Perspective pubs.acs.org/ac
Development of Gas Chromatographic Mass Spectrometry Ronald A. Hites* School of Public and Environmental Affairs, Indiana University, Bloomington, Indiana 47405, United States ABSTRACT: Gas chromatographic mass spectrometry is now widely used for the quantitation and identification of organic compounds in almost any imaginable sample. These applications include the measurement of chlorinated dioxins in soil samples, the identification of illicit drugs in human blood, and the quantitation of accelerants in arson investigations, to name just a few. How did GC/MS get so popular? It turns out that it required parallel developments in mass spectrometry, gas chromatography, and computing and that no one person “invented” the technique. This Perspective traces this history from the 1950s until today.
vapor pressure at temperatures below about 200−250 °C. This heated glass inlet was soon supplemented by a “solids probe,” with which a few milligrams of the sample could be inserted, through a vacuum lock, directly into the ion source, where it was heated. This approach allowed one to obtain mass spectra of relatively nonvolatile compounds, such as free amino acids. The shortcoming of these approaches was that the sample had to be reasonably pure so that one could relate the observed ion masses (actually their mass to charge ratios) and abundances to a specific compound structure. Meanwhile gas chromatography (GC), or as it was known at the time, vapor phase chromatography, had become a ubiquitous tool for the analytical and organic chemist. The invention of gas chromatography is commonly attributed to A. T. James and A. J. P. Martin in 1952.3 This was the first paper to demonstrate the partitioning of an analyte between a mobile gas phase and a stationary liquid phase; hence, the title of James and Martin’s paper included “gas-liquid partition chromatography.” As an aside, it is interesting to note that gas−solid (adsorption) chromatography had been pioneered by Erika Cremer at the University of Innsbruck in Austria, but this work was not published until well after World War II in 1951.4 In any case, James and Martin’s technique became popular throughout the world, and by 1960, almost every organic chemical laboratory had a gas chromatograph, many of them homemade. At the time, GC columns consisted of a 1−2 m long × 2−3 mm i.d. tube packed with particles, which had been coated with a nonvolatile oil or wax. A carrier gas (usually helium in North America or hydrogen in Europe) was caused to flow through the tube and around the particles. The mixture to be separated was injected into this gas stream, and the components were volatilized. Depending on their affinity for the oil or wax on the
W
hile viewing television recently, I accidentally came across one of the so-called crime scene investigation shows. One of the characters was asked a question, and her answer was “Let’s put it in the GC/MS!” It is not often that an analytical method that one worked on as a graduate student, as I did 50 years ago, shows up in today’s popular culture. This is a testament to the power of GC/MS and to its ubiquity in modern analytical chemistry laboratories. Thus, I thought it might be interesting to briefly outline the development of GC/ MS, which is actually a story of parallel developments in mass spectrometry, gas chromatography, and computing. The use of mass spectrometry for the analysis of organic compounds began (like many things) during World War II, when it was largely used by the petroleum industry for the quantitative analysis of refined products. It was soon realized that the technique had powerful applications for the structural identification of organic compounds. Pioneers included Klaus Biemann at the Massachusetts Institute of Technology, who published a ground-breaking book on the subject,1 and Fred McLafferty at Dow Chemical and now at Cornell University, who also published an important book on the subject.2 By the early 1960s, it was clear that mass spectrometry was taking its place alongside nuclear magnetic resonance (NMR) and infrared (IR) spectrometry as workhorse tools in organic chemistry laboratories seeking to verify the structures of synthesized compounds and to identify unknown structures. Other laboratories, notably that of Carl Djerassi at Stanford University got on the bandwagon, and by the late 1960, there was a flourishing literature documenting electron impact fragmentation mechanisms and explaining how one could deduce unknown structures from mass spectra. Because mass spectrometers operated under high vacuum, sample introduction was a little tricky. In the early days, the sample was vaporized into a heated glass container, which was connected to the mass spectrometer’s ion source by a pinhole leak. This approach required that the sample have a significant © XXXX American Chemical Society
Received: May 4, 2016 Accepted: June 24, 2016
A
DOI: 10.1021/acs.analchem.6b01628 Anal. Chem. XXXX, XXX, XXX−XXX
Perspective
Analytical Chemistry
spectrometer’s vacuum system. In Gohlke’s arrangement, less than 1% of the GC’s effluent was split off and sent into the mass spectrometer. Nevertheless, he was able to demonstrate results for the GC separation of acetone, benzene, toluene, ethylbenzene, and styrene. The data were recorded by photographing the oscilloscope output of the mass spectrometer. (Later a system was developed to sample the mass spectrum at specific times after ion acceleration to obtain the intensity at given masses. Scanning this time on successive ion extractions resulted in a series of data points that could be fed to an oscillographic recorder. In other words, the performance of time-of-flight instruments, lauded for producing a complete spectrum with each ion batch, was degraded into a scanning instrument because of the unavailability of a fast enough transient recorder.) Thus, Gohlke had demonstrated the concept of GC/MS, but it was not sustainable. The split arrangement sacrificed almost all of the mass spectrometer’s sensitivity or required overloading the GC column. In addition, the data acquisition system (photographing the oscilloscope) was too transitory; if one missed taking a photograph at the right time, the data were lost. Other papers using similar splitting arrangements, in which a small part of the GC effluent was directed into the mass spectrometer, were published between 1957 and 1963 by Holmes and Morrell (from Philip Morris, Inc.),10 Lindeman and Annis (from California Research Corp.),11 Ebert (from E. I. du Pont de Nemours & Co.),12 and Miller (also from du Pont).13 It is worth noting that all of these early GC/MS papers were from industrial laboratories probably because academic laboratories could not afford to buy a mass spectrometer. None of these early papers showed data for compounds with molecular weights in excess of about 200 amu, and sensitivities in terms of amount of analyte injected into the GC column were not given. By the mid-1960s attention had turned to maximizing the amount of analyte that entered the mass spectrometer from the GC, while at the same time minimizing the amount of carrier gas that accompanied it and to using mass spectrometers with a higher mass range. The goal was to get as much of the analyte of interest into the mass spectrometer as possible and to avoid the effluent splitting systems that had been previously used and which threw away >90% of the compounds exiting the GC column. In effect, the problem was to separate the analytes of interest from the carrier gas and in this way to maintain the sensitivity of the coupled methods. Two devices were developed independently and published more or less simultaneously in 1964. Schematics of these devices are shown in Figure 1. The top shows the jet separator, which came from Ragnar Ryhage’s laboratory in Sweden.14 This device was originally made from stainless steel, but eventually it was commercialized in glass. The operating principle was based on the idea that spraying the GC effluent into a vacuum chamber would cause the light carrier gas molecules to diffuse in a wider cone-shaped space than the heavier analyte molecules, which could be skimmed out of the middle of the cone and directed into the mass spectrometer. In this way, most of the analyte was routed to the mass spectrometer, and the flow of denuded carrier gas was reduced sufficiently to not overload the mass spectrometer’s pumping system. It turned out that this worked well, and Ryhage demonstrated good results for fatty acid methyl esters using a magnetic sector mass spectrometer, which had been modified to scan up to m/z 500 in a few seconds.14 Both gas chromatographic and mass
particles, the components of the mixture would be separated from one another and detected at the end of the column (in the early instruments by a thermal conductivity cell). Components that had a high affinity for the oil or wax would come out of the column after those that had a low affinity for this stationary phase. The entire system was usually operated in a thermostatically controlled oven. By measuring the peak areas of the detector’s output, one could quantitate the components in the mixture. It was almost immediately clear that GC was a remarkably useful tool for separating compounds that could be volatilized at temperatures less than about 250 °C. The problem was that GC could not provide much information on the structure of the separated compounds, and thus, qualitative analyses were done by comparing retention times (how long it took for a compound to emerge from the column at a given temperature) to those of authentic standards. Because too many compounds could have identical retention times, this mode of qualitative analysis was not satisfying. Only a few years passed before it was noticed that mass spectrometry (MS) (Never say “mass spec.” Actors do that on television. If you are a serious practitioner of the method, always say “mass spectrometry.”) and gas chromatography shared some common features. Both required the sample components to be volatilized before they were injected into the instrument and both were well suited to microgram to milligram amounts of sample (unlike NMR or IR spectrometry). On the other hand, there was a fundamental incompatibility of the two methods: GC operated at atmospheric pressures with carrier gas flow rates on the order of 30 mL/min and MS operated under high vacuum (
