Professional Documents
Culture Documents
SECOND EDITION
This page intentionally left blank
GAS
CHROMATOGRAPHY
SECOND EDITION
Edited by
Colin F. Poole
Department of Chemistry, Wayne State University, Detroit, MI, United States
Elsevier
Radarweg 29, PO Box 211, 1000 AE Amsterdam, Netherlands
The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom
50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States
No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical,
including photocopying, recording, or any information storage and retrieval system, without permission in writing from the
publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our
arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found
at our website: www.elsevier.com/permissions.
This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may
be noted herein).
Notices
Knowledge and best practice in this field are constantly changing. As new research and experience broaden our
understanding, changes in research methods, professional practices, or medical treatment may become necessary.
Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any
information, methods, compounds, or experiments described herein. In using such information or methods they should be
mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility.
To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any
injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or
operation of any methods, products, instructions, or ideas contained in the material herein.
ISBN: 978-0-12-820675-1
v
vi Contents
33.4 Organochlorine pesticides 848 34.5 Search for key chemical biomarkers: Mars
33.5 Halogenated flame retardants 851 exploration 869
33.6 Polybrominated diphenyl ethers 853 34.6 Search for chirality in space 871
33.7 Other halogenated flame retardants 853 34.7 Conclusions and perspectives 872
33.8 Perfluorinated compounds 859 References 872
33.9 Polycyclic aromatic hydrocarbons 859
33.10 Other compounds not specifically 35. Gas chromatographic analysis of chemical
discussed 861 warfare agents
33.11 Summary 861 PHILIP A. SMITH
References 862
35.1 Introduction and background 875
35.2 Analytical considerations for sampling and gas
34. Gas chromatography in space exploration
chromatographic analysis of CWA-related
MARIA CHIARA PIETROGRANDE
compounds 887
34.1 Introduction 865 35.3 GC applications for biomedical CWA
34.2 Technological and operating constraints in space analyses 895
GC 866 35.4 Conclusion 896
34.3 Prebiotic chemistry in Titan’s atmosphere: the References 896
CassinieHuygens mission 867
34.4 Prebiotic chemistry in comet environments: Rosetta Index 901
mission 867
Contributors
Lourdes Arce Department of Analytical Chemistry, Peter Dawes Analytical Science, SGE, Ringwood,
University of C
ordoba, Institute of Fine Chemistry VIC, Australia
and Nanochemistry, Cordoba, Spain alia de Aguiar Porto Institute of Chemistry,
Nath
Carlo Bicchi Dipartimento di Scienza e Tecnologia University of Campinas (UNICAMP), Campinas,
del Farmaco, University of Torino, Torino, Italy SP, Brazil
Leonid M. Blumberg Advachrom, Wilmington, DE, Alexandre de Andrade Ferreira Division of
United States Geochemistry, PETROBRAS Research and Devel-
Cecilia Cagliero Dipartimento di Scienza e Tecnolo- opment Center (CENPES), PETROBRAS, Rio de
gia del Farmaco, University of Torino, Torino, Italy Janeiro, RJ, Brazil
K. H€ u Can Başer Anadolu University, Eskisehir,
usn€ Bieke Dejaegher Department of Analytical Chemis-
TR, Turkey try, Applied Chemometrics and Molecular Model-
ling, Vrije Universiteit Brussel (VUB), Brussels,
Carlos Alberto Carbonezi Division of Geochem-
Belgium
istry, PETROBRAS Research and Development
Center (CENPES), PETROBRAS, Rio de Janeiro, Frédéric Destaillats Nestlé Nutrition Research,
RJ, Brazil Vevey, Switzerland
Maria Jose Cardador Department of Analytical Jaap de Zeeuw Restek Corporation, Middelburg,
Chemistry, University of C
ordoba, Institute of The Netherlands
Fine Chemistry and Nanochemistry, C ordoba, Frank L. Dorman Biochemistry and Molecular
Spain Biology, The Pennsylvania State University, Uni-
Joseph A. Caruso University of Cincinnati, Cincin- versity Park, PA, United States
nati, OH, United States Susan E. Ebeler Department of Viticulture and
Rogério Mesquita Carvalho Division of Chemistry, Enology, One Shields Avenue, University of Cali-
PETROBRAS Research and Development Center fornia, Davis, CA, United States
(CENPES), PETROBRAS, Rio de Janeiro, RJ, Brazil Amadeo R. Fernandez-Alba Pesticide Residue
Qilin Chan University of Cincinnati, Cincinnati, Reseach Group, University of Almeria, La Canada
OH, United States de San Urbano, Almeria, Spain
Eric Chun Yong Chan Department of Pharmacy, Kenneth G. Furton Department of Chemistry and
National University of Singapore, Singapore Biochemistry, Florida International University,
Miami, FL, United States
Chiara Cordero Dipartimento di Scienza e Tecnolo-
gia del Farmaco, University of Torino, Torino, Italy Leandro Wang Hantao Institute of Chemistry, Uni-
versity of Campinas (UNICAMP), Campinas, SP,
Juliana Crucello Institute of Chemistry, University
Brazil
of Campinas (UNICAMP), Campinas, SP, Brazil
David J. Harvey Target Discovery Institute,
Cristina Cruz-Hernandez Nestlé Research, Vers-
Nuffield Department of Medicine, University of
Chez-Les-Blanc, Lausanne, Switzerland
Oxford, Oxford, United Kingdom
xi
xii Contributors
Walter G. Jennings Food Science, University of Maria Chiara Pietrogrande Department of Chemis-
California, Davis, CA, United States try, University of Ferrara, Ferrara, Italy
Natividad Jurado-Campos Department of Analyt- Colin F. Poole Department of Chemistry, Wayne
ical Chemistry, University of C
ordoba, Institute of State University, Detroit, MI, United States
Fine Chemistry and Nanochemistry, C ordoba, Eric J. Reiner Biochemistry and Molecular Biology,
Spain The Pennsylvania State University, University
Abuzar Kabir Department of Chemistry and Park, PA, United States
Biochemistry, Florida International University, Patrizia Rubiolo Dipartimento di Scienza e Tecnolo-
Miami, FL, United States gia del Farmaco, University of Torino, Torino, Italy
Yuri Kalambet Ampersand Ltd., Moscow, Russian A.I. Ruiz-Matute Instituto de Química Org
anica
Federation General (CSIC), Madrid, Spain
Leesun Kim Ulsan National Institute of Science and Karen D. Sam Applications Lab, CDS Analytical,
Technology, Ulsan, South Korea Oxford, PA, United States
Sze Han Lee Department of Pharmacy, National M.L. Sanz Instituto de Química Org
anica General
University of Singapore, Singapore (CSIC), Madrid, Spain
Erica Liberto Dipartimento di Scienza e Tecnologia Kevin A. Schug Department of Chemistry &
del Farmaco, University of Torino, Torino, Italy Biochemistry, The University of Texas, Arlington,
Mainak Mal Department of Pharmaceutical Tech- TX, United States
nology, Brainware University, Kolkata, India John V. Seeley Oakland University, Department of
Philip J. Marriott School of Chemistry Monash Chemistry, Rochester, MI, United States
University, Clayton, VIC, Australia Barbara Sgorbini Dipartimento di Scienza e Tecno-
M. Angeles Martinez-Uroz Pesticide Residue logia del Farmaco, University of Torino, Torino,
Reseach Group, University of Almeria, La Canada Italy
de San Urbano, Almeria, Spain Michael J. Sithersingh KR Consultants, Parsippany,
A. Mena Instituto de Química Organica General NJ, United States
(CSIC), Madrid, Spain Johanna Smeyers-Verbeke Department of Analyt-
Milagros Mezcua Pesticide Residue Reseach Group, ical Chemistry, Applied Chemometrics and Molec-
University of Almeria, La Canada de San Urbano, ular Modelling, Vrije Universiteit Brussel (VUB),
Almeria, Spain Brussels, Belgium
Jeremy S. Nadeau Department of Chemistry, Philip A. Smith US Department of Labor, Salt lake
University of Washington, Seattle, WA, United States Technical Center, Sandy, UT, United States
Ariel M. O’Brien Department of Chemistry & Nicholas H. Snow Chemistry and Biochemistry,
Biochemistry, The University of Texas, Arlington, Seton Hall University, South Orange, NJ, United
TX, United States States
€
Temel Ozek Anadolu University, Eskisehir, TR, A.C. Soria Instituto de Química Org
anica General
Turkey (CSIC), Madrid, Spain
Kishore Kumar Pasikanti Accelerating Therapeu- Stanley D. Stearns Valco Instruments Co. Inc.,
tics for Opportunities in Medicine (ATOM) Con- Houston, TX, United States
sortium, San Francisco, CA, United States Robert E. Synovec Department of Chemistry, Uni-
Karisa M. Pierce Department of Chemistry and versity of Washington, Seattle, WA, United States
Biochemistry, Seattle Pacific University, Seattle, Andrew Tipler PerkinElmer Inc., Shelton, CT,
WA, United States United States
Contributors xiii
Timothy J. Trinklein Department of Chemistry, Adam Voelkel Institute of Chemical Technology
University of Washington, Seattle, WA, United and Engineering, Poznan University of Technology,
States Pozna
n, Poland
Yvan Vander Heyden Department of Analytical Elizabeth Woolfenden Markes International
Chemistry, Applied Chemometrics and Molecular Limited, 1000 Central Park, Western Avenue,
Modelling, Vrije Universiteit Brussel (VUB), Brus- Bridgend, United Kingdom
sels, Belgium
C H A P T E R
1
Milestones in the development of gas
chromatography
Walter G. Jennings1,y, Colin F. Poole2
1
Food Science, University of California, Davis, CA, United States; 2Department of Chemistry,
Wayne State University, Detroit, MI, United States
y
Author W.G. Jennings is deceased.
Gas Chromatography
https://doi.org/10.1016/B978-0-12-820675-1.00018-6 1 © 2021 Elsevier Inc. All rights reserved.
2 1. Milestones in the development of gas chromatography
theoretical basis for gas chromatography was University's Institute of Physical Chemistry. Pro-
first conceived by Erika Cremer, an Austrian sci- fessor Dr. Cremer, by all accounts a brilliant
entist at the University of Innsbruck, Austria, in woman scientist, died in 1996.
the late 1940s during the period of the Second
World War. As Ettre points out, this was a period 1.3 Early instrumentation
when women, especially in Germanic countries,
were expected to confine their activities to “chil- By 1953, several petroleum companies, pri-
dren, church, and kitchen”. In spite of her “su- marily in Great Britain and the Netherlands,
perb Ph.D. thesis work,” she had great were exploring this new analytical technique,
difficulty in finding a position. Her opportunity and in 1954, a few flavor chemists (including
came in 1940 when the war started, and univer- this author) were building crude chromato-
sity teachers were drafted. She obtained an aca- graphs, many of them based on an article by N.
demic position at the University of Innsbruck, H. Ray [5]. Ray inserted thermal conductivity
Austria (then a part of Germany), in the Institute cells into a Wheatstone bridge, whose outlet con-
of Physical Chemistry. It was here that she and nected to a strip chart recorder, thus generating a
her students (with major credit to Fritz Prior) Gaussian peak for each eluting solute; this was
constructed the first prototype of a gas chro- (to my knowledge) the first gas chromatogram
matograph (GC) (Fig. 1.1) and, after a long delay as we know it today. The schematics and chro-
that was probably attributable to the war, pub- matograms published by Ray encouraged a
lished the results of their research in 1951 [3,4]. number of readers (including the author) to
At this time she was promoted to professor build similar chromatographs. Almost every
and, some 20 years later, to director of the part of these crude instruments had to be
FIGURE 1.1 The gas chromatographic system used in Prior’ work, in 1945e1947. (A) adsorbent for purification of the
carrier gas (hydrogen); (B) sample inlet system; (C) buret containing mercury with niveau glass for sample introduction;
(D) Dewar flask; (E) separation column (containing silica gel on activated carbon); and (F) thermal conductivity detector.
From Ref. O. Bobleter, Exhibition of the first gas chromatographic work of Erika Cremer and Fritz Prior, Chromatographia 43 (1996)
444e446; Copyright Friedr. Vieweg and Sohn.
1.4 Early column developments 3
self-designed and self-made, but this introduced limits the practical length of the column, usually
the days of the packed column; supports, such as to a few meters. In much later days, some packed
granules of diatomaceous earth, were coated columns approached an efficiency equivalent to
with a variety of high-boiling fluids (e.g., diethy- our current 0.32 mm internal diameter capillary
lene glycol succinate, DEGS) and, in our earliest columns (ca. 3000 plates per meter), but because
efforts, packed into copper columns, typically of their length limitations, they could achieve
3e6 m long with internal diameters of about perhaps 35,000 theoretical plates while a 30 m
6 mm, and (usually) coiled. It was soon recog- open-tubular 0.25 mm internal diameter column
nized that copper columns were quite active, should be capable of three times that efficiency,
and most workers switched, first to stainless merely because it is three times longer.
steel and then to coiled glass tubing of similar di- It was at the 1958 Second International Sym-
mensions. The reminiscences of many of the posium on Gas Chromatography in Amsterdam
early pioneers in gas chromatographic instru- that Marcel Golay, who was then a consultant to
mentation are summarized in Ref. [6]. PerkinElmer, introduced the theory of open-
tubular capillary columns and demonstrated
their superiority to packed columns [9]. (There
1.3.1 Early commercial instruments is no connection between the old PerkinElmer referred
The first companies to manufacture GCs in to here and the PerkinElmer of today, they are entirely
Europe were Griffin and George (London) and different companies.) These columns demanded
Metropolitan Vickers Electric (Manchester), but much smaller sample injections, necessitating
the US instrument companies had a greater more sensitive detectors than the thermal con-
impact on the development of this new area ductivity detectors that were in common use at
[7]. PerkinElmer was one of the first companies the time. Fortunately, James E. Lovelock had
to market a GC; in May of 1954, they introduced invented an electron-capture detector in 1957
their Model 154 Vapor Fractometer. The temper- [10], and in 1958 the flame ionization detector
ature of the column oven was adjustable from appeared; some credit this to Harley, Nel, and
room temperature to 150 C, and it offered a Pretorious [11], and others to McWilliams and
“flash vaporizer” with a rubber septum permit- Dewer [12,13]; these increased detector sensitiv-
ting syringe injection into the carrier gas stream. ities by ca. 103e106. The invention and develop-
The detector was a thermal conductivity cell. The ment of the flame ionization detector are
instrument was a great success and sold widely discussed in more detail in Ref. [14]. Early capil-
[8]. In early 1956, PerkinElmer followed up lary columns were of plastic and copper tubing;
with their Model 154-B, with a temperature the former had serious temperature limitations
range from room temperature to 225 C and and the latter was active; this led to stainless steel
could be fitted with an optional rotary valve of- tubing. PerkinElmer had filed for and been
fering a variety of sample loops for the injection granted patents on the open-tubular concept,
of gas samples. essentially worldwide. I purchased two Perki-
nElmer wall-coated open-tubular (WCOT) col-
umns at different times, only to find that the
1.4 Early column developments columns available from them produced abomi-
nable results. PerkinElmer apparently recog-
All of the early instruments utilized packed nized this fact, because they essentially
columns, some coiled, some U-shaped. Packed abandoned their research efforts on WCOT col-
columns all have one thing in common: they umns, outsourced their production, and redir-
possess a high resistance to gas flow, and this ected their efforts to columns whose interiors
4 1. Milestones in the development of gas chromatography
were first coated with a support material (e.g., 1.4.2 The positive results of patent
diatomaceous earth) and then with the station- enforcement
ary phase. These were dubbed “support coated
open-tubular (SCOT) columns.” PerkinElmer The low quality of the PerkinElmer WCOT
continued to rigorously enforce the patent on columns and enforcement of their patent led
WCOT (and SCOT) columns, but under consid- many scientists to begin making their own col-
erable pressure (especially from the applicant) umns for their own use. This caused investiga-
they were forced to issue one license, to Hans- tors in many other fields, who would have
jurg Jaeggi, a former assistant of Kurt Grob in preferred to purchase useable columns from an
Switzerland. Under Swiss law, if a patent bars outside source and confine their research to
a Swiss from conducting his or her business, a li- some other field, to now have to make columns;
cense must be issued. Jaeggi had been and still scores of scientists were now studying and pub-
was making excellent glass WCOT columns. lishing on methods of pretreating, deactivating,
The high quality of his columns led PerkinElmer and coating columns. Their combined results
to later propose that they collaborate, but, prob- were responsible for many of the advances in
ably because of the acrimonious battle he had column improvements, and they soon surpassed
gone through to obtain his license, he wanted the results that had been generated by Perkin
nothing more to do with PerkinElmer and Elmer [17].
refused their offer. His obituary, written by Kon-
rad Grob, makes interesting reading [15].
1.4.3 Mileposts in coating WCOT
capillary columns
1.4.1 Do-it-yourself glass capillary
Golay had coated his original glass open-
columns
tubular columns by completely filling them with
In 1965, Desty et al. invented an elegantly simple a dilute solution of stationary phase in a low-
machine for drawing long lengths of coiled glass boiling solvent, sealing one end, and then drawing
capillaries [16]. Besides the fact that this was much the column, open end first, through an oven [18].
less expensive tubing, it also had a smoother interior As used by Golay, the column could be coiled
and it was transparent. For the first time, it was only after coating, a distinct drawback. The
possible to scrutinize the layer of stationary phase method was improved by Ilkova and Mistrykov
as it existed on the column wall; soon it was obvious [19], who coiled and then filled the column and
that when a new unused column exhibiting a thin sealed one end. The open end of the filled column
uniform film of stationary phase was exposed to was fed into a heated oven by supporting the col-
higher temperatures, the stationary phase collected umn on a rotating rod fitted with a drive roller at
into beads that were randomly scattered over the in- the entrance to the column oven, thus literally
ner surface. Almost immediately scores of scientists screwing the rest of the column into the oven.
realized that many of the low-viscosity fluids that Up until just a few years ago, all chemistry de-
worked well in packed columns were unsatisfac- partments at the multicampus Universities of
tory for WCOT columns and should be replaced California system frowned on applied research,
with high-viscosity materials that would retain their and chemistry department faculties drifting
high viscosity even at higher temperatures. Low- away from pure organic chemistry or pure phys-
viscosity silicone fluids (e.g., OV 101) were replaced ical chemistry rarely survived to tenure. Analyt-
with high-viscosity silicones (e.g., OV 30, a viscous ical chemistry was essentially forbidden. There
paste-like silicone), and experimenters soon began were many faculty members scattered through
producing much more stable columns. various other departments who were analytical
1.4 Early column developments 5
chemists, and eventually they formed a “Group This may have guided me as I attempted to
in Analytical and Environmental Chemistry,” modify our column coating apparatus. After
open to anyone in any department who had in- replacing the opaque lid of the oven with a glass
terests in the analytical side. At one time I chaired lid, it was obvious that the solution was super-
that group for several years and it still exists. At heated and evaporated in a series of minor
this time, my title was “Professor of Food Science explosions, leaving blotches of stationary phase
and Technology and Chemist in the Experiment randomly scattered over the surface of the col-
Station.” This had several advantages: for one umn. I decided to introduce the column into
thing, those with just the academic title worked the oven through a preheater made from a short
9 months per year, while the Experiment Station curved length of 1/8 inch stainless steel tubing
operated 11 months a year. My department lagged with an electrically heated wire and heat-
chairman tolerated what he called my ed to 150 C: the column's passage through the
“dabbling” in gas chromatography, but insisted curved preheater was not smooth; indeed, it
that I should also be working on subjects “more scraped the walls and vibrated. To avoid
aligned with the food industries”; he suggested breakage, I introduced graphite powder into
“circulation cleaning,” which was just emerging. the loop. The column continued to vibrate, but
Swallowing what I wanted to say, I assigned a more gently. In retrospect, that vibration at this
new Ph.D. student (Malcolm Bourne) to the proj- point on its passage into the oven would also
ect, built a miniature circulation pipeline, and discourage superheating. It may have been
helped him bake radiolabeled tristearin onto dumb luck, but from then on, the evaporation
glass microscope slides and strips of stainless of the solvent occurred smoothly. We were
steel of similar width and length. These were routinely producing very stable columns with
inserted into the rapid circulating system, and uniformly thin coatings.
the levels of radioactivity measured every 2 mi-
nutes. Inevitably, semilog plots of our cleaning
1.4.4 Commercial column
curves showed a rapidly dropping curve that
terminated in a straight line with a slight down-
manufacturers
ward slope. All at once Bourne realized that he At this point, one of my Ph.D. students, Rob-
was looking at two first-order reactions that ert Wohleb, suggested that we had a marketable
were occurring simultaneously. The baked film product and should establish a column produc-
of tristearin existed only at one of two energy tion company. When I broached this suggestion
levels, a loosely bound system, and a much with my department chairman, he called in the
more tightly bound systemdnothing in between. dean. After much discussion, they agreed that I
Spatial estimates indicated that the tightly bound could engage in this activity only if I agreed to
form was not a monolayer, but could be up to at several restrictions: (1) as long as I was a full-
least 20 molecular thicknesses [20e22]. He also time professor, I could receive no remuneration
discovered that by manipulating time and tem- from this venture; (2) there could be no connec-
perature, he could change the ratio of the two tion between my university research and
species. Bourne went on to study the kinetics of activities of the company; (3) I could not get
cleaning in much greater depth and made quite involved in the day-to-day activities of the
a name for himself. I went back to my company. I could, on my own time, answer
“dabbling,” then slowly realized that my newly trouble-shooting questions and engage in educa-
gained knowledge on removal of thin films was tional activities. J&W Scientific was founded in
just the reverse of achieving more stable coatings 1974; I constructed two drawing machines in
of stationary phases to the column surface. my machine shop at home and left them with
6 1. Milestones in the development of gas chromatography
Wohleb as I departed on my second sabbatical have happened. It turned out that Kurt Grob in
leave, this time in Karlsruhe, Germany. Sandy Switzerland had (independently) been working
Lipsky had launched his company, Quadrex, along these same lines. Checking on the dates
slightly ahead of J&W. that both parties had ordered vinyl chemicals,
Rand was just about 5 days ahead.
1.4.5 Bonded, cross-linked, and/or
immobilized stationary phases 1.4.6 Further improvements in
In 1975, J&W noticed a sudden drop in column
stationary phases
sales from several regular customers; fortunately, In the late 1950s and extending into the 1960s,
we had been using (and saving) bar codes to trace gas chromatographers dealt with a plethora of
every step each column went through. The bar stationary phases. Petroleum chemists, dealing
codes on all of the columns those customers had with nonpolar products, favored phases such
been buying showed they had all gone through as squalane, a fully saturated hydrocarbon
the same coating machine in the same general (C30H62) on the basis of like-attracts-like. The up-
time period. I called several of the buyers and per temperature limit for Squalane is 125 de-
asked if they were having problems with our col- grees, and a higher temperature paraffin,
umns and was shocked when they replied, “no, Apolane 87 (C87H176), was sometimes used
the columns are greatdthey never wear out.” instead. Chemists dealing with more polar prod-
This was serious, we were putting ourselves out ucts needed more polar stationary phases, and
of business. In checking that coating machine, it they turned to high-boiling esters such as dieth-
was discovered that the thermocouple on the pre- ylene glycol succinate (DEGS). As more and
heater through which the column entered the more scientists entered the field, the range of sta-
oven had failed some time ago and was giving tionary phases exploded to over 200 different
much lower readings. Over the past few weeks, phases, most of which were gradually discarded.
the technician on the machine had simply The polysiloxane stationary phases are now
compensated for the low reading by boosting widely used and have been reviewed by Blom-
the voltage. Instead of 150 C, that preheater berg [23] and Haken [24].
was close to 300 C. The stationary phase we
were using (SE-54) contained 1% vinyl. Our
R&D head, Rand Jenkins, realized that seren- 1.5 Interfacing glass capillary columns to
dipity had smiled on us! He immediately ordered injectors and detectors
materials and began adding vinyl-containing
compounds, to our coating solutions, hoping Attaching these relatively fragile glass
that the phase would not only cross-link but capillaries to injectors and detectors via leak-
also connect with some surface hydroxyls. With proof connections soon became a problem;
some pretty harsh testing, we found that the col- some workers punched holes in small disks of
umns exposed to these higher preheater tempera- silicone rubber, which were then substituted
tures were much more rugged, and the deposited for the ferrules supplied with 1/16 inch Swage-
phase could not be removed with solvent rinsing. lok fittings, but all too frequently, as the nut on
At the next “Advances in Chromatography” the Swagelok fitting was tightened to the point
meeting in Houston, we announced the “Bonded that there were no leaks, the compressive strain
Phase Column.” If the technician on that machine on the glass caused breakage at that point.
had reported the failing thermocouple, we would Others tried lead disks, with similar effects. I
simply have replaced it and none of this would was on sabbatical leave at the nuclear research
1.6 The Hindelang conferences and the fused-silica column 7
institute in Karlsruhe, Germany, when a clerk in on WCOT columns. It was then that Golay
the supply room called my attention to a sheet of approached me; he was feeble at the time and
plastic heavily infused with graphite. After steadied by Ettre. He wanted to talk about J&W0 s
sliding Swagelok caps onto both column ends, method of coating columns. I explained that the
I wound small strips of this around both ends first step was his idea of feeding a filled column,
of a glass capillary, removed the stock ferules open end first, into a heated oven. Ilkova and
from the Swagelok connectors to the inlet and Mistrykov's idea of screwing the coiled column
detector, and connected the column. As I tight- into the oven showed real ingenuity: my pre-
ened the connections to a point where there heater contribution would have been proposed
was no leakage and found that the column was by anyone that could actually see the operation
still intact, I was delighted. The company that taking place. Now that I have reached (and prob-
made these graphitized sheets was just a few ably surpassed) Golay's age at the time, I believe
miles outside of Karlsruhe, and I paid them a I can now understand his curiosity, and I hope
visit. They were very cordial and gave me a he realized that another of his brilliant ideas
number of samples. On my return to the United had borne fruit. He departed courteously, and I
States, J&W ordered several steel dies dimen- was left to continue my attempts to visit every
sioned on the inner measurements of a 1/16 table and meet every participant. I did meet
inch Swagelok fitting and began pressing fer- virtually all the experts, absorbed a great deal
rules. These went on the market in 1975; they of knowledge, and left filled with new ideas
were a great success, and I have always regretted and a desire to get home quickly where I could
my failure to patent that product. get back to work in my own lab.
There are many types of glass, but we will
restrict our interests to those studied for their
1.6 The Hindelang conferences and the chromatographic interest by Dandeneau and
fused-silica column Zerenner: soda lime, borosilicate, uranium, pot-
ash soda lead, and fused silica (Table 1.1) [25].
The first symposium restricted to open- At the third Hindelang conference, I presented
tubular (capillary) columns was organized by a paper on multiple short-pass capillary chroma-
Rudolf Kaiser and chaired by Dennis Desty in tography, in which one of my Ph.D. students
1975; this was the first of the four Hindelang con- (James A. Settlage) had recycled a butane injec-
ferences and was held in an isolated village in tion 16 times through two 7.5 m columns to
the Bavarian Alps. Ettre visited Kaiser, first to achieve 1170 theoretical plates per second and
voice his opposition to holding the meeting, generated 850,000 theoretical plates in those 16
and then proposed a delay, arguing that it was passes in less than 12 min [26]; as I finished,
being held “far too early,” but many other several members of the audience rushed to the
workers in the field were enthused by the idea. podium with congratulations; at a final synopsis
I still regard the first Hindelang conference as of the meeting, Desty cited also this paper, but I
the most exciting meeting I have ever attended. knew they were all wrong. The most significant
It attracted nearly everybody that was working paper was that of Dandeneau and Zerenner
with glass capillary columns. We listened to pre- [25], releasing the fused-silica column. When
sentations from our associates in the morning Dandeneau presented his paper, very few people
and spent afternoons and evenings gathered realized its importance. Sandy Lipsky, who had
around tables seating perhaps eight participants, started Quadrex about the same time that J&W
drinking strong German brews and exchanging was founded, was one, and I was another. Both
views on chromatography, with the emphasis Lipsky and I had been trying to convert
8 1. Milestones in the development of gas chromatography
TABLE 1.1 Approximate compositions for glass materials used for capillary column fabrication.
Glass type SiO2 Al2O3 Na2O K2O CaO MgO B2O3 PbO U2O3 BaO
Soda lime 68 3 15 6 4 2 2
Borosilicate 81 2 4 13
Uranium glass 76 3 4 2 14 1
Potash soda lead 56 2 4 9 29
a
Fused silica 100
a
Typically contains less than 1 ppm total metals.
industrial analysts to WCOT glass capillaries, they changed to a polyimide; this seals surface
with but scant success. Industrial analysts real- flaws in the tubing. It is usually applied in
ized the superiority of results generated by these several thin coats during the drawing process.
columns, but they also recognized their fragility. Not everyone recognized the advantages of
Downtime is rarely critical in academia, but in the fused-silica column; indeed, in spite of my
industry it can be very serious; glass capillaries vigorous protests, my partners at J&W dismissed
break easily, packed columns lasted longer. it as “a gimmick.” 2 weeks later, it was impos-
However, this fused-silica column changed sible to give, let alone sell, a glass capillary col-
everything! Here was a capillary column that umn, and J&W made the conversion to fused
was both strong and flexible, but because it silica. At this point, things became very
was both labor-intensive and materials- “touchy.” The vast majority of scientists working
intensive, it was much more expensive. in gas chromatography had careers based on
Fused-silica columns have immense tensile glass capillary columnsddrawing, deactivation,
strength, permitting an extremely thin column coating, etc.dnow those careers were passe. I
wall, which in turn makes a flexible column. had a seminar scheduled for Milan, Italy, where
This permitted drawing long lengths of straight I was warmly receivedduntil my first slide was
tubing, which can then be coiled. Interfacing shown. It was a fused-silica column. Several at-
these to detectors and injectors was now a simple tendees left the room, and the atmosphere took
task; no longer must we flame straighten the on a chill. Very few in this audience wanted to
ends of the column, it was already inherently hear how their futures might be affected. Shortly
straight! But the outer surface of drawn silica re- thereafter, Georges Guiochon organized a chro-
quires protection. Like all silica-based glasses, matography symposium in Cannes, France.
fused silica is subject to flaws at its surface, and The GC column portion of the meeting was
flaws grow at a rate determined by stressdand chaired by Gerhardt Schomburg. As he called
coiling this long straight piece of fused silica pla- the meeting to order, he announced that there
ces it under stressdas the flaw grows larger, the would be “no discussion of fused-silica col-
column snaps at that point. To protect the outer umns.” As the meeting proceeded, several at-
surface, Hewlett Packard first coated the col- tendees expressed their objections to the
umns with a coating of polysiloxane, but this opening restriction and announced that they
ended up as a very sticky column. DuPont then had come to the meeting just to learn more about
came forward with a polyamide coating that the fused-silica column. With obvious distaste,
was applied during the drawing process. Later Schomburg turned to me and said, “tell them
1.7 Increasing sophistication of instrumentation 9
something about fused silica.” At this time, I had communicating with a central processor. With
been exploring on-column injections with fused the introduction of robotic autosamplers at
silica, sometimes trapping samples within the about the same time, the GC could now operate
fused-silica tubing by folding the flexible tube without human intervention, 24 h operation
into a “U” shape in a Dewar flask filled with became standard practice for routine analysis
liquid nitrogen, then withdrawing the “U” and in high sample throughput environments, and
reinserting it into a flask of heated oil for injec- GCs were deployed to remote locations and
tion. This approach demanded a flexible tube monitored electronically with only occasional
and would have been impossible with glass visits for service and routine maintenance. Elec-
capillary tubing. An obviously uncomfortable tronic pressure control was central to a series of
Schomburg interrupted me with the observation advances in programmed and large-volume in-
that “flexibility appeals only to Americans, and jection techniques, multidimensional-column
only because they are too clumsy to handle chromatography, and retention-time locked
glass.” A few months later, J&W was selling methods. The complex functions of gas chroma-
Schomburg fused-silica columns. tography had been reduced to those of a black
box analyzer 50 years after its invention, but still
evolutionary changes continue in technology
1.7 Increasing sophistication of with the view of minimizing the importance of
instrumentation the skill and knowledge of the operator in the
production of data [27].
Table 1.2 provides a timeline for important
developments in instrumentation [8,26,27]. The
essential elements of instruments were devel-
1.7.1 Column heating
oped by the early 1960s, with further develop-
ments occurring in short bursts of innovation Apart from the use of liquid thermostats in
and advances in technology followed by longer the early days of gas chromatography, the col-
periods of evolutionary changes and consolida- umn heater in a GC is typically a forced-
tion. Many advances were catalyzed by ad- circulation air oven, the temperature of which
vances in column technology or electronics. For can be changed in a controlled manner with
example, the advent of the microprocessor time for temperature-programmed separations.
brought about a radical change in instrument Good temperature control is essential to obtain
design and use. From this point onward, instru- reproducible retention times and to avoid peak
ment functions were monitored and controlled distortions associated with temporal and spatial
by networks of circuits communicating with oven temperature gradients. A low thermal mass
each other and with a central controller. This for the oven is also important since it allows
paved the way for the emergence of the rapid cooling after temperature programming.
keyboard instrument controlled by software Typically, forced air-circulation ovens can main-
running on a personal computer, which domi- tain a higher rate of heating at lower tempera-
nated instruments for laboratory use by the early tures than higher temperatures due to the
1990s. A significant milestone in achieving full greater amount of heat lost to the surrounding
automation of instrument operation was the environment as the temperature ramp pro-
introduction of electronic pressure control in gresses. For fast gas chromatography, some
the early 1980s. This allowed carrier gas and sup- manufacturers have addressed this problem
port gas pressure and flow rates to be set and with the oven inside an oven concept or more
monitored by electromechanical devices directly using resistive heating achieved through
10 1. Milestones in the development of gas chromatography
Resistive heating of columns introduced as an alternative means of temperature control to convection ovens
the use of a metallic heating element to transfer maximum detector sampling rate, detector sensi-
heat to the column by conduction [28]. Fast gas tivity and noise level, and sample introduction
chromatography makes use of short columns of (initial bandwidth). Contemporary laboratory
small internal diameter, thin film columns, instruments can usually meet the requirements
higher carrier gas flow rates, and fast tempera- for fast gas chromatography (first entry in
ture program rates (Table 1.3) [29]. For the fastest Table 1.3), but special-purpose instruments are
separations, the limiting instrument conditions required for very fast gas chromatography
become the available column inlet pressure, (second entry in Table 1.3). The entries in
maximum oven temperature program rate, Table 1.3 reflect what is within the capability of
commercial instruments with little modification
(first entry) and the limits of current technology
TABLE 1.3 Instrumental parameters for fast gas requiring special purposing of instruments
chromatography.
(second entry). From the perspective of separa-
(i) Fast gas chromatography (separation times <10 min) tion speed, the capability of gas chromato-
Peak widths 0.5e2 s graphic instruments limits performance more
Columns 5e15 m with I.D. 0.25e0.1 mm
Temperature program rates 20e60 C min1
so than column technology in contemporary
Column inlet pressure <15 bar Data acquisition rate 50 Hz practice.
Mobile phase velocity 1.4 uopt
vaporization chamber in which the evaporated large sample volumes (usually <1 mL) for trace
sample is mixed with carrier gas and divided be- analysis. The PTV injector was originally intro-
tween a stream entering the column (carrier gas duced in Europe in the 1980s, but it was over a
flow) and a stream vented to waste (split flow) decade later before it gained traction in the
was the first practical solution to this problem. United States. For much of this time it was an op-
Split injection can be used to generate extremely tion but not heavily promoted or supported by
small bandwidths for high peak capacity separa- instrument companies based in the United
tions but discriminates against less volatile com- States. This was probably due to a combination
pounds and the quantitative analysis of wide of commercial preferences and the not-
boiling-point mixtures is difficult, and for sam- invented-here syndrome.
ples present in dilute solution, detectability is The production of wide-bore fused-silica
limited by the small amount of sample trans- capillary columns coated with immobilized sta-
ferred to the column. The splitless injection tech- tionary phases in the early 1980s allowed the sy-
nique was devised (actually discovered ringe needle to be introduced directly into the
accidentally [30]) to overcome some of the defi- column and eliminated the problem of removal
ciencies of split injection for the analysis of mix- of the stationary phase by the relatively large
tures in dilute solution through the transfer of volume of solvent released inside the column
relatively large sample volumes to the column. by the syringe [32]. These advances in column
The gas flow through a splitless injector is rela- technology facilitated the development of cold
tively low, and the sample is introduced into on-column injection where the sample is intro-
the column over a comparatively long time duced as a liquid into the column inlet and sub-
(30e60 s), relying on cold trapping and/or sol- sequently vaporized. Discrimination because of
vent effects to refocus the compounds at the volatility differences was virtually eliminated
head of the column. The importance of these and the risk of sample decomposition mini-
refocusing mechanisms was not fully under- mized. With secondary cooling of the injector,
stood at first and much of our current knowl- the oven temperature could be kept well above
edge of the injection mechanism owes a great the boiling point of the solvent while maintain-
deal to the exemplary studies of Konrad Grob ing the column inlet at a much lower tempera-
and the publication in 1986 of his classic book ture. This is important for using on-column
on split and splitless injection [30]. injection in high-temperature gas chromatog-
The programmed-temperature vaporizer raphy. Dirty samples present a problem owing
(PTV) injector is perhaps the most versatile to contamination of the sample introduction
injector developed for gas chromatography. It fa- zone, which leads to poor chromatography and
cilitates split and splitless injection as well as new unreliable quantification. Both on-column and
approaches for large-volume injection with sol- PTV injection afford high accuracy and preci-
vent elimination [31]. The PTV injector has a sion. These injectors also facilitated the direct
low thermal mass to allow rapid heating and coupling of other chromatographic systems to
cooling. The sample is introduced at a relatively gas chromatography, such as liquid chromatog-
low temperature and then raised ballistically to a raphy and supercritical fluid chromatography
temperature sufficient for rapid volatilization of [33]. In the 1980s, Carlo Erba manufactured an
the highest-boiling sample components. Slow instrument for on-line liquid chromatographye
sample introduction at a low temperature with gas chromatography, but the uptake was poor
solvent elimination facilitates the injection of and the system was discontinued.
1.7 Increasing sophistication of instrumentation 13
The array of sample introduction methods for in wet chemical processes, is less used today.
gas chromatography and an understanding of Interfacing of modern detectors to capillary col-
the mechanistic details on which they rely were umns is not normally difficult except for fast
complete by the end of the 1980s and modifica- gas chromatography where the detector volume
tions since then have been evolutionary. The and data acquisition rates limit the use of some
most visible change is the shift from manual to detectors. The flame ionization detector facili-
automated sample introduction systems that tated many of the early developments in capil-
accept samples in vials and can be programmed lary columns (Section 1.4). This detector has a
for sequential analysis of each or selected vials in low dead volume, a high sensitivity for nearly
a batch. The introduction of solid-phase microex- all carbon-containing compounds, and an
traction in the early 1990s simplified sample extremely wide linear response range. Other de-
preparation and handling steps using a syringe tectors were developed in response to the need
containing a retractable sorbent fiber compatible for element-selective or structure-selective detec-
with solvent-free sample introduction [34]. By tion. These detectors allowed the analysis of
the turn of the century, this had become one of target compounds at low concentrations in com-
the most popular sampling/sample introduction plex samples and made an important contribu-
techniques in gas chromatography. It can be seen tion to the rapid acceptance of gas
as the forerunner of the liquid-phase microex- chromatography in laboratories performing
traction techniques developed over the last routine analysis as well as research laboratories.
decade or so [35]. Together these microextraction A special mention should be made of the
formats are responsible for achieving a better coupling of gas chromatography with mass
scaling of sample size requirements to sample spectrometry (GCeMS). This coupling dates
utilization capabilities of capillary columns and almost from the beginning of gas chromatog-
are having a considerable impact on laboratory raphy, but in the early days the practical prob-
working practices. lems and high cost meant that the combination
was confined to a few research laboratories.
When packed columns dominated the practice
1.7.3 Detectors of gas chromatography, separators were
Gas chromatography is blessed by a number required to reduce typical column flow rates to
of robust and sensitive detectors based on gas- the vacuum-handling capacity of the ion source
phase ionization processes (e.g., flame ioniza- of the mass spectrometer. Separators based on
tion, thermionic ionization, electron capture, a jet, glass fritted tube, or diffusion membrane
and photoionization), bulk property (thermal design allowed sample enrichment while simul-
conductivity), optical (flame photometric, chem- taneously reducing the gas flow to the ion
iluminescence, and atomic emission), electro- source. Separators disappeared nearly
chemical (electrolytic conductivity), and completely with the introduction of fused-silica
advanced spectroscopic (mass spectrometry open-tubular columns, which could be routed
and infrared and far-UV spectroscopy) detection from the column oven through a heated transfer
principles. Many of these detectors were intro- line directly into the ion source of the mass spec-
duced during the initial phase of the develop- trometer. During the time that instrumentation
ment of gas chromatography and are still used for gas chromatography was advancing, so
today in a modified form reflecting changes in were all aspects of mass spectrometry, which
enabling technology. The exception is the (Hall) became more powerful, affordable, reliable,
electrolytic conductivity detector, which, and available to laboratories with low-skilled
because of its large detector volume and basis personnel. The great advantage of the mass
14 1. Milestones in the development of gas chromatography
spectrometer as a detector is its ability to provide personal computer with information being
structural information for identification using archived locally and/or centrally on a laboratory
different ionization approaches and through information network server. Today, the large
techniques such as tandem mass spectrometry. amount of data produced by capillary columns
Another advantage is its high sensitivity and (especially when coupled to a mass spectrom-
selectivity as a detector in the selected ion or eter) can be handled by a personal computer,
high-resolution mode. which simultaneously functions as the system
manager controlling and monitoring all aspects
of the chromatograph. What has not changed is
1.7.4 Data handling that the methods used to manipulate the chro-
In the early days of gas chromatography, peak matographic data for display and calculation
quantification was done by purely manual mea- are considered proprietary information,
surements of peak heights, although it was requiring an act of faith that what is provided
recognized that peak area measurement was as a printout or on screen is exactly what
fundamentally better but more difficult. The happened in the column [36]. What is clear, how-
product of the peak height and the peak width ever, is that it would be nearly impossible to
at half height could be used as an approximate persuade scientists to return to the status quo
area measurement, but this was neither fast nor before the electronic age of data handling as lab-
(usually) accurate, especially for narrow peaks. oratory productivity would be severely limited.
Approaches based on planimetry and cut-and-
weigh procedures could afford reasonable accu-
1.7.5 Comprehensive gas
racy but demanded considerable skill to achieve
chromatography
acceptable results and were slow and tedious.
Electromechanical devices such as the ball-and- Sometimes the separating power of gas chro-
disk integrators and integrating amplifiers were matography, even when augmented by the iden-
early attempts at automating this time- tification power of spectroscopic detectors such
consuming process. These devices were rather as mass spectrometry, is still insufficient to accu-
limited in range and speed of response. The rately determine the composition of a sample.
advent of microprocessors in the early 1970s ush- This problem is tackled in contemporary studies
ered in the computing integrator, which took using comprehensive gas chromatography.
over the labor-intensive process of data manage- Comprehensive separations have their origins
ment, eventually resulting in devices that could in multidimensional gas chromatography devel-
record a chromatogram simultaneously with oped in the early 1960s after the introduction of
data acquisition and, when the separation was the Dean's switch for controlling the flow direc-
complete, instantly generate reports in tabular tion of a gas stream at a T-piece using pressure
form of all sorts of descriptive information changes [37]. Modern microfabricated versions
from the chromatogram, including the calcula- of this device remain at the heart of many devel-
tion of peak areas. It was not long before main- opments in multicolumn separation systems.
frame computers were in use to handle data Multidimensional (typically two-dimensional)
produced from a number of instruments simul- gas chromatography employs two independent
taneously. The development of the personal columns connected to each other by an interface
computer and the rapid growth in processor [38]. In early versions a fraction of the chromato-
speed and memory, however, had a greater gram, a heartcut, from the first column was iso-
impact and resulted in data management being lated by cold trapping and then released by
handled by software running on a dedicated thermal desorption for separation on the second
1.8 Decline in the expertise of the average gas chromatographer 15
column. The disadvantage of this approach is time measured in seconds. This requires incred-
that the fraction collected at the interface con- ibly small injection bandwidths, short second
tains no information about the separation on columns of small internal diameter, fast detector
the first column. In the 1980s, Siemens marketed response times, and repetitious and fast changes
a remarkable instrument for its time, the in instrument operating conditions. Comprehen-
SiChromat-2, which had two independently sive gas chromatography generates a large
controlled air-circulation ovens, housing the amount of data, especially when mass spectrom-
two columns, connected by a T-piece to isolate, etry is used as a detector, and challenges remain
focus, and reinject fractions from the first column in how best to make these data available to the
to the second by live switching based on the user in a form convenient for analysis. In these
Deans' switching mechanism. This instrument instruments, one can perhaps see a vision of
likely contained the most complex pneumatic the future of single-column gas chromatography
system of any instrument of its generation and (based on technology developed for the second-
required considerable skill in its operation. It dimension separation) while at the same time
had to compete for preference with a new gener- highlighting the limitations of current technol-
ation of affordable and easy-to-use gas ogy for gas chromatography.
chromatographyemass spectrometry systems,
however, and was not a commercial success. In
the early 1990s, Phillips introduced comprehen- 1.8 Decline in the expertise of the average
sive gas chromatography in which the whole gas chromatographer
separation on the first column was transferred
to a second column of different selectivity, the At present, the success in the commercial
two columns being connected through a modu- sector in gas chromatography has resulted in a
lation interface [39]. In this case, the chromato- situation not all that different from a narcotic de-
gram is recorded as a two-dimensional contour pendency. To be classed as an expert in gas chro-
plot with the planar axes consisting of the inde- matography, all that seems to be necessary is the
pendent retention times for each compound on wherewithal to purchase a GC and a column.
the two columns and the vertical axis is related Today, those tasked to operate a GC do not
to the amount of each compound. This technique need to even make an injection, columns have
was not an instant success, but after further re- become classified as laboratory supply items,
finements in modulator design, a decade later and data as something that appears on a printout
it had entered the main stream of commerce. at the end of an experiment. When I was first
These instruments are at the cutting edge of in- introduced to gas chromatography in the
strument development where capabilities are 1970s, it was necessary to make, modify, and
pushed to match column performance. The func- continuously improve columns and instruments
tion of the modulator is to arrest, concentrate, to meet the needs of each application. Expert
and launch a fraction of the material contained chromatographers of this time were of necessity
within each peak of the chromatogram from well versed in troubleshooting and thoroughly
the first column and transfer them as a series of knowledgeable of the principles and practice at
pulses of constant frequency to the second col- what often appeared to outsiders as the dark
umn [40]. To avoid overlap of individual separa- arts of separations. On account of the immense
tions on the second column, each chromatogram powers of separations, virtually any unskilled
must be complete in less time than the cycle time individual can produce data today; even the
of the modulator. Thus, the second-column sep- village idiot can inject a sample and obtain peaks
arations must be very fast with a total separation that are interpreted as “usable data.” This has
16 1. Milestones in the development of gas chromatography
led to a continuing decline in the expertise of the [8] L.S. Ettre, The early development and rapid growth of
average practicing chromatographer from the gas chromatographic instrumentation in the United
States, J. Chromatogr. Sci. 40 (2002) 458e472.
mid-1980s to the present time. This can be peril- [9] M.J.E. Golay, in: D.H. Desty (Ed.), Gas Chromatog-
ous, because everything from column selection raphy 1958 (Amsterdam Symposium), Butterworths,
to troubleshooting skills is based on a funda- London, 1958, pp. 139e143.
mental knowledge of chromatographic princi- [10] J.E. Lovelock, S.R. Lipsky, Electron affinity spectros-
ples, the absence of which degrades the quality copy e a new method for the identification of func-
tional groups in chemical compounds separated by
and usefulness of the information acquired by gas chromatography, J. Am. Chem. Soc. 82 (1960)
these instruments. To address these problems re- 431e433.
quires a massive educational effort before the [11] J. Harley, W. Nel, V. Pretorious, Flame ionization de-
knowledge is lost and the usefulness of gas chro- tector for gas chromatography, Nature 181 (1958)
matography to decision-makers is called into 177e178.
[12] I.G. McWilliam, R.A. Dewer, in: D.H., Desty (Eds.), Gas
question. When I first started to make my way Chromatography 1958 (Amsterdam Symposium), But-
in gas chromatography, I never thought that terworths, London, 1958, p. 142.
such a reversion would occur so quickly, and [13] I.G. McWilliam, The origin of the flame ionization
as retirement approaches, that intellectually I detector, Chromatographia 17 (1983) 241e243.
might leave the field in a poorer state than [14] L.S. Ettre, The invention, development and triumph of
the flame ionization detector, LC GC 15 (2002)
when I entered it. There is no doubt that ad- 364e373.
vances in technology have delivered better in- [15] K. Grob, Hansjurg Jaeggi e obituary, J. High Resolut.
struments and columns but now in the hands Chromatogr. 12 (1989) 459.
of less qualified operators; this does not always [16] D.H. Desty, J.N. Haresnape, B.H.F. Whyman, Con-
lead to better-quality data. struction of long lengths of coiled glass capillary,
Anal. Chem. 32 (1960) 302e304.
[17] K. Grob, The role of column technology in capillary gas
References chromatography, J. High Resolut. Chromatogr. 7 (1984)
252e257.
[1] A.T. James, A.J.P. Martin, Gaseliquid partition chro- [18] M.J.E. Golay, in: D.H. Desty (Ed.), Gas Chromatog-
matography: the separation and microestimation of raphy 1958 (Amsterdam Symposium), Butterworths,
volatile fatty acids from formic to dodecanoic acid, Bio- London, 1958, pp. 36e55.
chem. J. 50 (1952) 679e690. [19] E.L. Ilkova, E.A. Mistrukov, A simple versatile method
[2] L.S. Ettre, The beginnings of gas adsorption chroma- for coating glass capillary columns, J. Chromatogr. Sci.
tography 60 years ago, LC-GC N. Am. 26 (2008) 48e51. 9 (1971) 569e570.
[3] F. Prior, Determination of Adsorption Heats of Gases [20] M.C. Bourne, W.G. Jennings, Existence of two soil species
and Vapors by Application of the Chromatographic in detergency investigations, Nature 197 (1963) 1003.
Method in the Gas Phase, Doctoral theses, University [21] M.C. Bourne, W.G. Jennings, Kinetic studies of deter-
of Innsbruck, Innsbruck, Austria, 1947 (in German). gency. I Analysis of cleaning curves, J. Am. Oil
[4] O. Bobleter, Exhibition of the first gas chromatographic Chem. Soc. 40 (1963) 517e523.
work of Erika Cremer and Fritz prior, Chromatogra- [22] M.C. Bourne, W.G. Jennings, Kinetic studies of deter-
phia 43 (1996) 444e446. gency. II Effect of age, temperature, and cleaning
[5] A.N.H. Ray, Gas chromatography, IeII, J. Appl. Chem. time on rates of soil removal, J. Am. Oil Chem. Soc.
4 (1954) 21e25, 82e85. 40 (1963) 523e530.
[6] L.S. Ettre, A. Zlatkis (Eds.), 75 Years of Chromatog- [23] L. Blomberg, Current aspects of the stationary phase in
raphy e a Historical Dialogue, Elsevier, Amsterdam, gas chromatography, J. High Resolut. Chromatogr. 5
1979. (1982) 520e533.
[7] H.A. Laitinen, G.W. Ewing, A History of Analytical [24] J.K. Haken, Developments in polysiloxane stationary
Chemistry, Division of Analytical Chemistry, Amer- phases in gas chromatography, J. Chromatogr. 300
ican Chemical Society, Washington, 1977, pp. 303e304. (1983) 1e77.
Further reading 17
[25] R.D. Dandeneau, E.H. Zerenner, An investigation of [36] N. Dyson, Chromatographic Integration Methods, The
glasses for capillary chromatography, J. High Resolut. Royal Society of Chemistry, Cambridge, UK, 1998.
Chromatogr. 2 (1979) 351e356. [37] K.M. Sharif, C.T. Chin, C. Kulsing, P.J. Marriott, The
[26] W. Jennings, J.A. Settlage, R.J. Miller, Multiple short microfluidic Deans switch: 50 years of progress, inno-
pass glass capillary gas chromatography, J. High Reso- vation and application, Trends Anal. Chem. 82 (2016)
lut. Chromatogr. 2 (1979) 441e443. 35e54.
[27] K.D. Bartle, P. Myers, History of gas chromatography, [38] P.Q. Tranchida, D. Sciarrone, P. Dugo, L. Mondello,
Trends Anal. Chem. 21 (2002) 547e557. Heart-cutting multidimensional gas chromatography:
[28] A.Z. Wnag, H.D. Tolley, M.L. Lee, Gas chromatog- a review of recent evolution, applications, and future
raphy using resistive heating technology, prospects, Anal. Chim. Acta 716 (2012) 66e75.
J. Chromatogr. A 1261 (2012) 46e57. [39] C. Meinert, U.J. Meierhenrich, A new dimension in sep-
[29] C. Cruz-Hernandez, F. Destaillates, Recent advances in aration science: comprehensive two-dimensional gas
fast gas chromatography: application to the separation chromatography, Angew. Chem. Int. Ed. 51 (2012)
of fatty acid methyl esters, J. Liq. Chromatogr. Relat. 10460e10470.
Technol. 82 (2009) 1672e1688. [40] H.D. Bahaghighat, C.E. Freye, R.E. Synovec, Recent ad-
[30] K. Grob, Classical Split and Splitless Injection in Capil- vances in modulator technology for capillary two
lary Gas Chromatography, second ed., Huethig, Hei- dimensional gas chromatography, Trends Anal.
delberg, 1986, 1992. Chem. 113 (2019) 379e391.
[31] W. Engewald, J. Teske, J. Efer, Programmed tempera-
ture vaporizer-based injection techniques for injection
in capillary gas chromatography, J. Chromatogr. A
856 (1999) 259e278. Further reading
[32] K. Grob, On-column Injection in Capillary Gas Chro-
matography, second ed., Huethig, Heidelberg, 1991. [1] A. Gerontas, New technologists of Research in the 1960s:
[33] K. Grob, On-line Coupled LCeGC, Huethig, Heidel- the case of the reproduction of automated chromatog-
berg, 1991. raphy specialists and practioners, Sci. Educ. 23 (2014)
[34] J. Pawliszyn, Solid-phase Microextraction: Theory and 1681e1700.
Practice, Wiley-VCH, New York, 1997. [2] G.M. Gross, V.R. Reid, R.E. Synovec, Recent advances in
[35] C.F. Poole (Ed.), Liquid-phase Extraction, Elsevier, instrumentation for gas chromatography, Curr. Anal.
Amsterdam, 2020. Chem. 1 (2005) 135e147.
C H A P T E R
2
Theory of gas chromatography
Leonid M. Blumberg
Advachrom, Wilmington, DE, United States
Gas Chromatography
https://doi.org/10.1016/B978-0-12-820675-1.00026-5 19 © 2021 Elsevier Inc. All rights reserved.
20 2. Theory of gas chromatography
Chromatograph
Inlet Column Outlet
Sample
Data
Sample introduction Detector Detector signal Information
Analysis
device
Gas flow
Connecting
assemblies
opt optimal parameter, except for speed-optimized Kc Distribution constant, Eq. (2.6)
conditions
L Column length
R parameter at retention time
[ Dimensionless column length, Eq. (2.50)
ref reference parameter
M Molar mass
st parameter at standard temperature and pressure
N Plate number (apparent), Eq. (2.121)
N Intuitive plate number, Eqs. (2.117) and (2.125)
2.2.3 Symbols (local symbols used only n Running peak capacity, Eq. (2.180)
shortly before or after their introduction Dn Incremental peak capacity, Eq. (2.180)
are not listed) nc Peak capacity of chromatographic analysis,
Eq. (2.180)
Dpst Solute diffusivity in gas at standard pressure and pg Gauge pressure, Eq. (2.26)
arbitrary temperature Pi Performance parameter, Eq. (2.240)
DS Solute diffusivity in stationary phase pst Standard pressure, 1 atm
F Gas flow rate, Eq. (2.43)
Q Transport efficiency, Eqs. (2.126) and (2.153)
f Gas specific flow rate, Eq. (2.45) ℛ Molar gas constant, 8.31447 J/(K mol)
fcrit Critical specific flow rate, Eq. (2.54) Resolution of two consecutive peak,
Rs
G Gibbs free energy, Eq. (2.7) Eq. (2.172)
22 2. Theory of gas chromatography
RT Heating rate, Eq. (2.63) Xr Sensitivity of DTR to relative change in rT, Eq.
(2.90)
rT Dimensionless heating rate, Eq. (2.69)
XT Parameter in Eqs. (2.167) and (2.220)
s Running separation capacity, Eq. (2.176)
Xq Sensitivity of DTR to Dqchar, Eq. (2.89)
Ds Separation of two peaks, incremental separation
capacity, Eqs. (2.175) and (2.178) Xs Gas parameter, Table 2.6
sc Separation capacity of chromatographic analysis, z Distance from column inlet
Eq. (2.177)
b Phase ratio, Eq. (2.8)
Dschar Characteristic separation capacity, Eq. (2.197)
z Dimensionless distance from column inlet,
T Temperature Eq. (2.31)
Tchar Characteristic temperature, Eq. (2.13) h Gas viscosity
DTchar Difference in characteristic temperatures of two hst Gas viscosity at standard temperature
solutes
qchar Characteristic thermal constant, Eq. (2.13)
DTM Void temperature increment, Eq. (2.73)
Dqchar Difference in characteristic thermal constants of
Tnorm Normal temperature, 298.15K (25 C) two solutes
TR Solute elution temperature qcharG Generic characteristic thermal constant
Eq. (2.18)
DTR Difference in elution temperatures of two solutes
L Linear dynamic range, Eq. (2.256)
Tst Standard temperature, 273.15K (0 C)
l Mean free path
t Time
m Solute mobility, Eq. (2.3)
tanal Analysis time
meff Solute effective mobility, Eq. (2.60)
tM Void time, Eq. (2.48)
mR Mobility of eluting solute
tm Gas propagation time, Eq. (2.47)
x Gas empirical parameter, Table 2.2
tM,char Void time at characteristic temperature
rn Noise spectral density
tM,R tM in static analysis under conditions existed at
time tR in dynamic analysis s Peak width (standard deviation) in time
domain
tM,ref Void time at reference temperature
e
s Width (standard deviation) of a solute zone
tm,sol Solute mobile time, Eq. (2.58)
sT Peak width (standard deviation) in temperature
tmol Mean time between molecular collisions
domain
tR Retention time sz Temporal width of migrating solute zone
DtR Retention time difference of two solutes y Average molecular speed
tR,stat Retention time in static analysis Function defined in Eq. (2.42)
F
u Gas velocity, Eq. (2.28) 4 Dimensionless film thickness, Eq. (2.9)
u Gas average velocity, Eq. (2.38)
U Gas pneumatic resistance, Eq. (2.30)
v Solute velocity
u Solute immobility, Eq. (2.3)
Xall Parameter in Eq. (2.255) u Distance-averaged immobility, Eq. (2.87)
XD Gas parameter, Table 2.2 w Parameter defined in Eq. (2.101)
Xi Parameter in Eq. (2.238) wS Parameter defined in Eq. (2.102)
2.4 Soluteecolumn interaction 23
2.3 General definitions and conventions (migration). A solute migrating through a column
might interact with the column stationary phase
Conditions of GC analysis are static (as in (might be partially absorbed by liquid stationary
isothermal isobaric GC) if they do not change phase or adsorbed on solid stationary phase). As
with time. Otherwise, the conditions are dynamic a result, the net velocity,
(as in temperature- and/or pressure-
dz
programmed GC). The conditions could be uni- v¼ (2.1)
form (the same at any place along the column dt
length) or nonuniform. Gas decompression along of migration of the zone at some location (z)
the column is a typical cause of nonuniform con- along the column can be smaller than the carrier
ditions [2] causing gas velocity to increase in the gas local velocity (u) at the same location. The
direction from the inlet to the outlet. relationship between v and u can be expressed as
Throughout this chapter, temperature (T) can
be expressed in degrees Celsius ( C) or in units of v ¼ mu (2.2)
absolute temperature (in kelvin, K). However, T in where m is the mobile phase fraction of the solute
all equations is assumed to be the absolute tem- (defined below).
perature unless the contrary is explicitly stated. Different solutesdcomponents of a sampled
On the other hand, temperature intervals and might be differently distributed between the sta-
the differences between two absolute tempera- tionary phase and the inert carrier gas in a col-
tures can be expressed in C units. umn. As a result, different solutes might
Two types of open-tubular columns (OTC) also migrate through the column with different ve-
known as the capillary columns are used in GC. locities (v) and elute from the outlet at different
The internal walls of the wall-coated open- retention timesdthus being separated from each
tubular (WCOT) column are coated with other. This means that.
absorbing liquid polymer film. The internal
Different distribution of different solutes between the
walls of the porous-layer open-tubular (PLOT)
stationary phase and the carrier gas resulting from
column are covered with a solid porous layer
the different interaction of the solutes with the sta-
having large adsorbing surface. The WCOT col-
tionary phase in a column is the root cause of the so-
umns play the dominant role in GC, and only
lute separation.
they are considered in this chapter. In this
context, the terms OTC, WCOT column, capil- The distribution of a solute between the col-
lary column, and column are synonyms. Typi- umn phases can be expressed in several equiva-
cally, the thickness of a stationary liquid film in lent ways [2]:
OTC is much smaller than the column tubing in- astat amob astat
ternal diameter (dc). In these cases, the difference k¼ ; m ¼ ; u ¼ (2.3)
amob a a
between dc and the diameter (d) of the internal
open space can be ignored. where a ¼ amob þ astat, amob and astat are, respec-
tively, the total amount of the solute and its
2.4 Soluteecolumn interaction amounts in the mobile and the stationary phases.
Parameters k, m, and u are, respectively, the
2.4.1 Solute distribution between mobile solute retention factor [2,20] (capacity factor [4,5],
capacity ratio [20], partition ratio [3,7,17]), mobility
and stationary phases
[2] (originally introduced by Consden et al. [21],
Let z and t be, respectively, the distance trav- it is also known as the retardation factor [6,20], the
eled by a solute zone (a packet of solute mole- retention ratio [5], and the frontal ratio [4]), and
cules) from the inlet and the time of that travel immobility [2] (interaction level [2]).
24 2. Theory of gas chromatography
It follows directly from their definitions in (GLC) and as partition GC [24]. This is the most
Eq. (2.3) that parameters k, m, and u relate to frequently used type of GC by far and the subject
each other as of primary attention in this chapter.
The equilibrium of a solute distribution be-
1m u 1
k¼ ¼ ; m ¼ 1u ¼ ; tween two phases (stationary liquid polymer
m 1u 1þk phase and mobile gas phase in partition GC)
k can be described by the distribution constant
u ¼ 1m ¼
1þk [2,20] ( partition constant [3e5,7,8,24,25]):
(2.4)
Cstat
and are bound by conditions: Kc ¼ (2.6)
Cmob
0 k N; 0 m 1; 0u1 (2.5)
where Cmob and Cstat are the concentrations of
Parameters m and u are complementary to the solute in the mobile and the stationary phase,
each other. If m ¼ 0, then u ¼ 1. Conversely, if respectively. The dependence of Kc on tempera-
u ¼ 0, then m ¼ 1. This justifies the complemen- ture (T) can be described by the integrated van't
tary terms for these parameters such as the Hoff equation representing an ideal thermodynamic
mobility (m) and the immobility (u). The latter equilibrium model [5,26] (Fig. 2.2A):
represents a measure of a solute interaction
with a column stationary phase (its affinity to G G GH
Kc ¼ eg ; g ¼ ¼ Sþ
the phase). As such measure, the solute immo- T ℛ ℛT (2.7)
bility (u) is similar to the retention factor (k). ðideal thermodynamic modelÞ
However, u is a normalized measure (changing
from 0 to 1) while k is not. where G, GS, GH, and g are the Gibbs free energy,
As parameters k, m, and u are interdependent entropy, enthalpy, and dimensionless Gibbs free en-
with each other, either one is sufficient for all ergy [2,23], respectively, of a solute transfer
evaluations that depend on a solute distribution from stationary to mobile phase, and
between the phases. However, a need for the ℛ ¼ 8.31,447 J/(K mol) is the molar gas constant.
simplicity and transparency of interpretation of As an example, parameters GS and GH of several
theoretical results justifies preferential choice of solutes are listed in Table 2.1. Generally, T can be
one parameter over the others depending on nonuniform, i.e., it can be different at different
the type of evaluations. Thus, the mobility (m) locations (z) along the column. In this case, Kc
most directly affects, Eq. (2.2), the solute velocity is also nonuniform, i.e., it is a function Kc(z) of
(v); the immobility (u) most directly affects their z. This condition exists in, e.g., thermal gradient
separation [22,23]; and the retention factor most GC (discussed later). It is assumed throughout
directly relates to thermodynamics of a solutee this chapter that, unless the contrary is explicitly
column interactiondthe next topic. stated, T and Kc are uniform.
It follows from Eqs. (2.3), (2.6), and (2.7) that
Kc eg stationary phase volume
k¼ ¼ ; b ¼
2.4.2 Characteristic parameters of a b b gas volume
soluteecolumn interaction (2.8)
GC utilizing columns with liquid stationary where b is the phase ratio. Generally, the
phase is known as gaseliquid chromatography stationary-phase film thickness and/or
2.4 Soluteecolumn interaction 25
(A) (B)
25 15
c32 c30
c20
20
10
g = lnKc
15
c10 5
lnk
10
c10 0
5
-5
0
2 2.5 3 300 400 500 600
1/T (1000/K) T (K)
FIGURE 2.2 Thermodynamic properties of n-alkanes in 5% diphenyl-dimethyl polysiloxane polymer: (A) dimensionless
Gibbs free energies (g, Eq. 2.7) of a solute transfer from stationary to mobile phase versus inverse temperature (1/T), (B) lnk
versus T (Eq. 2.10 at 4 ¼ 0.001). Within shaded area in (B) where 0.1 k 10, a tangent line to lnk(T)-curve at k ¼ 1 closely
approximates the curve itself. From L.M. Blumberg, Temperature-Programmed Gas Chromatography, Wiley-VCH, Weinheim,
2010. Adapted with permission.
thermodynamic properties can be nonuniform and GH) is not the only known thermodynamic
[28]. In this case, the numerator and the denom- model in GC [2,5,29e34]. More complex three-
inator of Eq. (2.8) represent local volumes in nar- parameter models [31e34] can provide more
row vicinity of each coordinate z, and b can be accurate retention time prediction. However, as
nonuniform. However, it is assumed throughout stated in the Introduction, accurate retention
this chapter that b is uniform. time prediction is out of scope of this chapter.
For capillary column having its internal walls Simpler models are sufficient for theoretical
coated with thin liquid stationary-phase film, study of performance of GC analysesdthe
b can be expressed as main focus of this chapter.
1 df The model in Eq. (2.7), although relatively
bz ; 4 ¼ (2.9) simple, has its own shortcomings. Its two param-
44 d
eters, GH and GS, change with temperature and
where df and 4 are, respectively, the stationary- solute concentration. The dependence of GH
phase film thickness and dimensionless thickness and GS on a solute concentration can be negli-
[2]. Eq. (2.8) becomes (Fig. 2.2B): gible if stationary-phase overloading is avoided
k ¼ 44eg (2.10) when relatively small sample amounts are
analyzed. It is assumed throughout this chapter
which, in the case of ideal thermodynamic that, unless otherwise explicitly stated, a column
model, Eq. (2.7), becomes is not overloaded and the solute concentration
GH GS does not affect parameters GH and GS. As for
ln k ¼ þ lnð44Þ the dependence of these parameters on tempera-
RT R
ðideal thermodynamic retention modelÞ ture, it is negligible if the solute migration is eval-
uated within a relatively narrow temperature
(2.11)
range (less than 50 C [2]) around the tempera-
The ideal thermodynamic model of Eq. (2.7) ture at which the parameters were found
(two-parameter model based on parameters GS (measured, calculated, etc.).
26 2. Theory of gas chromatography
TABLE 2.1 Thermodynamic parameters of several solute and the tangent line to that curve at
solutes in 5% phenyl-dimethyl polysi- k ¼ k0 is small (Fig. 2.2B) so that, in column per-
loxane polymer [34]. Also included are formance evaluations, the tangent line can be
the solute molecular weights (M) and
boiling points (Tb). used as a linear retention model [2] described
below.
M [27] GH GS Tb [27] Tchar qchar The linear model has chromatographic justifi-
# Name g/mol kJ/mol J/mol/K C C C cation. During the heating ramp covering a wide
temperature range, all solutes elute with approx-
1 Heptane 100.2 30.2 49.3 98.4 43.7 27.7 imately equal retention factors (kR) [2,14]
2 Octane 114.2 34.5 55.2 125.7 67.8 28. (see also Section 2.7.6). The solutes also migrate
3 Nonane 128.3 38.8 61.5 150.8 88.4 28.
through the major portion of a column length
with retention factors that are not much larger
4 2-Octanone 128.2 40.2 60.3 173.6 105.7 29.7 than 10kR [2]. This suggests that a linear reten-
5 Decane 142.3 43.1 67.5 174.1 107.3 27.9 tion model can be used for evaluation of the
6 2-Octanol 130.2 41. 61.9 180. 107.5 29.4
key elution parameters of the solutes in
temperature-programmed GC analyses. A
7 1-Octanol 130.2 43.9 65.7 194.4 120.1 29.3 choice of k0 ¼ 1 leads to the simplest transforma-
8 Undecane 156.3 40. 55.1 196.8 122.5 32.6 tion (see further) of parameter GS and GH into
9 5-Nonanol 144.3 44.6 66.5 193. 123.5 29.3
parameters of a linear retention model, and
vice versa [2]. It also helps that, at optimal or
10 2-Nonanol 144.3 45.2 67.6 193 125.4 29.2 near-optimal heating rate, retention factors of
11 1-Nonanol 48.1 71.3 136.9 29.1 eluting solutes are not far from unity [2].
12 5-Decanol 48.8 72.2 139.7 29.1
Let lnk(T) be some (not necessarily ideal)
model of a soluteecolumn interaction where
13 2-Decanone 156.3 48.5 71.4 210.1 140.3 29.3 functions lnk(T) for all solutes are roughly
14 Dodecane 170.3 43.7 59.8 216.3 140.4 32.5 similar to the ones in (Fig. 2.2B). The tangent
15 Tridecane 184.4 47.5 64.6 235.4 156.8 32.3
line at k ¼ 1 for each solute can be treated as
the solute linear retention model [2]:
16 Tetradecane 198.4 51.3 69.4 253.7 171.7 32.1
T Tchar
17 Pentadecane 212.4 55. 74.1 270.6 185.3 31.8 ln k ¼ ðlinear retention modelÞ
qchar
Sorted by ascending Tchar. Tchar, and qchar are at 4 ¼ 0.001. (2.12)
where
Even when its parameters (GH and GS) are
1
fixed quantities, the use of the ideal thermody- dðln kðTÞÞ
namic model in theoretical studies of Tchar ¼ Tjk¼1 ; qchar ¼
dT k¼1
temperature-programmed GC is highly prob-
(2.13)
lematic because it, in Giddings' words [35], “al-
ways leads to rather formidable integrals Both parameters (Tchar and qchar) of the
whose solutions are not easily obtained” straight line in Eq. (2.12) are measured in units
[14,17]. Fortunately, within a moderately wide of temperature. Their examples for several sol-
range, 0.1k0 k 10k0, of retention factors (k) utes are listed in Table 2.1.
around some predetermined retention factor k0, Quantity Tchardthe intercept of the line in
the difference between actual lnk(T)-curve for a Eq. (2.12)dis the characteristic temperature of the
2.4 Soluteecolumn interaction 27
soluteecolumn interactiondthe temperature at • Functions Kc(T) ¼ Exp(G/T), lnKc (T) ¼ G/T,
which lnk ¼ 0. In a typical isothermal and and k(T) ¼ Exp(G/T)/b, Eqs. (2.7) and (2.8),
temperature-programmed analysis, the solutes represent interdependent quantities (Kc and
generally elute in the order of increasing their k), but all have different slopes. Thus, the
Tchar. slope of Kc(T) is different from those of lnKc(T)
Parameter qchar is the characteristic thermal con- and k(T). Not only that, but the slope of the
stant of the interaction. The slope of the line in same parameter depends on the way it is
Eq. (2.12) is equal to 1/qchar. Following are described. Thus, the slope of Kc(T) is different
the reasons for describing the soluteecolumn from that of lnKc(T). On the other hand, the
interaction via parameter qchar rather than via characteristic thermal constant (qchar) of all
the slope. these quantities is the same. When T changing
from Tchar to Tchar þ qchar, both Kc and k
• Quantity qchar is measured in temperature
become e-times smaller regardless of
units, while the slope is measured in less
equations describing these parameters as
intuitive inverse temperature units.
functions of T.
• Temperature (T) increase reduces k. As a
result, lnk(T) has a negative slope, while Example 2.1. Quantity qchar typically ranges
parameter qchar is conveniently a positive between 25 and 40 C [2] (see also examples in
quantity equal to the temperature increase Table 2.1). If qchar ¼ 30 C, then raising T from
that reduces k in Eq. (2.12) by a factor Tchar to Tchar þ 30 C reduces both Kc and k by
e z 2.72. a factor of 2.72, raising T by 1 C, from Tchar to
• From a more general perspective, k is an Tchar þ 1 C, reduces both Kc and k by (1 C)/
almost exponentially declining function of T. (30 C) z 3.3%, in order to reduce k or Kc by a
In some disciplines (physics, medicine, etc.) factor of 2, T should be raised from Tchar by
the exponentially (or almost exponentially) 30 C ln2 z 21 C. The latter is close to pub-
declining quantities (mass, concentration, lished experimental results [35e37].
strength, etc.) are characterized by the half-life
The topic of this chapter is theoretical evalua-
timedthe time during which the quantity
tion of the factors affecting performance of GC
declines by a factor of 2. (It is easier to
analyses and their optimization. Among these
comprehend that the half-life strength of
factors are the effects of a column temperature
some medicine, e.g., is 12 h rather than that
and heating rate not on each particular solute,
the slope of reduction of natural logarithm of
but general trends in such effects on all solutes.
the strength of the medicine is 0.06 h1).
The linear retention model is sufficiently accu-
In other disciplines (electronics, mechanics,
rate for such evaluations and is the only known
etc.), quantities exponentially declining with
model that leads to broad theoretical solutions
time are characterized by the time
[2]. The theory reviewed in this chapter is based
constantdthe time during which the quantity
on that linear retention model.
declines by a factor of e z 2.72. As a concept,
The significance of Tchar and qchar goes beyond
the time constant is less intuitive than the
their role of parameters in linear retention model
half-life time. On the other hand, the time
of Eq. (2.12). Their definitions in Eq. (2.13) are not
constant is more convenient to handle
related to any particular model and are valid for
mathematically. They relate to each other as
any model. Quantities Tchar and qchar can serve as
(half-life time) z 0.7(time constant). The
parameters in an ideal two-parameter thermody-
characteristic thermal constant (qchar) is
namic model, as well as parameters of a three-
analogous to the time constant.
parameter thermodynamic model [34].
28 2. Theory of gas chromatography
If parameters, GH and GS, of the ideal two- Several applications of parameters Tchar and
parameter model in Eq. (2.7) are known, then qchar are demonstrated in this chapter. Among
parameters Tchar and qchar could be found from them are evaluations of optimal heating rate,
Eqs. (2.11) and (2.13) as [2]: prediction of reversal of a solute elution order
as a function of the heating rate, evaluation of
GH
Tchar ¼ ; diffusivity of an arbitrary solute with a known
GS ℛ lnð4fÞ Tchar. Outside this chapter, the characteristic
(2.14)
ℛ GH parameters were also used for prediction of the
qchar ¼
ðGS ℛ lnð4fÞÞ2 peak distribution maps in GC GC [38]. The
concept of characteristic parameters extended to
These transformations were used for calcula-
LC (liquid chromatography) was used for evalu-
tion of Tchar and qchar in Table 2.1. If necessary,
ation of optimal mixing rate (the rate of program-
Tchar and qchar could be transformed back to GH
mable change in solvent composition) in gradient
and GS [2]:
LC. This suggest that characteristic parameters of
ℛTchar
2
T a solute-column interaction can serve as a com-
GH ¼ ; GS ¼ ℛ char þ lnð4fÞ mon ground for representation of different reten-
qchar qchar
tion mechanisms. There are other applications for
(2.15)
these parameters. One of them could be quantita-
Substitution of Eq. (2.15) in Eq. (2.11) yields tive evaluation of the effects of column aging. In a
[2]: recent publication [39], the effect of a column
heating to high temperature on parameters GH
Tchar Tchar
ln k ¼ 1 and GS of several solutes was measured and tabu-
qchar T (2.16) lated. Transformation of these parameters into
ðideal thermodynamic modelÞ Tchar and qchar using Eq. (2.14) led to a chromato-
which is the same ideal thermodynamic model graphically meaningful conclusion that the heat-
as the one in Eq. (2.11) only differently ing reduced Tchar by 3e4 C and reduced qchar
expressed. Eq. (2.16) makes it obvious that, as by 4%e6%. For a solute eluting during a typical
mentioned earlier, k ¼ 1 at T ¼ Tchar and that heating ramp, reduction of Tchar by some DTchar
(1/qchar) is the slope of lnk at T ¼ Tchar. results in about the same reduction in elution
Having clear chromatographic interpretation, temperature [2], reduction in qchar is an indication
characteristic parameters Tchar and qchar provide of proportionally higher sensitivity of k to the
insight into chromatographic properties of the change in column temperature, Eq. (2.13).
solutes (their expected elution temperatures in As follows from Eq. (2.14), characteristic pa-
temperature-programmed analyses, dependence rameters depend on film thickness. If two col-
of the temperatures on the heating rate, etc.). umns have the same stationary-phase type of
This cannot be said about thermodynamic different film thicknesses, then Eqs. (2.14) and
parameters GH and GS in Eq. (2.11). While hav- (2.15) can be used for transforming Tchar and
ing clear thermodynamic interpretation, they qchar of each solute in one column into these pa-
tell little about chromatographic properties of rameters in another column. However, such
the solute. Thus, the fact that decane in Table 2.1 transformations are somewhat cumbersome [2].
has GH z 43.15 kJ/mol and GS z 67.5 J/mol/K A simple approximate transformation for Tchar
tells little about, say, its elution temperature in can be obtained from the approximation [2]:
a typical temperature-programmed analysis, 0:07
Tchar2 42
while Tchar ¼ 107.3 C indicates that decane will ¼ (2.17)
Tchar1 41
elute in such analysis at about 107.3 C.
2.5 Properties of ideal gas 29
It shows that Tchar is a weak function of the Tchar (ºC)
film thickness. Thus, doubling the film thickness, 26.85 126.85 226.85 326.85
i.e., doubling 4, causes about 5% increase in Tchar
of each solute. Similar transformation for qchar is (A)
40
described later.
TcharG
30
θchar (ºC)
0
of GC analyses and beyond. As a general trend,
the solutes with larger boiling points have larger (B)
40
Tchar. The values of characteristic thermal con-
TcharG
stant (qchar) are typically confined within
30
25e40 C range [2] (see also examples in
Table 2.1). For a given solute, parameters Tchar 20
and qchar are independent of each other and
essentially are the parameters of the solute chro- 10 10000 solute-liquid pairs,
matographic identity in relation to a give col- θchar,st = 22.3 ºC, rstd = 10.5 %
umn. However, there is a general trend, 0
300 400 500 600
Fig. 2.3. The solutes having larger Tchar (and
Tchar (K)
eluting at higher temperatures) have generally
larger qchar. In addition to that, film thickness FIGURE 2.3 Maps of (Tchar, qchar) pairs (the dots) for
slightly affects qchar. The trends in Fig. 2.3 can (A) 2412 combinations of more than 1000 solutes in eight liquid
polymers [2] and (B) about 10,000 combinations generated
be described as [2]:
from the data in Refs. [30,40]. In all cases, 4 ¼ 0.001. Each dot
in (B) represents one of 212 solutes in one of 50 stationary
Tchar 0:7 phases characterized in Ref. [40]. (Some 20 or so solutes repre-
qcharG ¼ qchar;st ; sented in (B) appear in both sources [30,40].) Quantity qchar,st
Tst (2.18)
0:09 listed in each graph corresponds to respective least square fits,
qchar;st ¼ 22 C$ 103 4 solid lines, of Eq. (2.18). The dashed lines mark 30% distances
from the solid lines, rstd is relative standard deviation of verti-
where qcharG is the expected value of qchar for the cal departures of the dots from the solid line. Panel A from L.M.
solute with a given Tchar. A solute for which Blumberg, Temperature-Programmed Gas Chromatography, Wiley-
qchar ¼ qcharG (i.e., its actual qchar is equal to the VCH, Weinheim, 2010. Adapted wiht permission.
expected qchar for the solute's Tchar) is called equal diameter, but twice as thick stationary
herein as the generic one [2]. In view of that, phase of the same type has qcharG z 32 o C.
qcharG can be interpreted as the characteristic
thermal constant of a generic solute and can be
called as the generic characteristic thermal constant. 2.5 Properties of ideal gas
Example 2.2. In a column with 4 ¼ 0.001, a 2.5.1 Theoretical relations
generic solute having Tchar ¼ 423.15K (150 o C)
has qcharG z 30 o C. A (possibly different) generic It is assumed throughout this chapter that the
solute that has the same Tchar in a column of carrier gas is an ideal gas [2,26,41]. This means
30 2. Theory of gas chromatography
that relationship between its mass (mg), pressure The values of some of these parameters and
( p), temperature (T), and volume (V) is governed others, introduced later in this section, are listed
by the ideal gas law: in Table 2.2.
ℛ mg T
pV ¼ (2.19)
M
where M is the gas molar mass. Several molecular 2.5.2 Viscosity
properties of ideal gas are important for GC. Among gas parameters in Eq. (2.20), only the
Among them are average molecular speed (y), average molecular speed (y) can be calculated
mean free path (l), mean time between collisions from a priori known molar mass (M) of the gas.
(tmol), self-diffusivity (Dg), and viscosity (h). They Other parameters are expressed via the gas vis-
relate to each other and to other gas properties cosity (h) d a measure of its resistance to flow.
as [2,26,41,42]: Ideally, as is assumed throughout this chapter,
rffiffiffiffiffiffiffiffiffiffiffi h is independent of pressure. Although theoret-
8ℛT 4yh ph
y¼ ; l ¼ ; tmol ¼ ; ical expressions for h are known, they are either
pM 5p 4p complex or not sufficiently accurate for practical
3p hy2 applications [26,41,43]. Instead, empirical ex-
Dg ¼ $ (2.20) pressions are used in GC [2,43e45].
20 p
Gas He H2 N2 Ar
Molecular properties:
M (molar mass), g/mol 4.003 2.016 28.01 39.95
lst (mean free path at p ¼ pst, T ¼ Tst), mm 0.177 0.112 0.0604 0.0641
yst (average molecular speed at T ¼ Tst), km/s 1.20 1.69 0.454 0.38
hst (viscosity at T ¼ Tst), mPa,s 18.69 8.362 16.84 21.35
Dg,st (self-diffusivity at p ¼ pst, T ¼ Tst), cm /s 2
1.26 1.12 0.162 0.144
Empirical parameters:
x (parameter in Eq. 2.21) 0.685 0.698 0.710 0.750
XD (parameter in Eqs. 2.24 and 2.25) 0.456 0.544 0.693 0.686
Combinations:
XD lst, mm 0.081 0.061 0.042 0.044
(XD lst)/(XD lst) hydrogen 1.33 1 0.689 0.721
XD yst, km/s 0.548 0.92 0.315 0.261
(XD yst)/(XD yst) hydrogen 0.596 1 0.342 0.284
2
XD2 Dg;st , cm /s 0.262 0.331 0.078 0.068
.
D/Dhydrogen ¼ XD2 Dg;st XD2 Dg;st hydrogen (Eqs. 2.24 and 2.25) 0.79 1 0.24 0.21
8. Koralli.
9. Kangashärkä.
Kun kuopan sivut ovat tulleet niin luisuiksi ja sileiksi, ettei mikään
hyönteinen pääse kiipeämään niitä ylöspäin, kaivautuu toukka
kuopan pohjalle, niin että siitä näkyy vain uhkaavat leuvat, jotka
selkoselälleen avattuina odottavat saalista.