC of the future might be the size of today’s smartphone. Such a system would also fit well into process analysis and control situations.

After leaving PerkinElmer in 1999, I ran a training and consulting business for a few years. I very much enjoyed these activities, but I missed working with a team on new instrumental challenges. An opportunity to participate in a start-up business based on online GC came my way, and in 2002 I signed on with Serveron, in Hillsboro, Oregon, USA, to engineer a new miniature GC system as the measurement component of an online analyzer to be deployed in remote locations. This is a non-laboratory system that had to perform as well as its in-laboratory counterparts. At least three-quarters of the space inside the instrument is dedicated to sample preparation and gas extraction. The work was challenging, but the project was successful and yielded a remotely deployed instrument with better-than-laboratory sensitivity, along with column and gas tank service intervals averaging about four years. Many of them are installed hundreds of miles from anywhere in particular, and left alone until servicing is needed. The entire GC instrument-columns, detector, and electronics-is replaceable as a single component by removing two screws and two cables. I think this is a good example of the melding of sample preparation and analysis into a practical application-specific system that can be deployed remotely.

Today, my work has shifted into the spectroscopic realm: The team is engineering a new analyzer to supplement GC-based ones in areas where managing large quantities of carrier gas tanks is impractical. Who knows, I may publish something in Spectroscopy before too long.

Conclusion

After 40 years in GC, 32 of them writing this column, I’m signing off and moving on to keep innovating and creating anew. “GC Connections” will keep going under the capable watch of Nick Snow, who will bring new insights and his own experiences to bear. I want to thank Nick for taking on this new endeavour, and I wish him the best of everything that comes with a long career in separation science. When I’m not beaming light around the laboratory, look for me outside as I head to my next marathon, bike ride, or road trip on the electrified highway.

References

1. J.W. Dolan, LCGC Europe 30(10), 544–549 (2017).

“GC Connections” editor John V. Hinshaw is a Senior Scientist at Serveron Corporation in Beaverton, Oregon, USA, and a member of LCGC’s editorial advisory board. Direct correspondence about this column to the author via e-mail: LCGCedit@mmhgroup.com