Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
High-speed narrow-bore capillary gas chromatography :theory, instrumentation and applicationsCitation for published version (APA):Ysacker, van, P. G. (1996). High-speed narrow-bore capillary gas chromatography : theory, instrumentation andapplications. Technische Universiteit Eindhoven. https://doi.org/10.6100/IR468539
DOI:10.6100/IR468539
Document status and date:Published: 01/01/1996
Document Version:Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)
Please check the document version of this publication:
• A submitted manuscript is the version of the article upon submission and before peer-review. There can beimportant differences between the submitted version and the official published version of record. Peopleinterested in the research are advised to contact the author for the final version of the publication, or visit theDOI to the publisher's website.• The final author version and the galley proof are versions of the publication after peer review.• The final published version features the final layout of the paper including the volume, issue and pagenumbers.Link to publication
General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright ownersand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.
• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal.
If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, pleasefollow below link for the End User Agreement:www.tue.nl/taverne
Take down policyIf you believe that this document breaches copyright please contact us at:[email protected] details and we will investigate your claim.
Download date: 11. Sep. 2020
HIGH SPEED NARROW BORE CAPILLARY
GASCHROMATOGRAPHY
Theory, Instrumentation and Applications
PROEFSCHRIFT
ter verkrijging van de graad van doctor
aan de Technische Universiteit Eindhoven,
op gezag van de Rector Magnificus,
prof.dr. M. Rem,
voor een commissie aangewezen door
het College van Dekanen
in het openbaar te verdedigen op
donderdag 31 oktober 1996 om 16.00 uur
door
Peter Gilbert Van Y sacker
geboren te Roeselare (België)
Dit proefschrift is goedgekeurd door de promotoren:
prof.dr.ir. C.A.M.G. Cramers
en
prof.dr. P.J.F. Sandra
Copromotor: dr.ir. J.G.M. Janssen
Van Y sacker, Peter Gilbert
High-speed narrow-bore capillary gas chromatography
Theary instrumentation and applications I
Peter Gilbert Van Ysacker.- Eindhoven:
Eindhoven University ofTechnology
Thesis Eindhoven University ofTechnology.
ISBN 90-386-0468-8
Table of Contents
1 GENERAL INTRODUCTION AND SCOPE 1 References . . . 5
2 FACTORS DETERMINING THE SPEED OF ANALYSIS IN GAS CHROMATOGRAPHY. 7
2.1 Introduetion . . 8
2.2 Golay-Giddings equation and routes towards faster separations 11
2.3 Parameters affecting the chromatographic efficiency . 15
2.3.1 Influence ofthe column inside diameter. 15 2.3.2 Influence ofthe carrier gas. . 17 2.3.3 Influence of the outlet pressure . 19 2.3.4 lnfluence ofthe flow profile . 21 2.3.4.1 Turbulent flow conditions . 21 2.3.4.2 Coiling induced secondary flow. 23 2.3.5 Packed columns versus capillary columns . 25 2.3.6 Conclusions . 26
2.4 Practical consequences of the use of narrow-bore capillary columns. 26
2.4.1 Detection limits. 27 2.4.2 Sample capacity 29 2.4.3 Instromental band broadening 30
2.5 Conclusions 32
2.6 References . 34
3 NON-SPLITTING INJECTION TECHNIQUES FOR NARROW-BORE CAPILLARY GAS CHROMATOGRAPHY. 37
3.1 Introduetion . 38
ii
3.2 Overview of injection devices enabling high-speed separations 39
3.3 Concept of non-splitting injection techniques 42
3.4 Overview of non-splitting injection techniques for narrow-bore
capillary chromatography . . 43
3.5 Experimental. . . . 45 3.5.1 Practical considerations on hot splitless injections 46 3.5.2 Practical considerations on cold splitless injections. 48 3.5.3 Practical considerations on on-column injections. 49
3.6 Results and discussion . 50 3 .6.1 Splitless times . 50 3.6.2 Liner capacity for splitless injections . 52 3.6.3 Focusing effects on narrow-bore columns 54 3.6.4 Focusing effects of retention gaps in cold splitless and on-column
injections 58 3.6.5 Discrimination. 61 3.6.6 Thermal degradation. 66
3. 7 Conclusions 68
3.8 References . 69
4 HIGH-SPEED NARROW-BORE CAPILLARY GAS CHROMATOGRAPHY COUPLED TO ELECTRON CAPTURE DETECTION 73
4.1 Introduetion . 74
4.2 Overview of detection devices in combination with high-speed
narrow-bore columns . . 76
4.3 Electron capture detection in capillary chromatography 78
4.4 Instrumentation 79
4.5 Theory 80
4.6 Results and discussion . 84 4.6.1 Detection band broadening 85 4.6.2 Sensitivity, detection limits and dynamic range 86 4.6.3 Applications. 90
4. 7 Conclusions 93
4.8 References . 94
Table of contents iii
5
5.1
5.2 5.2.1 5.2.2 5.2.3 5.2.3.1 5.2.3.2
5.2.3.3 5.2.3.4 5.2.3.5 5.2.4
5.3 5.3.1 5.3.2 5.3.3 5.3.3.1 5.3.3.2 5.3.3.3 5.3.4
5.4 5.4.1 5.4.2 5.4.3 5.4.3.1 5.4.3.2 5.4.3.3 5.4.3.4
5.5
5.5.1 5.5.2
5.6
5.7
HIGH-SPEED NARROW-BORE CAPILLARY GAS CHROMATOGRAPHY COUPLED TO V ARIOUS MASS SPECTROMETRIC DETECTION METBODS 97
Introduetion . 98
Ion trap mass analyser. . 100 Introduetion. . 100 Experimental 101 Results and discussion . 1 04 Instrumental parameters 104 Influence ofthe helium background pressure on mass resolution and sensitivity. • • . • 1 04 Deleetion limits and dynamic range 107 Quality of mass spectra.. 108 Applications . . . 108 Conelusions. . . . . 112
Fast scanning double focusing sector instrument . 113 Introduetion . 113 Experimental. 113 Results. 115 Scan speed . 116 Detection limits and dynamic range 120 Applications. • • • 126 Conelusions. 127
Time-of-Oight mass analyser 129 Introduetion. 129 Instrumentation. 129 Results. . 134 Scan speed . 134 Mass resolution and quality of mass spectra 13 7 Applications . . . 141 Conclusions . • . 143
Comparison of different mass analyser as detectors for high-speed narrow-bore capillary gas chromatography . . . . 145 Other mass analysers. . . . • . . 145 Comparison of several mass spectrometers as detectors for high-speed gas ehromatographie separations. 147
Conclusions 151
References . 152
iv
6 COMPREHENSIVE TWO-DIMENSIONAL GAS CHROMATOGRAPHY. . 155
6.1 Introduetion . . 156
6.2 Experimental. . 158
6.3 Results . . 161
6.4 Conclusions . 164
6.5 References. . 164
SUMMARY. . 167
SAMENVATTING . . 171
LIST OF SYMBOLS .. . 175
LIST OF ABBREVIATIONS . . 179
DANKWOORD ..... . 181
CURRICULUM VITAE . . 183
BffiLIOGRAFIE . . . . . 185
Chapter 1
General introduetion
and scope
Chromatography was discovered in 1906 by Michael Tswett, a Russian botanist,
when he was attempting to separate coloured leaf pigments by passing a solution
containing them through a column packed with adsorbent chalk particles [1]. Since
these early experiments many scientists have made substantial contributions to the
theory and practice of chromatography [2,3].
Gas chromatographic separations were first described by Martin and James in 1952
[4]. This technique has become a widely used separation technique for the analysis of
mixtures of gases or volatile liquids and solids. lts impact on modem analytica!
chemistry has clearly been immense; gas chromatography (GC) has been used to
solve a large number of significant problems in medicine, biology and environmental
sciences, as well as in an impressive number of industrial applications. In spite of the
developments in analytica} chemistry, GC remains one of the most widely used
analytica! tools. No other analytical technique can provide equivalent resolving
power and low concentration detection limits for such a wide variety of compounds
[5].
Since the introduetion of GC the performance of columns and chromatographic
equipment has considerably improved. The major breakthrough was the introduetion
2 Chapter 1
of open tubular columns by Golay in 1958 [6]. Cornpared to packed columns the
better perrneability of capillary columns allows a substantially higher obtainable
plate nurnber. It lasteduntil the end ofthe seventies, when Dandeneau and Zerenner
introduced fused silica columns, before capillary columns became widely accepted
[7]. The flexible and strong fused silica columns allow easy installation in the gas
chromatograph, in contrast to glass columns which were too fragile. Additionally,
deactivation methods have become available which provide an excellent inertness of
the column inner wall. This enables the analysis of a large range of substances,
including very polar or unstable ones. The stationary phases nowadays are
remarkably stabie and allow sample introduetion of a few microlitres of liquid
sample onto the column. Moreover, a wide variety of columns is available, including
columns of different inside diameters, stationary liquid phases with different
polarities, and adsorption based stationary phases. The combination of high
resolution, speed of analysis and sensitive detection has made GC a routine
technique used in almost every chemicallaboratory.
Current research topics in GC are focused on the further refinement of the separation
methods and the hyphenation to other chromatographic techniques and spectroscopie
methods in order to improve the performance with regard to selectivity, detectability,
identification and separation speed.
Since the introduetion of capillary GC, there has been a demand for increased speed
of analysis. Already in the early sixties, a practically feasible approach towards faster
and/or higher chromatographic resolution separations was suggested by Desty [8]. In
the seventies, several theoretica! studies established the influence of various
parameters for obtaining faster and/or more efficient chromatographic separations.
Despite many improvements in the quality of capillary GC columns and GC
equipment, it is surprising to note how little has been done to realise the true
potentials of high-speed separations. The lack of adequate instrumentation has
hindered the application of high-speed separations in daily practice. Only Gaspar et
al. [9,10] and the Eindhoven group [11-13] have studied the potentials and
limitations of high-speed GC in detail.
Nowadays, the industrial and governmental demands for faster and more detailed
analyses of trace impurities in complex mixtures become more pronounced. The
General introduetion and scope 3
tremendous growth in the number of samples to be analysed urgently calls for an
increased speed of analysis in combination with highly sensitive and selective
detection methods for virtually all application areas of GC. To meet these
requirements, many groups became a ware of the possibilities offered by high-speed
GC and started research projects on this topic.
The objective of this thesis is to briefly review the theoretica} aspects of the various
methods for high-speed GC and to investigate the possibilities and limitations of
several injection and detection devices to exploit the true potentials offered by
narrow-bore capillary columns for fast chromatographic separations.
In chapter 2, the various routes towards shorter analysis times are discussed in detail.
The basic common background of these methods is that they enhance radial mass
transfer equilibration inside the column. The effects of using reduced inside diameter
columns, hydrogen carrier gas, vacuum outlet conditions, turbulent flow and the use
of geometrically deformed separation tubes will be evaluated. Additionally, the
practical consequences conceming the detection limits, sample capacity and the
instrumental band broadening of these methods will be emphasised.
Sample introduetion is one of the most critica} steps for the successful application of
high-speed separations. Several injection devices that yield small input band widths
suffer from the disadvantage that only very small sample volumescan be introduced
onto the separation column. Hence, the minimum detectable concentration is far too
low in daily practice. Other injection techniques based on non-splitting principles of
sample introduetion have to be used in order to improve the minimum detectable
concentration. In chapter 3, several of these techniques are discussed in detail. Hot
splitless injections in combination with narrow-bore columns suffer from several
disadvantages. To overcome these problems other injection techniques including
cold splitless and on-column injection are evaluated and guidelines for optimisation
are discussed. The performance of these sampling techniques is illustrated by various
applications.
Since the chromatographic peak broadening of narrow-bore columns is very small,
also the peak broadening caused by the detector must be extremely small to preserve
a high column efficiency. Moreover, a considerable increase in the sampling
frequency of the data acquisition system is required for accurate registration of the
4 Chapter 1
chromatogram. Finally, due to the low loadabilities of narrow-bore columns, a very
sensitive detection device is required to preserve an acceptable working range. In chapter 4, the combination of high-speed narrow-bore capillary GC with electron
capture detection (ECD) is evaluated. Because the sensing volume of the ECD is
relatively large, the use of high make-up flow rates is required to eliminate peak
tailing. In the ECD, however, the make-up gas actively participates in the detection
mechanism. F or this reason, it is important to investigate the influence of the higher
make-up flow rates on the sensitivity and the detection limits ofthe ECD.
The combination of gas chromatography with mass speetrometry is without any
doubt the most powerfut hyphenated technique for the separation and identification
of unknown samples. In chapter 5, the compatibility of various types of mass
spectrometers with narrow-bore columns is investigated. Special emphasis is placed
on scan rates, detection limits, mass resolution and quality of the spectra obtained.
Quadrupole instruments suffer from a number of disadvantages, the most important
one being the limited sensitivity and the relatively long time required to complete a
speetral scan. Ion trap mass spectrometers and sector instruments are inherently more
sensitive detectors. Additionally, for sector instruments, a higher selectivity can be
obtained by using high mass resolving powers. The scan speed of these mass
scanning instruments is rather limited. With a time-of-flight mass analyser,
theoretically a spectrum can be recorded in less than 100 ~sec. Hence higher
acquisition rates can be obtained. Two time-of-flight mass spectrometers are
evaluated. The differences between both instruments are explained and the
consequences of the experimental set-up is described in detail. In order to have a
complete overview of the potentials of all mass spectrometers as detection device for
high-speed GC, the possibilities of some other mass spectrometers are briefly
summarised. The performance of the mass spectrometers is compared to each other
and the most suitable for high-speed GC is selected. Further more, several
modifications are suggested to improve the performance of this instrument.
In chapter 6, a system for comprehensive two-dimensional gas chromatography is
developed and evaluated. The first column generates a normal speed, one
dimensional chromatogram. This column is serially coupled to a second column that
generates a series of high-speed chromatograms as the separation in the first column
General introduetion and scope 5
proceeds. By suitably selecting the stationary phase of the second column,
orthogonal separations can be obtained. In this way high peak capacities are possible
thereby enabling the analysis of complex mixtures in a relative short time.
REFERENCES
1. V.G. Berezkin, Chem. Revs., 89 (1989) 277. 2. L.S. Ettre, A. Zlatkis (eds.), in "75 Years of Chromatography. A Historica}
Dialogue", Elsevier, Amsterdam, 1979. 3. F. Bruner (Ed.), in "The Science ofChromatography", Elsevier, Amsterdam, 1985. 4. A.T. James, A.J.P. Martin, Biochem. J., 50 (1952) 679. 5. C.F. Poole, S.K. Poole, Anal. Chim. Acta, 216 (1989) 109. 6. M.J.E. Golay, in "Gas Chromatography", V.J. Coates et al.(Eds.), Academie Press,
New York, 1958. 7. R.D. Dandeneau, E.H. Zerenner, J. High Resolut. Chromatogr. Chromatogr. Comm.,
2 (1979), 351. 8. D.H. Desty, A. Goldup, W.T. Swanton, in "Gas Chromatography", N. Brenner, J.E.
Callen, M.D. Weiss (Eds.), Academie Press, New York, 1962, (Lansing Symposium 1961), p. 105.
9. G. Gaspar, R. Annino, C. Vidal-Madjar, G. Guiochon, Anal. Chem., 50 (1978) 1512.
10. G. Gaspar, C. Vidal-Madjar, G. Guiochon, Chromatographia, 15 (1982) 125. 11. C.P.M. Schutjes, Ph.D. Thesis, Eindhoven University of Technology, the
Netherlands, 1983. 12. C.A. Cramers, P.A. Leclercq, CRC Crit. Rev. Anal. Chem., 20 (1988) 117. 13. A.J.J. van Es, Ph.D. Thesis, Eindhoven University of Technology, the Netherlands,
1990.
6
Chapter 2
Factors determining the speed of analysis
in gas chromatography
SUMMARY
The performance of a gas chromatographic system in Jast separations depends both
on the column efficiency as well as on the design of the equipment. The analysis
time, under normalised conditions, is a complex function of amongst others,
stationary phase selectivity, solute dijfusion coefficients in the mobile and the
stationary phase, flow profile, column dimensions, etc. A theoretica/ study shows the
importance of the various parameters involved in the optimisation of a GC system.
The most efficient means to reduce the analysis time is to use open tubular narrow
bore columns and hydragen as carrier gas. Reduced outlet pressure conditions only
result in a considerable gain in analysis time for wide-bore and/or short columns.
The practical use of coiled columns or turbulent flow conditions as a means of
reducing the analysis time is limited. Bestdes a.ffecting the analysis time, the
varfation of the column dimensions also strongly affects the minimum detectable
amount, the sample capacity, the minimum detectable concentra/ion and the
requirements imposed on the instrumental design.
8 CMprer2
2.1 INTRODUCTION
The objective of chromatography is the separation, identification and quantitation of
compounds in a mixture. The separation process in gas chromatography (GC) is
based on the partitioning of the solutes between two phases, the stationary phase
(solid or liquid) and the mobile phase. Each compound is distributed between the
mobile phase and the stationary phase in a characteristic way, and hence is travelling
through the column at its own characteristic speed.
Therefore, the solutes will elute at the end of the column at different times,
designated by the total retention time tR, which can be written as:
(2.1)
where k is the retention factor which is related to the distribution constant K
governing the distribution process and the phase ratio ~ which represents the ratio of
the volume of the mobile phase over that of the stationary phase. tM is the gas hold
up time which can be calculated according to:
(2.2)
where L is the column length and u the average linear carrier gas velocity. As a
sample traverses a column, its zone width increases in proportion to its migration
distance or time spent in the column. The extent of zone broadening determines the
chromatographic efficiency, which can be expressedas either the plate number N, or
the plate height H. If it is assumed that the chromatographic peak bas a Gaussian
distribution, then the column efficiency can be expressed in terms of the peak
retention time and the standard deviation of a chromatographic peak cr according to:
(2.3)
The height equivalent to a theoretica! plate can be calculated according to:
H=L/N (2.4)
Factors determining the speed of analysis in gas chromatography 9
A property called peak resolution, Rs, is used to describe quantitatively the
separation between two adjacent peaks. Rs is defined as:
(2.5)
where tR,t and tR.2 are the total retention times ofthe fust peak and the second peak,
respectively. cr1 and cr2 are the standard deviations for the first and the second peak.
i\tR is the difference in total retention time between the two adjacent peaks. For two
closely adjacent peaks, cr1 and cr2 are almost equal and can be substituted by cr, the
average standard deviation of the two peaks. For peaks of equal size, baseline
separation is achieved at Rs > 1.5, but separation is usually considered to be
sufficient for Rs > 1.
The resolution of two adjacent peaks is related to the adjustable chromatographic
variables of selectivity, efficiency and time. The substitution of equations (2.1) and
(2.3) into equation (2.5) leads to:
R = 1 (a-1) N s 4 (l+k2) a
(2.6)
where k2 is the retention factor of the later eluting peak of the critical pair and a is
the separation factor of the two solutes. This latter parameter can be calculated
according to:
(2.7)
where k1 is the retention factor of the first eluting peak. The separation factor is a
constant for a given set of analytica! conditions.
By rearranging equation (2.6) it is possible to calculate the plate number required,
Nreq. to obtain a given chromatographic resolution Rs according to:
N = 16R2(~)2(1+ k2)2 req s l k a- 2
(2.8)
10 Chapter 2
This equation shows how important it is to optimise the chromatographic resolution,
the separation factorand the retention factor [1]. It is clear that in order to obtain a
good separation, conditions in which k2 is small as well as a separation factor close
to one must be avoided.
Once the desired resolution is specified and the required number of plates is
calculated, the analysis time for separating a critical pair can be calculated by
combining equation (2.1), (2.2), (2.4) and (2.6). The analysis time is given by:
(2.9)
Equation (2.9) was derived as such by Ettre [2] in 1973 and already conceived by
Pumell [3] and Giddings [4] in the early 1960s. When the objective is to perform fast
separations, equation (2.9) states that an overly large value of the resolution will
unnecessarily proiongate the analysis time. By differentiation of equation (2.9), an
optimal value of k2 ~::> 2 is calculated, assuming that H is independent of k, which is
only approximately true. Additionally, the separation factor a should be maximised
for the pair most difficult to separate by careful selection of the stationary phase and
temperature. The evaluation of the ratio H/U: is very complicated since there are
numerous operational and extra column parameters affecting the velocity of the
mobile phase and the corresponding plate height.
In the next paragraphs, the various strategies for optimisation of the analysis speed
will be discussed in detail. The influence of various operational parameters
determining the speed of analysis will be descri bed. The main aim of this chapter is
to compare the various strategies in terms of practical applicability. As most of the
theory is already discussed by several authors, it will only be briefly summarised
here. Additionally, the operational parameters that determine the speed of analysis
also affect the minimum detectable amount, the sample capacity of the column, the
minimum detectable concentration and the requirements imposed on the instrumental
design. This will be discussed as well.
Factors determining the speed of analysis in gas chromatography 11
2.2 GOLAY-GIDDINGS EQUATION AND ROUTES TOWARDS FASTER
SEPARATIONS
The chromatographic separation is counteracted by dispersive processes which cause
band broadening of the compound zones during migration of the solutes through the
column. It is known from theory [5,6] that peak dispersion is dependent on
instrumental parameters. The relative band broadening caused by the
chromatographic process in a capillary column is accurately described by the Golay
equation [6]. Later on, this formula was extended to take pressure drop over the
column into account [7]. The resulting equation has demonstrated to give an accurate
estimate of the chromatographic band broadening. Including the decompression
effects, the plate height equation for a uniformly coated column reads:
(BMo )
H = ~ + CM,oUo f1 + Csu0 f2 (2.10)
with
BM,o =2 DM,o (2.11)
CM,o K _ll_k_2_+_6-::-k_+_l_d--=~- = K f(k) d~
(l+k)2 2DM,o 2DM,o (2.12)
Cs (2.13)
In these equations Uo is the linear velocity of the carrier gas under outlet pressure
conditions, de the inside diameter of the capillary column, dr the film thickness of the
stationary phase, DM,o the diffusion coefficient of a component in the mobile phase at
column outlet pressure and Ds the diffusion coefficient of the component in the
(liquid) stationary phase. K is a function of the shape of the velocity profile which is
1/48 for straight columns and is called the velocity profile factor.
Defining the relative pressure P = p/p0 as the ratio of the inlet pressure Pi over the
outlet pressure Po of the column, the Giddings [7] and James-Martin [8] pressure
drop correction factors can be calculated according to:
12
f _ 9 (P4 - 1) (P2
- 1) l- 8 (P3 -1)2
f _ 3 (P 2 -1) 2-2 (P3 -1)
Chopter 2
(2.14)
(2.15)
For smalt pressure drops over the column (P~l), ft and f2 approach unity. For
columns operated at large pressure drops (P~oo ), ft will increase to its maximum
value of 9/8. In this situation f2 decreases drastically and can be approximated by
3/2P. The factor f2 relates the carrier gas velocity at the column outlet, u0 , to the
average linear carrier gas velocity, u by [8]:
(2.16)
For gases showing ideal gas behaviour, the carrier gas velocity under column outlet
conditions in case of viscous laminar flow through a capillary column with an inside
diameter dç can be calculated from the Hagen-Poiseuille equation:
(2.17)
where 11 is the mobile phase viscosity.
The plate height in a chromatographic separation depends on several dispersive
broadening processes. The first term in equation (2.1 0) represents the contri bution
from longitudinal diffusion. The contribution of this term to the overall band
broadening is reduced considerably at higher veloeities of the mobile phase. The
second and third term describe the resistance to mass transfer of the solute in the
mobile phase and the stationary phase, respectively. The stationary phase mass
transfer term becomes increasingly important as the film thickness increases.
Increasing the film thickness can considerably reduce the number of theoretica}
plates. Because of their high retention factors, thick-film columns will only be the
column type of choice for the analysis of very volatile compounds.
Factors determining the speed of analysis in gas chromatography
---u
Figure 2.1
reduced column diameter
A
increased diffusion coefficient
B
flat velocity profile and increased radial transport by convection
c .~
13
Schematic overview of the various routes towards faster separations. A: Reduction of the column inside diameter, B: Increasing the diffusion coefficient by applying vacuum outlet conditions or by using hydrogen as carrier gas, C: enhancement of the radial dispersion by geometrically deformed tubes or turbulent flow conditions.
There are various routes to increase the analysis speed in gas chromatography. These
possibilities are schematically illustrated in Figure 2.1. In capillary columns
operating under normal chromatographic separation conditions, the flow profile is
laminar. In this situation, fluid flows through the column in layers where two
adjacent elemental gas volumes move parallel, each with its own characteristic
velocity. The velocity profile of the carrier gas in the separation tube is parabolic.
The velocity close to the column wall is almost zero while the gas velocity is
maximised in the centre of the tube. This velocity gradient results in
chromatographic band broadening. Inequalities in the migration veloeities of the
compounds in the gas stream have to be overcome by radial diffusion. The
14 Chapter 2
characteristic time constant 1: for radial diffusion coefficient of a compound can be
calculated according to:
(2.18)
As can be seen from this equation, there are various routes to reduce the time
required for radial equilibration. First of all, the time constant can be reduced by
reducing the column inside diameter (option A). In this way, the diffusion distance is
reduced so that the time required for exchange of compounds between fluid layers in
the gas stream is shorter. In this situation 1: is reduced considerably.
Another approach to reduce the time constant is to increase the diffusion coefficient
of the compound in the mobile phase ( option B). This can be obtained by applying
vacuum outlet conditions. Under vacuum outlet conditions, the average pressure
inside the column is reduced. As the diffusion coefficient of a compound in the
carrier gas is inversely proportional to pressure, the pressure reduction increases
radial diffusion resulting in faster radial equilibration and less chromatographic band
broadening. Also the use of hydragen as the carrier gas results in higher diffusion
coefficients and hence faster radial equilibration.
The last option to reduce the analysis time is by changing the velocity profile inside
the column by working under turbulent flow conditions or by using geometrically
deformed (coiled, stitched, etc.) separation tubes (option C). Under these conditions,
the velocity profile is largely flattened. The inequalities in the velocity profile are
now significantly reduced. Additionally, radial transport of the compounds is
increased by convection. In this way radial equilibrium conditions are quickly
obtained so that faster separations can be achieved.
Each of the four above mentioned possibilities to increase the analysis speed will be
described in detail in the following sections. Here it is opted to discuss the influence
of these factors on the speed of analysis under optima! column efficiency conditions
for a constant number of theoretica! plates.
Factors determining the speed of analysis in gas chromatography
2.3 PARAMETERS AFFECTING THE CHROMATOGRAPHIC
EFFICIENCY
15
As described in previous sections, there are basically four alternative routes towards
faster separation. In the subsequent paragraphs, these methods will be described. The
practical consequences will be discussed insection 2.4.
2.3.1 Influence of the column inside diameter
Already in the early sixties, Desty et al. demonstrated the advantages of the use of
narrow-bore columns (I.D. ::; 100 J.tm) to speed up the chromatographic analysis [9].
Surprisingly, since then the use of such columns has received little attention mainly
because of the lack of dedicated instrumentation. In 1982 Schutjes [10] proved
theoretically and experimentally that the reduction of the column inside diameter is
an attractive route for improving the speed of analysis in isothermal and programmed
temperature capillary GC.
Cramers et al. derived that the retention time required to obtain N theoretica! plates
under optimal separation conditions can be expressedas [11,12]:
(2.19)
where p is the average pressure of the column. In deriving this equation the
contri bution of the Cs term to plate height was neglected. From the equation, it can
be seen that the column pressure drop will have a considerable influence on the
relationship between the retention time and the column inside diameter. Here two
extreme situations can be discemed. For small pressure drops over the column
(P~ 1 ), equation (2.19) can be greatly simplified because for these conditions, f1
equals 1 and p is Pi· The retention time now becomes proportional to the square of
the column inside diameter. For large pressure drops over the column (P~oo),
f1 ~ 9/8 and p ~2p/3. For this situation, it can be shown that the retention time is
given by [11,12]:
16 Chapter 2
(2.20)
From this equation, it follows that the analysis time is proportional to the column
inside diameter.
In Figure 2.2, the relationship between the optimal average carrier gas velocity and
the column inside diameter for a given number of theoretica! plates is shown. For
these calculations, a computer program described by Leclercq and Cramers was used
[11]. The optimum carrier gas velocity is increased continuously up toa limit as the
column inside diameter is reduced. The magnitude of the increase is strongly
dependent on the number of plates. Again, two extreme situations can be
distinguished. First of all, for short columns, there is a small pressure drop over the
column and the optimum average carrier gas velocity tends to increase inversely to
the column inside diameter. For long columns, the pressure drop over the column is
higher and the optimum average carrier gas velocity becomes almost independent of
the column inside diameter.
250 ~ (.) Cl) til
] 200 '-'
~ '(3 150 0 -Cl)
> !a 100 Cl)
;§ Cl)
i 50
~ 0
0 100 200 300 400 500 600 700 800
Column inside diameter (J..lm)
Figure 2.2 Influence of the column inside diameter on the optimal average linear carrier gas velocity for different plate numbers ( operating under maximum separation efficiency conditions ). Experimental conditions: C12, Toven = 373 K, ~ = 200, k = 5.76, stationary phase: SE-30, carrier gas: He, Po= 1 bar.
Factors determining the speed of analysis in gas chromatography 17
Schutjes proved that the dependenee of the analysis time on the column inside
diameter is the same in both isothermal and linear temperature programmed analyses
[10]. lt is important, however, to indicate that the programming rate must be varied
proportional to the average carrier gas velocity and inversely proportional to the
column length, if the same retention temperatures are to be obtained on columns
having the same phase ratio and stationary phase.
2.3.2 Influence of the carrier gas
In the previous paragraph, it is demonstrated that the separation can be speeded up
by reducing the time required for radial equilibration by reducing the column inside
diameter. Another possibility to reduce the radial equilibration time is obtained by
increasing the diffusion coefficient of the analyte in the mobile phase. The diffusion
coefficient of the analyte in the mobile phase can be estimated according to the
equation ofFuller et al. [13] which reads:
(2.21)
where D A,B is the binary diffusion coefficient of the analyte in the mobile phase at
pressure p. mA and mB are the molecular weights of the carrier gas and the analyte,
respectively. Aui and Bui are the diffusion volumes that depend on the molecular
structure ofthe diffusing species. From equation (2.21) it is clear that high diffusion
coefficients can be obtained for low molecular weight carrier gases. Additionally, the
diffusion volumes of these gases arealso small [13]. The smallest diffusion volumes
are observed for the noble gases He and Ne (2.88 and 5.59 respectively). The
diffusion coefficient of the diatomic molecule of hydrogen is slightly higher (7 .07).
Although the diffusion volume of H2 is higher, the diffusion coefficient of the
analytes in the H2 carrier gas will be higher because of the lower molecular weight of
this gas.
For thin-film columns operated at negligible pressure drops, the analysis time is
proportional to 1/DM,o (see equation 2.19). For this reason, the fastest separations are
obtained with hydrogen as carrier gas. For large pressure drops (low column LD.),
18 Chapter 2
the retention time is also influenced by the viscosity of the carrier gas (see equation
2.20). In this situation, the retention time is proportional to the square root of the
carrier gas viscosity over the diffusion coefficient. The viscosity is lowest for
hydrogen. Because of the low viscosity and the high diffusion coefficient, also in this
situation, the fastest separation is obtained by using hydrogen as the carrier gas. The
ratio of the total retention times for maximum separation efficiency conditions for
different carrier gases as a function of the column inside diameter, is illustrated in
Figure 2.3. The exact shapes ofthe lines in this figure are difficult to understand as
both viscosities and diffusion coefficients affect the exact values. Numerically, the
lines can easily be calculated using the equation 2.19, the Golay·Giddings equation
and the Hagen-Poiseuille equation. From the figure it is clear, however, that the use
ofhydrogen as carrier gas results always in the fastest separations.
In going from a thin-film to a thick-film column the maximum efficiency of the
column decreases substantially. The optimum linear carrier gas velocity is reduced
and the slopes of the ascending branch of the Golay-Giddings curves at high carrier
gas veloeities are much steeper, depending primarily on the diffusion time of the
4 tR,N/ tR,H2
"" <U e 3 '.;:j
"" ~ § 0
·~ 2
1 L_~~~~~-r~~~~r-~~~~~-r~ 0 100 200 300 400 500 600 700 800
Column inside diameter (Jlm)
Figure 2.3 Comparison of the ratio of total retention times for several carrier gases ( operating under maximum separation efficiency conditions) as a function of the column inside diameter.
Experimental conditions: C12, Toven = 373 K, ~ = 200, k = 5.76, stationary phase: SE-30, Po
= 1 bar, N 100000.
Factors determining the speed of analysis in gas chromatography 19
analyte in the stationary phase. In contrast to thin-film columns, where the minimum
theoretica! plate height is almost comparable when either hydrogen or nitrogen is
used, for thick-film columns the minimum plate height is considerably smaller for
nitrogen. However, the corresponding optimum linear velocity of the carrier gas is a
few times higher for hydrogen resulting in a lower Hl u ratio and faster separations.
2.3.3 Influence of the outlet pressure
An alternative metbod to increase the diffusion coefficient in the mobile phase is
obtained by reducing the column outlet pressure. As can be seen from equation
(2.21) the local diffusion coefficient in the mobile phase is inversely proportional to
the local pressure. By decreasing the column inlet pressure, the average pressure in
the column is reduced. As a consequence, the average diffusion coefficient is
increased with the effect of a faster radial equilibration enabling faster analyses.
As demonstrated by Giddings in the earlier 60s, operation of a capillary column at
reduced pressures will favour the speed of analysis [14]. Cramers et al. described and
evaluated the possibilities of vacuum outlet conditions in detail [11,12,15-18]. By
operating the column under vacuum outlet conditions, the maximum attainable
number of theoretica! plates (per length unit) is reduced, theoretically by 8/9 through
the f1 factor. The column has to be lengthened to compensate for this loss.
In Figure 2.4 the gain in separation speed is illustrated for thin-films columns(~=
200) as a function of the column inside diameter at different plate numbers. Also
here, for the theoretica! calculations, the program of Leclercq and Cramers was used
[11]. From this tigure it is obvious that the gain is most pt:onounced for wide-bore
andlor short columns. For such columns, the average pressure in the column is
significantly reduced by operating under reduced outlet pressure conditions. The
gain in separation speed in particular increases strongly at sub-atmospheric optimum
inlet pressures. For narrow-bore columns andlor long columns, the gain by operating
under vacuum outlet conditions is limited because the inlet pressure required for the
maximum separation efficiency is relatively high. As a consequence, the average
pressure is only moderately affected by reducing the outlet pressure from
atmospheric to vacuum. Hence, the analysis times for these columns remain almost
unaffected. For standard GC columns (e.g. 320 ~-tm I.D., N = 105), only small
20 Chopter 2
tor-----------------------------------~
N= 103
8
2
0 0 100 200 300 400 500 600 700 800
Column inside diameter (Jlm)
Figure 2.4 Ratio of the total relention times under atmospheric conditions over that under vacuum outlet conditions as a function of the column inside diameter for different plate numbers (operating under maximum separation efficiency conditions). Experimental conditions: C12,
Toven = 373 K, ~ 200, k = 5.76, stationary phase: SE-30, carrier gas: He.
improvements are observed upon going from atmospheric outlet to reduced outlet
pressures.
Because for short and/or wide-bore columns the contri bution of the CM,0-term to the
plate height is reduced drastically when operating the column under reduced outlet
pressures, the contri bution of the Cs-term becomes more pronounced. The effect of
the film thickness and the column inside diameter on the gain in separation speed is
illustrated in Figure 2.5. From this figure it can again be seen that the gain is
negligible for narrow-bore columns. Moreover, it is evident from the tigure that the
gain for wide-bore columns is the greatest at a low film thickness (high B).
Factors determining the speed of analysis in gas chromatography
5
4
Figure 2.5
100 200 300 400 500 600 700 800
Column inside diameter (J.lm)
21
Influence of the film thickness on the ratio of the total retention times under atmospheric conditions over that under vacuum outlet conditions as a function of the column inside diameter (operating under maximum separation efficiency conditions). Experimental conditions: C12, Toven 373 K, k = 5.76, stationary phase: SE-30, carrier gas: He, N = 10000.
2.3.4 Influence of the flow profile
2.3.4.1 Turbulent flow conditions
An important contribution to chromatographic band broadening is the parabolic
velocity profile in the column. Solute molecules in the centre of the column have a
higher effective velocity than the molecules near the column wall. Considerable band
broadening is observed because the radial diffusion of the solutes in the mobile
phase is relatively slow. The time required for radial mass transport between fluid
layers with different veloeities is appreciable compared to the residence time in the
column. By changing the flow profile of the mobile phase, enhanced radial
equilibration of the solutes can be obtained.
One way to increase radial transport is obtained by creating turbulent flow [19-24].
For turbulent flow conditions, the velocity profile is largely flattened, thereby
decreasing flow inequalities. Moreover, the effective diffusion coefficient is
considerably increased by convective contributions. As a consequence, peak:
broadening arising in the mobile phase as a result of the velocity profile is expected
22 Chapter 2
to be reduced. The possibilities and limitations of turbulent flow conditions as a
means to increase separation speed in capillary columns were studied in detail by
Van Es et al. [24]. The conclusions of this work can be summarised as follows:
transition of laminar to turbulent flow occurs as foreseen at Re = 2300. Beyond the
critica! Re value the plate height is reduced drastically. In Figure 2.6, a plot of the
reduced plate height which is defined as the theoretica! plate height over the column
inside diameter, as a function ofthe Reynolds number is shown for k = 0 and k = 1.
It was found that this curve was in excellent agreement with earlier theoretica! and
experimental results for unretained compounds presented by Flint et al. back in 1969
[20]. The large influence of the retention factor on the plate height is an intrinsic
property of turbulent flow. It is caused by the presence of a laminar sublayer close to
the column wall which arises from the shape of the velocity profile. The total
chromatographic band broadening is depending on the turbulent dispersion in the
centred turbulent core and the radial molecular dispersion in the small laminar
sublayer at the wall. At higher Re-values, the band broadening caused by the laminar
sublayer becomes more pronounced for the more retained compounds because
transfer of these compounds through this layer to the stationary phase has to occur
more frequently. Hence the contribution to the total band broadening caused by
molecular diffusion through this layer will increase for highly retained compounds.
For this reason the gain in separation speed in real separations is limited.
Additionally, the inlet pressure required to obtain turbulent flow conditions is
extremely high. For example, for a column with a length of 5 m and an I.D. of 320
J.lm, the inlet pressure required to obtain a Re-value of 104 with N2 as carrier gas is
approximately 36 bar. This pressure is even higher if He or H2 are used as carrier
gases. Therefore, for practical chromatography, turbulence is not a useful
phenomenon. There are better routes to obtain the same or even higher separation
speeds.
Factors determining the speed of analysis in gas,chromatography
Figure 2.6
2.0 .-----------------,
1.8 1.6 1.4 1.2 1.0
..= 0.8 ~ 0.6
0.4
k=I
0.2 O.O~--"!;---T----4-------èti,.-----1
-0.2 -0.4 k=O -0.6 0
2 3 4 5 6
log Re
23
Comparison of the theoretica! and experimental results for the reduced plate height h as a function of Re with k 0 and k = 1 [24].
2.3.4.2 Coiling indoeed secondary flow
A second way to flatten the velocity profile and enhance radial diffusion, and hence
increase the separation speed in open tubular columns, is to introduce a so-called
secondary flow by tightly coiling the separation column. In these coiled tubes
centrifugal forces produce a secondary flow in radial direction [22,25]. This flow
significantly enhances radical equilibration. At low veloeities of the mobile phase,
these forces are weak. Now the secondary flow manifests itself in the formation of
two radial circular pattems which tend to divide the cross-section of the tube into
two equal parts. At higher veloeities the centrifugal forces increase sharply which
results in a more linear axial velocity profile. At very high veloeities the axial
velocity profile tends to plug flow conditions while turbulence sets in.
Detailed theoretica! studies and experimental descriptions of axial dispersion in
helically coiled columns for both GC and LC have been publisbed by Tijssen [22,25-
28]. Hofinann and Halàsz [29] evaluated peak dispersion in squeezed, twisted and
waved tubes. The positive effects of coiling in open-tubular SFC have been
demonstrated by Novotny et al. [30] and Janssen et al. [31 ]. Coiled columns have
24 Chapter 2
also been shown to be advantageous for post-column reactions in LC [32,33], for
flow injection analysis [34,35] and for on-line sample enrichment for GC [36].
In straight tubes radial dispersion is detennined by molecular diffusion solely. In
defonned tubes radial dispersion is the combined effect of molecular diffusion and
convection by the secondary flow. For this reason (and assuming that the gas
behaves ideal), the radial dispersion coefficient, DM,o in the CM,o tenn of the plate
height equation should be replaced by DR.o which is the sum of DM,o and Dsp, the
latter parameter being the secondary-flow dispersion coefficient. The value of the
secondary flow dispersion coefficient depends on the velocity parameter De2Sc,
where De (Dean number) and Sc (Schmidt number) are defined as:
De= Re..fÀ (2.22)
Sc= (2.23) PoDM,o
with
Re= Pollode (2.24) 11
À.=~ (2.25) dcoil
where Po is the density of the carrier gas under outlet conditions, Re the Reynolds
number and À. the aspect ratio which is the ratio of the column radius (de) and the coil
radius ( dooi1). Tijssen gave qualitative and quantitative descriptions of the radial
dispersion coefficients in coiled columns as a function of the De2Sc number. The
results are listed in Table 2.1 [25]. As can beseen from the equations in Table 2.1,
the radial dispersion coefficient can be increased by a factor of 6.4 by working at
De2Sc values above 2300. Unfortunately however, it can be calculated that for GC
the practical consequences of coiling are limited. To obtain a significant gain in
radial dispersion due to secondary flow effects, the column has to be operated at very
high linear velocities. Not only does this require very high pressure drops, it also
Factors determining the speed of analysis in gas chromatography
Table 2.1 Theoretica} relations for radial dispersion in helically coiled open tubular columns.
Conditions
DéSc<lO,ÎI. 0
De2Sc <2300
equation
DR,o = DM,o
DR,or = I+ De2Sc
DM,o 270
DR,o = 6.39 DM,o
25
implies that the absolute plate height is still very high. For this reason the practical
use of geometrically deformed tubes in gas chromatography is very limited.
2.3.5 Packed columns versus capillary columns
In the previous sections, various methods for increasing the speed of analysis in gas
chromatography have been described. So far, the discussion was confined to open
tubular columns. Fast separations can also be obtained with packed columns. For
packed columns, the van Deemter-Giddings equation reads [37]:
(2.26)
where dp is the partiele diameter of the packing material, À' is a geometry factor
indicating how uniformly the column is packed and y is a correction factor which
brings the tortuosity of the gas channels in the column into account. The first term in
this equation is known as the Eddy diffusion contribution which represents the band
broadening caused by differing path lengths for the carrier gas along the particles in
the packed bed. The van Deernter equation suggests that the use of small-size
supporting particles can result in very low plate heights. Especially if narrow mesh
ranges are used. In practice, however, the use of small partiele sizes is limited
because unpractically high inlet pressures (more than 50 bar) are required in order to
reach the optimal flow conditions of such columns.
26 Chapter 2
Jonker et al. [38] showed a separation of a 4 component mixture within 0.15 seconds
using a 32 mm long column packed with 10 J.tm partiel es. In this separation 650
plates were obtained (k = 2) at an inlet pressure of 64 bar. From theoretica!
calculations, it can be proven that the same separation efficiency can be obtained in
0.15 sec using a 120 J.tm LD. open tubular of 12 cm. When vacuum outlet conditions
are applied, the same result can also be obtained with a 300 J.tm I.D. column of 24
cm. This example clearly illustrates the great practical advantage of capillary
columns over packed columns. Only for some applications and from the point of
view of sample capacity, packed columns are superior.
2.3.6 Conclusions
As convincingly demonstrated above, open tubular columns provide the highest
separation efficiencies. Generally speaking, thin-film columns have to be favoured
since for thick-film columns the plate height is increased and the corresponding
optimal carrier gas velocity is decreased considerably. In open tubular gas
chromatography, there are various routes towards faster analyses. The most efficient
way to reduce the analysis time is to use narrow-bore open tubular thin-film columns
with hydrogen as the carrier gas. Forthese columns, the gain in separation speed by
operating the column under vacuum outlet conditions is negligible. Vacuum outlet
operation only results in a significant reduction of the analysis time for short/wide
bore columns. The use of coiled columns, turbulent flow conditions or small-size
packed columns offers only limited practical advantages. Moreover, each of these
options requires extremely high inlet pressures.
2.4 PRACTICAL CONSEQUENCES OF THE USE OF NARROW-BORE
CAPILLARY COLUMNS
In the previous sections it has been demonstrated that the best method to increase the
speed of analysis for a given separation efficiency is decreasing the column inside
diameter. The increased number of plates per time unit results in very narrow peaks.
Apart from efficiency and analysis time, also the minimum amount that can be
Factors determining the speed of analysis in gas chromatography 27
detected (Q0) is favoured sirree narrow peaks result in a better signal-to-noise ratio.
Unfortunately, however, the sample volume that can be injected onto narrow-bore
columns is very low and the injection band width that can be tolerated is very small.
As a consequence, narrow-bore capillary GC poses high demands on the injection
technique, the detector and the data acquisition system (sampling rate ). The injection
system has to be able to produce narrow input plugs. The detection system has to be
sufficiently fast to reconstruct the chromatographic resolution. Additionally, because
of the low sample capacity of a narrow-bore column, the detection devices have to be
very sensitive in order to obtain an acceptable working range. All these
consequences of the use of narrow-bore capillaries are discussed in detail in the
following subsections.
2.4.1 Defection limits
Detectors for capillary chromatography must have a high sensitivity. The term
sensitivity refers to the change in the detector response per unit concentration or unit
mass flow of a substance in the mobile phase entering the detector. Some detectors
respond to concentration whereas others respond to mass flow. Hence, two types of
detectors have to be considered: mass flow sensitive detectors and concentration
sensitive detectors. The detection limit for a given compound is not only a function
of the detector sensitivity but also of the detector noise RN. The noise level is the
amplitude of the envelope around the baseline which includes all random variations
ofthe detector signal. The detection limits fora compound are strongly dependent on
the column and detector properties. The minimum detectable amount (Q0) is the
smallest quantity of a solute that can be detected (signal-to-noise ratio = 2). For a
mass flow sensitive detector, Q0, can be calculated according to:
m r;;;-- 2RN Qo = '\/21t--<>tot sm (2.27)
where sm is the sensitivity of a mass flow sensitive detector and <>tot the total standard
deviation of a chromatographic peak. For a concentration sensitive detector, the
minimum detectable amount is given by:
28 Chopter 2
(2.28)
where se is the sensitivity of a concentration sensitive detector and F det is the
volumetrie flow rate in the detector at detector pressure and temperature. O"tot is the
total standard deviation ofthe chromatographic peak (see section 2.4.3).
Noij demonstrated the large influence of the column inside diameter for thin-film
columns under normalised chromatographic conditions [39]. He found that,
depending upon the column in- and outlet pressure ratio, a second to third power
dependenee exists for concentration sensitive detectors, whereas for mass flow
sensitive detectors the minimum detectable amount is proportional to de up to d~.
The minimum detectable.concentration ofthe sample, C0, is defined as:
(2.29)
where Q0 is the minimum detectable amount for a mass flow sensitive or a
concentration sensitive detector and Vinj is the sample volume introduced onto the
column. Although the minimum detectable amount is favoured by the reduction of
the column inside diameter, this does not necessarily imply that very small sample
concentrations can be detected. This because in narrow-bore GC only very small
sample volumes can be injected. During the last decade, considerable effort was
devoted to the development of injection devices able to produce narrow input bands,
compatible with the small chromatographic band broadening of narrow-bore
columns (see chapter 3). Unfortunately, however, with these introduetion systems,
only very small sample quantities are introduced onto the column and most of the
sample is splitted. As a consequence, the minimum detectable concentration is high.
In order to overcome this problem and to obtain full profit of the true potentials
offered by narrow-bore columns, non-splitting injection techniques have to be used.
The possibilities and limitations of a number of newly developed non-splitting
injection techniques will be discussed in detail in chapter 3.
Factors determining the speed of analysis in gas chromatography 29
2.4.2 Sample capacity
F or reasons of detectability it is often desirabie to inject as large a quantity of sample
as possible on a chromatographic column. However, high mass or volume loads may
cause an undesirable decrease in column efficiency. lt is therefore useful to have a
theoretica! understanding of the re lation between sample load and efficiency and to
have practical guidelines for the maximum sample amount that can be injected.
Many theoretica! models have been reported. However, only two studies publisbed
by Jaulmes et al. [40] and Ghijsen et al. [41], respectively, have proven to be in good
agreement with practice. Although the models of Jaulmes and Ghijsen have been
derived following completely different approaches, the conclusions derived from the
equations are essentially the same. Below the conclusions from these models will be
briefly discussed.
The Jaulmes and Ghijsen theories both state that the sample capacity Qs is
proportional to the square of the column inside diameter. Both theories predict that
the sample capacity is directly proportional to the film thickness. Hence, the sample
capacity is even proportional to d~ if the phase ratio is kept constant. The sample
capacity is thus drastically reduced when using narrow-bore columns. As long as the
contribution of the C8-term is relatively small, it might therefore be interesting to
increase the film thickness to increase the sample loadability. However, when the
influence of the C8-term becomes apparent, the sample capacity will show an
additional increase yielding a considerable increase of the plate height. In this
situation, the sample capacity is increased at the expense of column efficiency and
separation time.
Another important factor is the interaction between the solute and the stationary
phase. In general, if the solute and the stationary phase have a similar structure, the
solubility will be greater and thus the sample capacity will be larger.
Finally, also the retention factor has a significant effect on the sample capacity. Only
for very small retention factors (k<l), high sample quantities can be injected before
overloading starts to occur. The sample capacity is drastically reduced for higher k
values. For largervalues ofthe retention factor (k>3), the sample capacity is almost
unaffected by a further increase of k.
30 Chapter 2
The limited sample capacity of narrow-bore columns in itself should not be regarded
as a major drawback of these column. Problems, however, may arise when a
decreased sample capacity results in a significantly reduced woricing range, W, here
defmedas:
(2.30)
where Qs is the sample amount of a solute that can be injected on a column giving a
limited (e.g. 10%) increase in peak width and Q0 is the minimum detectable amount
of substance that can be reliably detected. It will be clear that it is important to have
a very sensitive detector available in ordertopreserve an acceptable woricing range.
2.4.3 Instromental band broadening
The fundamental assumption of the Golay-Giddings model is that band broadening
occurs only by the chromatographic process. The peak profile recorded during a
separation should depend only on the operating characteristics of the column and
should be independent of the instrumentation. Under experimental conditions,
however, additional broadening can arise from dispersion and mixing phenomena in
the injector, connecting tubes and the detector cell, as well as from electronic
components which govem the response speed of the detector and the data system.
The various contributions to band broadening can be treated as independent factors
additive intheir variances (cr2) [42-44]. For this reason the total peak width can be
calculated according to:
2 2 2 2 2 2 O'tot = O'chrom + O'inj + O'det + O'con + O'eJ (2.31)
where O'tot is the total standard deviation of the chromatographic peak, O'chrom the
standard deviation due to chromatographic dispersion, O'inj and O'det the standard
deviation due to the volume, geometry and flow stream through the injector,
respectively the detector, O'con the standard deviation due to connecting tubes, unions,
frits, etc., and O'e1 the standard deviation caused by the finite response time of the
Factors detennining the speed of analysis in gas chromatography 31
electrooie circuits of the detector and data acquisition system. For an evaluation of
the total peak broadening, equation (2.36) can be simplified to:
2 2 2 cr tot = cr chrom + cr extra (2.32)
where crextra is the total standard deviation due to instromental contributions.
Under high-speed conditions, the chromatographic band broadening is reduced
drastically so that the contribution of the instromental band broadening becomes
more pronounced. Gaspar et al. [44] showed that the extra column band broadening
can be brought into account by adding a factor to the Golay-Giddings equation. This
factor reads:
2 H D-2 O'extra -2 = u = u
extra (1 + k)L (2.33)
where Rextra is the increased plate height due to instrumental band broadening and D
the term descrihing extra band broadening effects. It will be clear that the efficiency
loss caused by the instrumentation becomes smaller for longer columns and when the
solute is more retained. For high-speed narrow-bore capillaries (short columns), the
contribution of the extra column band broadening becomes more important. To
obtain the true potentials offered by these columns, the band broadening caused by
instrumentation has to be extremely small. The lack of compatible instromentation,
so far, seriously hinderedtheuse of narrow-bore columns(< 100 Jlm I.D.) in daily
laboratory operation. During the last decade, considerable effort was devoted to the
development of injection and detection devices compatible with the small
chromatographic band broadening of narrow-bore columns. The injection system has
to be able to generate narrow input bands. The detection system should have a low
detection volume in order to avoid additional band broadening in the detector.
Additionally, the detector and its electronics should be sufficiently fast to reconstruct
the chromatographic separation and to preserve the chromatographic integrity. When
the detector electronics are too slow, tailing and distorted peaks will be observed. An
overview of the injection and detection devices described in literature for use in
combination with narrow-bore capillaries, will be given in the chapters 3 (injection
systems) and 4 ( detection systems ).
32 Chapter 2
2.5 CONCLUSIONS
In Table 2.2, the possibilities of alternative routes toward faster separations and the
practical consequences for detection limits, inlet pressures, etc. are summarised.
From a theoretica! point of view, there are various ways to perform high-speed
separations. The most efficient and easiest way to increase the separation speed is to
reduce the column inside diameter and to use hydrogen as carrier gas. The
applicability of vacuum outlet conditions as a means to increase the separation speed
is limited to short wide-bore columns. The use of turbulent flow conditions or
geometrie deformation of the column as a means to enhance radial diffusion, and
hence enable faster separations, is unpractical since very high inlet pressures are
required and the gain in analysis speed is limited, especially for the more retained
analytes.
Since the analysis time is proportional to the chromatographic band broadening,
under normalised separation conditions (N constant), a reduction ofthe analysis time
results in a reduced peak width and hence smaller sample amounts can be detected.
Unfortunately, the instromental requirements become more stringent the smaller the
column inside diameter. The contribution of the instromental band broadening
becomes more pronounced the higher the separation speed. First of all, the injection
device has to be able to produce narrow input plugs in order to be compatible with
the small chromatographic peak width and to preserve the high column efficiency
provided by the narrow-bore column. An extra complication is the high inlet pressure
required to obtain the optimal column flow for these narrow-bore columns. The inlet
pressures for these narrow-bore columns are typically between 5 and 20 bar.
Additionally, the sample capacity is drastically reduced. For this reason, the detector
should be sufficiently sensitive topreserve an acceptable working range. Moreover,
the detection system, the detector electronics and the data acquisition system have to
be sufficiently fast to reliably reconstroct the chromatographic separation.
Table 2.2 ~ Influence ofvarious routes towards faster separations ~
~ ~
reduced de vacuum outlet hydrogen carrier turbulence/coiling ~
~-~-
separation speed ++ + for short andlor moderate for He -+ H2 + s. wide-bore columns + from other gases to H2
(!;,
~ almost no influence for (!;,
narrow-bore and/or 2. long columns <S1.
~ inlet pressure (p;) moderate almost no influence for low veryhigh ~
typically 5 to 20 narrow and/or long columns t;·
s· bar sub-atmospheric for short i and wide·bore columns
k dependenee moderate moderate moderate strong
g. ~ ~
minimum detectable amount (Qo) ++ + + ++ ~ .ij
sample capacity (Qs) ~d3 0 0 0 ~
c
minimum detectable + + ++ concentration (Co) acceptable values
still possible with non-splitting injection techniques
Instrumental requirements difficult easy/moderate easy/moderate almost impossible
Legend:++/·- : strong effect,+/- : important effect, 0: almost no effect w w
34 Chapter 2
2.6 REFERENCES
1. C.F. Poo1e, S.K. Poole, in "Chromatography Today", Elsevier, Amsterdam, 1991, p. 32.
2. L.S. Ettre, in "Open Tubular Columns: an Introduction", Perkin Elmer, Norwalk, 1973, p. 13.
3. J.H. Pumell, C.P. Quinn, in "Gas Chromatography", R.P.W. Scott (Ed.), Butterworths, London, 1960, p. 184.
4. J.C. Giddings, Anal.Chem., 32 (1960) 1707. 5. J.J. van Deemter, F.J. Zuiderweg, A. Klinkenberg, Chem. Eng. Sci., 5 (1956) 271. 6. M.J.E. Golay, in "Gas Chromatography", D.H. Desty (Ed.), Butterworths, London,
1958, p. 36. 7. J.C. Giddings, S.L. Seager, L.R. Stucky, G.H. Stewart, Anal. Chem., 32 (1960) 867. 8. A.T. James, A.J.P. Martin., Biochem. J., 50 (1952) 679. 9. D.H. Desty, A. Goldup, W.T. Swanton, in "Gas Chromatography", N. Brenner, J.E.
Callen, M.D. Weiss (Eds.), Academie Press, New York, 1962 (1961 Lansing Symposium), p. 105.
10. C.P.M. Schutjes, E.A. Vermeer, J.A. Rijks, C.A. Cramers, J. Chromatogr., 253 (1982) 1.
11. P.A. Leclercq, C.A. Cramers, J. High Resolut. Chromatogr. Chromatogr. Comm., 8 (1985) 764.
12. C.A. Cramers, P.A. Leclercq, CRC Crit. Rev. Anal. Chem., 20 (1988) 117. 13. E.N. Fuller, P.D. Schettler, J.C. Giddings, Ind. Eng. Chem., 58 (1966) 19. 14. J.C. Giddings, Anal. Chem., 34 (1962) 314. 15. C.A. Cramers, G.J. Scherpenzeel, P.A. Leclercq, J. Chromatogr., 203 (1981) 203. 16. P.A. Leclercq, G.J. Scherpenzeel, E.A. Vermeer, C.A. Cramers, J. Chromatogr., 241
(1982) 61. 17. P.A. Leclercq, C.A. Cramers, J. High Resolut. Chromatogr. Chromatogr. Comm., 10
(1987) 269. 18. P.A. Leclercq, C.A. Cramers, J. High Resolut. Chromatogr. Chromatogr. Comm., 8
(1988) 845. 19. J. Giddings, W. Man, M. Myers, Science, 154 (1966) 146. 20. L. Flint, P. Eisenklam, Can. J. Chem. Eng., 47 (1969) 101. 21. F. Doue, G. Guiochon, Sep. Sci. Technol., 5 (1970) 197. 22. R. Tijssen, thesis, University ofTechnology Delft, the Netherlands, 1982. 23. M. Martin, G. Guiochon, Anal. Chem., 54 (1982) 1533. 24. A van Es, J. Rijks, C.A. Cramers, J. Chromatogr., 477 (1989) 39. 25. R. Tijssen, Sep. Sci. Technol., l3 (1978) 681. 26. R. Tijssen, Chromatographia, 3 (1970) 525. 27. R. Tijssen, R.T.Wittebrood, Chromatographia, 5 (1972), 286.
Factors determining the speed of analysis in gas chromatography 35
28. R. Tijssen, N. van den Hoed, M.E. van Kreveld, Anal. Chem, 59 (1987) 1007. 29. K. Hoffman, I. Halàsz, J. Chromatogr., 199 (1980) 3. 30. S.R. Springston, M. Novotny, Anal. Chem., 58 (1986), 2699. 31. H.-G. Janssen, J.A. Rijks, C.A. Cramers, J. High Resolut. Chromatogr., 13 (1990)
475. 32. B. Lillig, H. Engelhardt, in "Reaction Detection in Liquid Chromatography", l.S.
Krull (Ed.), Marcel Dekker, New York, 1986, Chapter 1. 33. R.S. Deelder, M.G.F. Kroll, A.J.B. Beeren, J.H.M. van den Berg, J. Chromatogr.,
149 (1978) 669. 34. R. Tijssen, Anal. Chim. Acta, 144 (1980) 71. 35. J.H.M. van den Berg, R.S. Deelder, H.G.M. Egberink, Anal. Chim. Acta, 144
(1980) 91. 36. H.G.J. Mol, H.-G. Janssen, C.A. Cramers, U.A.Th.Brinkman, J. MicrocoL Sep., 7
(1995) 247. 37. C.F. Poole, S.K. Poole, in "Chromatography Today'', Elsevier, Amsterdam, 1991, p.
23. 38. R.J. Jonker, H. Poppe, J.F.K. Huber, Anal. Chem., 54 (1982) 2447. 39. T. Noij, J. Curvers, C.A. Cramers, J. High Resolut. Chromatogr. Chromatogr.
Comm., 9 (1986) 752. 40. A. Jaulmes, I. Ignatiadis, P. Cardot, C. Vidal-Madjar, J. Chromatogr., 395 (1987)
291. 41. R.T. Ghijsen, H. Poppe, J.C. Kraak, P.P.E. Duysters, Chromatographia, 27 (1989)
60. 42. J.C. Stemberg, Adv. Chromatogr., 2 (1966) 205. 43 J.C. Giddings, Separ. Sci., 4 (1969) 181. 44. G. Gaspar, R. Annino, C. Vidal-Madjar, G. Guiochon, Anal. Chem., 50 (1978)
1512.
36
Chapter 3
Non-splitting injection techniques for
narrow-bore capillary gas
chromatography
SUMMARY
In this chapter, the use of hot splitless, cold splitless and on-column injections for
trace analysis in narrow-bore capillary GC is evaluated. Despite the low column
flow rates for the columns used (0.5 to 1 mi/min), the required splitless times for
splitless injections can be surprisingly short if Ziners are used with a smal! inside
diameter. The effect of the experimental conditions as sample volume, injection and
oven temperatures on the focusing effects, dis crimination and degradation behaviour
of the analytes are discussed. The possibilities to obtain sensitive and Jast
separations are illustrated by various applications.
38 Chapter3
3.1 INTRODUCTION
Narrow-bore columns have great potentials for improving the chromatographic
performance in terms of column efficiency, resolving power and speed of separation.
However, to realise the high potential offered by these columns, the extra column
band broadening caused by instrumentation has to be minimised accordingly.
Sample introduetion is a very critical step in capillary GC. Stringent requirements are
put on the introduetion devices used. The composition of the sample injected onto
the column should be the same as in the original sample, i.e. discrimination,
adsorption, rearrangements or other reactions should he avoided. To obtain
quantitative results, the amount of sample injected should be highly reproducible and
the sampling procedure should not degrade the column efficiency. For narrow-bore
columns, the increased inlet pressure required is an extra complication for sample
introduction.
In this chapter, first a short overview of injection devices described in literature
capable of producing very narrow injection plugs is presented. Although narrow
input bands can be obtained with these injection devices, they all suffer from one
common disadvantage i.e. only minute sample volumes are actually injected onto the
column. Despite the attractive minimum detectable amounts (MDA) that can be
obtained with narrow-bore columns, the minimum detectable concentration (MDC)
is far too low for many practical applications. For this reason, these methods are not
suited for the analysis of sample concentrations less than 100 ppm. To improve the
MDC, larger sample volumes have to be injected.
In literature, so far only a limited number of different approaches for the introduetion
of larger sample volumes onto narrow-bore columns has been descri bed. Ho wever, a
detailed discussion of focusing efficiencies, band widths, recoveries, repeatabilities
and detection limits for hot splitless, cold splitless and on-column injection
techniques has notbeen presented yet. In this study, the possibilities and limitations
of these injection techniques in combination with narrow-bore columns are
evaluated. The experimental consequences of using narrow-bore columns in
combination with non-splitting techniques will be discussed in detail and illustrated
by various applications.
Non-spUtting injection techniques for narrow-bore capillary GC 39
3.2 OVERVIEW OF INJECTION DEVICES ENABLING HIGH-SPEED
SEP ARA TI ONS
Desty and Goldup [1,2] reported on the first fast injection which was obtained by
hitting a syringe with a hammer. In this way it was possible to produce input band
widths with a duration of a few tens of milliseconds. Syringe consumption was not
reported.
For rapid injection of gaseous samples, a conventional split injector operatingat very
high flow rates can be used. Typical gas flow rates are in the range of a few litres per
minute! In this way, narrow input bandscan be obtained. The limiting contribution to
band broadening of this type of injector, crinj• can be estimated by {assuming an
exponentially shaped injection plug):
{3.1)
where Vinj is the volume of the gas sample occupied by the sample under injector
conditions and Finj is the total flow rate in the injector, assuming that the introduetion
of the sample in the injector occurs infinitely fast. The above equation is only valid
for gaseous samples. For liquid samples, the evaporation rate ofthe sample in the hot
injector is slow and refocusing techniques are often required. More detailed
information on the injection of samples in high-speed GC is given below.
Gaspar and co-workers [3-6] described a fast injection device for gaseous injections
basedon a fluid logic gate. With this injection device, originally suggested by Wade
and Cram [7], the injection of sample vapour into the column is controlled by a gas
flow switching system. A fluid stream can be switched very fast between two paths.
In this way, injection bands with a duration in the order of a few milliseconds can be
obtained.
A revolutionary new approach for sample introduetion resulted from the
miniaturisation of a gas chromatograph using an etching technique called silicon
micromachining. This technology permits the integration of carrier gas, sample
channels, valve seats and a thermal conductivity detector on a silicon wafer [8-12].
The capillary GC column is directly mounted onto this wafer. Due to the negligibly
40 Chapter 3
small volumes of the interfacing, this micro-GC is an i deal instrument for high-speed
narrow-bore GC. At the start of the injection cycle, fresh gas sample is pumped by an
extemal vacuum pump into the etched silicon channel which forms the sample loop.
Next, the inlet valve is closed and the sample gas is compressed. When the pressure
reaches a pre-set value slightly higher than the inlet pressure of the carrier gas, the
injection valve is opened forabout 10 msec, injecting the sample (splitless) into the
carrier gas stream. The input band of such a valve can he considered rectangular. It
can he calculated that for this situation the standard deviation caused hy injection is
about 2.9 msec [11 ].
So far, the injection devices described only allow the injection of gaseous samples.
More detailed information on the injection of liquid samples in high-speed GC is
given helow.
For the injection of liquids directly onto the on-column, high pressure injection
valves as applied for supercritical-fluid chromatography [13] and micro liquid
chromatography [14] can he used [12,15,16]. A schematic representation of such an
injection device is shown in Figure 3 .1. The injection valve consists of a high-speed
switching valve with an internat sample loop. The sample volume varies from 40 to
200 nl. By rapid switching ofthe rotor, a small fraction ofthe contents ofthe sample
sample ...
Figure 3.1.
waste carrier
internat rotor
fused silica tubing
Schematic diagram of a switching valve, including the flow splitter [16].
-- t separation
column
Non-splitting injection techniques jor narrow-bore capillary GC 41
loop is introduced onto the column as a narrow band. In order to avoid problems
with dead volumes, a split flow can be applied. In this way it is possible to generate
the required narrow input bands with a width in the range of milliseconds. Such
bands are compatible with the small chromatographic peak broadening of narrow
bore capillaries.
An alternative route to obtain narrow input bands is to use cold trapping on the GC
column followed by a fast thermal desorption [15,17-21]. The sample is reinjeeled
from the cold trap by a fast thermal desorption step made by means of high-power
electrical heating. An example of an extremely fast separation is shown in Figure 3.2.
The complete separation of the nine components present in the sample was finished
within 660 msec after reinjection. For example for cyclohexane, about 24000 plates
per second have been generated [15].
With all these injection techniques enabling small input band widths, only minute
sample amounts can be introduced onto the column. As a consequence, the minimum
detectable concentrations are far too low for many daily laboratory analyses.
Problems arise with the sampling of mixtures that are highly diluted. To improve the
0 100 200 300 400 500 600 700
Time (msec)
Figure 3.2
High-speed chromatogram. GC column: OV-1, L = 0.3 m, I.D. = 50 fJ.ffi. Experimental conditions: Toven = 72°C, Pi= 4.5 bar, trap temperature = -75°C, heating ofthe cold trap: 11 V, 50 ms [15].
42 Chapter 3
minimum detectable concentration, larger sample volumes have to be injected. In the
next paragraphs, the basic principles of non-splitting injection techniques will be
described. First of all, the basic principles, advantages and disadvantages of these
injection techniques will be briefly discussed. In the subsequent sections, the
applicability of these non-splitting injection techniques in combination with narrow
bore columns is described in detail.
3.3 CONCEPT OF NON-SPLITTING INJECTION TECHNIQUES
In the splitless mode, the sample is introduced by syringe injection into the same or a
similar device as used for split sampling. In the hot splitless injection mode, the
sample is injected into a hot injector with the split exit closed. During the "splitless
period" ( ciosure time of the split exit after the injection) alrnost the entire vaporised
sample volume is transferred to the column by the carrier gas flow. The duration of
the splitless period is comrnonly between 40 and 80 seconds. After this time, the split
is re-opened to purge the remaining sample material out of the injector. The
successful performance of splitless injections is due to the reconcentration effects
which sharpen the solute bands before the actual separation process starts.
Refocusing of the most volatile compounds by the so-called "solvent effect" can be
obtained if the column temperature during the sample introduetion is far enough
below the (pressure corrected) boiling point of the solvent so that the latter
recondenses in the column inlet. High boiling compounds are focused by cold
trapping. The most important disadvantage of this injection technique is
discrimination of high boiling compounds due to selective evaporation from the
syringe needie and thermal degradation of thermally unstable solutes.
To overcome discrimination, the sample can be injected into a cold injector. Only
after the liquid sample has been injected, the injector is rapidly heated to achieve
vaporisation and subsequent transfer into the column under splitless conditions. Here
the splitless time is the time the split remains closed after the injector temperature
has reached its final temperature. Also here, to obtain good focusing effects, the
initial oven temperature should be below the boiling point of the solvent. The oven
Non-splitting injection techniques for narrow-bore capillary GC 43
temperature is increased after the sample is transferred to the column and the split
exit is opened.
Another injection technique that will be discussed in this chapter is the on-column
injection technique. This technique offers increased utility for thermally labile and
higher boiling solutes because the sample is deposited directly in the unheated
column. With increasing column temperatures, solute vapour pressure increases and
the chromatographic process commences. In this way overheating of the sample, as
can occur during extemal vaporisation, is avoided.
3.4 OVERVIEW OF NON-SPLITTING INJECTION TECHNIQUES FOR
NARROW-BORE CAPILLARY CHROMATOGRAPHY
In previous years a (limited) number of publications have been published dealing
with the subject of non-splitting injection techniques in combination with narrow
bore columns.
In 1983 Schutjes et al. demonstrated some applications of splitless and on-column
injection techniques in narrow-bore columns [22]. On-column injections were
performed by forcing a sample plug through the inlet liner of the injector by a flow
of carrier gas. In this way, a small fraction of the liquid sample was transferred to the
column, while the majority of the sample was vented via the split outlet Good peak
shapes were obtained with this method. Schutjes also reported that obtaining sharp
peaks became more difficult when the column inside diameter was reduced and
proposed modified injector designs.
In 1984, Onuska demonstrated the advantages of the use of narrow-bore columns for
environmental analysis ofPCBs, toxaphene and tetrachlorodibenzo-dioxins [23]. The
separations obtained with narrow-bore columns were compared to separations with
250 J.l.m or 320 J.l.m LD. columns. Sample introduetion was performed in the splitless
or in the on-column mode. Direct sample introduetion onto the column in practice is
restricted to column diameters larger than about 200 IJ.m. Onuska performed direct
introduetion into narrow-bore columns by coupling a wider bore fused-silica
retention gap (250 or 320 J.l.m LD. column) to the 100 IJ.m LD. separation column.
44 Chapter3
Other applications of on-column introduetion with a wide-bore pre-column coupled
to a narrow-bore analytica! column were also demonstrated by Damascenco et al. to
perform high-temperature GC [24] and Attaran Rezaii et al. for liquid-modified
adsorption GC [25] .
In other publications, the possibilities of improving the MDC by using other
injection techniques have been described. Snijders et al. [20] demonstrated on-line
enrichment of the sample by using cryofocusing techniques. In those instances where
the difference in volatility between the solvent and the sample components is
sufficiently large, selective solvent elimination could be achieved by accurate control
of the trap temperature. Sample enrichment on the column was obtained by repetitive
injections. Another way to obtain sample enrichment is by performing solid phase
micro-extraction. Gorochi et al. demonstrated the potentials of this solventless
introduetion technique in combination with fast GC for the rapid analysis ofVOCs at
low concentrations [26,27]. The extraction fibre was inserted into a wide-bore
capillary (0.53 mm LD.) which was connected to the narrow-bore column through a
union. A fast injection was obtained by using flash heating of a segment of the 0.53
mm fused-silica capillary. A short length of this tubing was heated at a rate of
roughly 1 000°C/sec by capacitive discharge during a very short period.
As will be clear from the above overview, the possibility to use non-splitting
injection techniques in combination with narrow-bore capillaries bas been
demonstrated by different authors. However, a detailed discussion bas not been
presented. In this chapter, the possibilities and limitations of these injection
techniques are evaluated. First, some practical aspects of hot splitless, cold splitless
and on-column injections will he discussed in more detail. The experimental
consequences of using narrow-bore columns in combination with these non-splitting
injection techniques will be discussed and various applications will he shown as
ill ustration.
Non-splitting injection techniques for narrow-bore capillary GC 45
3.5 EXPERIMENTAL
The experimental work was performed on a Fisons 8000 gas chromatograph (Fisons,
Milan, Italy) equipped with an EL 980 FID operated at 360°C. For the experiments
of large volume sampling, another Fisons 8000 gas chromatograph enabling faster
heating rates of the oven, was used. The higher programming rate enables a faster
analysis. The two sample introduetion systems available on the instruments are a
SSL 71 split/splitless injector and a cold split-splitless injector (Figure 3.3)
- Carrier gas
Liner
Splitting line
Capillary column
Figure 3.3
Schematic diagram ofthe Fisons cold SSL injector
46 Chapter 3
controlled by the Multi Function Actuator model 815 (Fisons, Milan, Italy). For the
experiments different columns were used. The first column was a 5 m x 67 f.tm LD.
CP-Sil 5 fused silica column coated with a 0.1 f.tm film of stationary phase
(Chrompack, Middelburg, the Netherlands). Also a 10 m x 50 f.tm LD. capillary
column coated with a 0.17 f.tm film of DB-1 (J&W scientific, Folsom, CA, USA)
and a 10 m x 100 f.tm LD. HP-1 column with a 0.17 f.tm film (Hewlett Packard, Palo
Alto, CA, USA) were used. In order to obtain the optimal carrier gas flows forthese
narrow-bore columns, high inlet pressures are required. The required inlet pressure
ofthe helium carrier gas for the 67 and the 100 f.tm LD. columns was 6 bar. Pressures
up to 18 bar were used for the 50 ~-tm column. The gas chromatograph allowed inlet
pressures up to 7 bar. To enable the use of higher inlet pressures, the gas
chromatograph carrier gas inlet was by-passed and a Tescom 44-1100 high pressure
regulator (Tescom Inc., Minnesota, USA) was installed. Data acquisition was
performed using a VG Xchrom data system running under Microsoft WindowsNT
(VG Data Systems, Cheshire, UK).
3.5.1 Practical considerations on hot splitless injections
Lower minimum detectable concentrations are obtained with splitless injections
relative to split injection because virtually the entire sample material is transferred to
the column. During splitless injections, the carrier gas transfers almost the entire
vaporised sample plug from the injector into the column. The success of splitless
injections relies on reconcentration effects, i.e. solvent focusing and cold trapping
effects, which sharpen the solute bands before the separation process starts.
Incomplete sample transfer and poor peak focusing are the most widely observed
problems in splitless injections.
The optimal column flow for narrow-bore columns is in the range of 0.25 to 1
ml/min (measured under atmospheric conditions). Because ofthe high inlet pressure,
the volumetrie flow under injector conditions is even smaller. This is illustrated in
Figure 3.4. The optimal separation conditions (maximum separation efficiency) for
several column inside diameters were calculated with the computer program of
Leclercq and Cramers [28]. In this figure it can be observed that the optimal flow
under atmospheric conditions behaves almost linear with the I.D. of the separation
Non-splitting injection techniques for narrow-bore capillary GC 47
50 2.5
- Optimal Inlet Pressure
i 40 2.0 - Optimal flow under
~ -atmospheric conditions c .... rJl 30 1.5 :§ ~ s ö '-" - 20 1.0 ~ .s 'a ~
·~ 10 0.5 8 liner conditions
00 100 200 300 400 500 0.0
Column Inside Diameter (J.lm) Figure 3.4 Inlet pressure, column flow under atmospheric conditions and column flow under liner conditions as a function of the column inside diameter ( operating under optimal separation efficiency conditions). Experimental conditions: C20, T = 473K, ~ = 200, k = 3.9, stationary phase: SE-30, carrier gas : He, Po= 1 bar, N = 100000.
column. The inlet pressure, however, increases sharply for columns with an LD.
smaller than 200 J.lm. F or this reason, the optimal flow under inlet conditions is very
low for narrow-bore columns. For 100 J.lm LD. columns for example, the optima!
column flow rate is about 0.4 mVmin under atmospheric conditions. Under inlet
conditions, the flow is about 5 times smaller. For 50 J.lm I.D. columns, this ratio is
increased to about 10. Since in splitless injections only the column flow is utilised
for sample transfer from the injector to the column, the transfer times required can be
very long. For this reason, Grob Jr. postulated in 1986 that narrow-bore columns
(I.D.<200 J.lm) are ruled out for splitless injections [29]. According to Grob, durlog
sample transfer the vapour plug in the liner is steadily growing. At low gas flow
rates, this broadening is more pronounced than the transfer to the column. Even if
high gas veloeities are applied, in Grob's expectations, sample transfer is insufficient.
In contrast to Grob's predictions, the successful application of narrow-bore columns
in combination with splitless injection techniques is possible if the experimental
conditions are carefully optimised. First of all, in order to accelerate sample transfer,
48 Chapter 3
liners with a small inside diameter have to be used. In this way, the volume of the
liner is reduced and sample transfer is accelerated. Additionally, if the oven
temperature is morethansome 30°C below the (pressure corrected) boiling point of
the solvent, the vacuum created upon recondensation of the solvent in the column
accelerates the sample transfer noticeably [30]. In this way, the flow rate into the
column is increased. Finally, due to the high inlet pressure, the diffusion coefficients
of the solutes are low. Under these conditions, it is possible to obtain good sample
transfer by splitless injections into narrow-bore columns within a short time, even at
low gas flow rates. In this chapter the influence of transfer times on the recoveries
seen for different liner diameters in splitless injections in combination with narrow
bore columns is studied.
3.5.2 Practical considerations on cold splitless injections
Already almost 20 years ago Vogt and co-workers [31 ,32] constructed a temperature
programmabie (PTV) injector and applied it for the introduetion of large volumes (up
to 250 J.tl) in biomedical and environmental applications. Although the PTV injector
in the years following that pubHeation hardly received any interest for large volume
sampling, it received considerable attention as an alternative injector to hot
split/splitless injectors. The groups of Schomburg [33] and Poy et al. [34]
demonstrated that PTV sample introduetion offers a number of advantages in
comparison with hot splitless injections. The most important advantage of
temperature programmabie injections is that discrimination of high boiling
compounds is virtually absent. The quantitative performance of a PTV injection is
generally better compared to hot splitless injection techniques.
An important factor to optimise in PTV sample introduetion is the position inside the
vaporising chamber where the syringe ends. Care should be taken to introduce the
sample in the properly heated region of the injector, otherwise discrimination can
occur for the high boiling compounds. In the top and bottorn section of the liner
temperature gradients will be present. If the sample is released in a region that is not
adequately heated or if the entrance of the column is positioned at a location where
the temperature is too low, losses can occur for the high boiling compounds. With a
thermocouple, the temperature profile over the length of the liner was measured. In
Non-splitting injection techniques for narrow-bore capillary GC 49
the region from 4 to 9 cm below the upper end of the liner, for the present injector
the temperature was found to correspond to the set temperature. In order to eliminate
discriminatien because of insufficient heating in the upper part of the injector, a long
needie (about 8 cm) was used. Also the original septurn head was replaced by a
custom-built septurn head with a smaller height than the original version. In this way,
the syringe tip reached the correct temperature zone. A liner with a low inside
diameter (0.8 mm) was used for the reasons outlined above.
3.5.3 Practical considerations for on-column injections
For hot and cold splitless injection techniques, the vaporisation of the sample inside
the syringe needie or in the injector vaporisation tube, as well as the splitting and
transfer of the aerosol or sample vapour formed in the injector, may cause problems
in quantitative analysis [35]. The on-column injection technique is therefore
generally the preferred technique for the analysis of high boiling compounds. Also
for the analysis of thermally labile species on-column injection shows a better
performance. On-column injection offers an improved performance because the
(liquid) sample is injected directly into the column. There is no intermediate
evaporation step and contact with other extemal column devices is avoided.
Overheating of the sample, as can occur during extemal vaporisation, is avoided.
F or these reasons, the on-column injection technique is a superior injection technique
for quantitative analysis.
In order to enable direct on-column injection onto a narrow-bore column, a retention
gap with a wider inside diameter has to be used. The conneetion between the
retention gap and the separation column can not be made using standard glass
pressfit column connectors. Due to the high inlet pressures, leaks occur frequently.
Apart from being leak tight, the column conneetion should also have a very low
internat volume. The use of conventional low dead volume unions results in
broadened peaks. This is most likely because the outside diameter of the narrow-bore
column is smaller than the inside column diameter of the retentien gap. Therefore, if
the narrow-bore column is pushed too far, it penetrates into the retentien gap. In this
way, an unswept dead volume is created. One way to overcome this problem is
allowing a small teak flow. Reliable quantitative analysis with such systems is
50 Chapter 3
Figure 3.5 Schematic drawing of the union used for coupling a normal-bore retention gap and the narrow-bore separation column.
virtually impossible due to changes in the split ratio as the temperature program
progresses. To couple the retention gap to the narrow-bore column, a special
conneetion was developed. It is shown schematically in Figure 3.5. The device is
similar to a conventional pressfit connector. The bore, however, is very narrow, so
that the narrow-bore column cannot be pushed through. Proper sealing is obtained by
using polyimide ferrules.
3.6 RESULTS AND DISCUSSION
3.6.1 Splitless times
As pointed out before, in splitless injections only the column flow is utilised for
sample transfer from the injector liner to the column. Since the optima} flow for
narrow-bore columns is in the range from 0.25 to 1 mi/min (measured under
atmospheric conditions ), the transfer times required for quantitative transfer yields
can be very long. In order to avoid too long splitless times and to obtain good
transfer yields for splitless injections in combination with narrow-bore columns, a
liner with a small inside diameter has to be used. We investigated the peak area for
Non-splitting injection techniques for narrow-bore capillary GC 51
n-nonane dissolved in hexane as a function of the splitless time for hot splitless
injections with liners of different inside diameters. The results are shown in Figure
3.6. The column flow rate under atmospheric conditions was approximately 0.7
ml/min (column: 5 m x 67 J.lm LD., Pi= 6.5 bar). It can beseen that despite the low
carrier gas flow rate, the time required for quantitative transfer of the vaporised
samples to the column is surprisingly short for the liner with an LD. of 0.8 mm. For
the wider-bore liner (I.D. = 1.2 mm) the splitless time required to obtain complete
sample transfer is significantly longer. For the liner with an LD. of 1. 7 mm, complete
sample transfer was no longer possible. It seems that under these conditions, and in
accordance to Grob's expectations, diffusional spreading exceeds sample transfer to
the column. From this, it can be concluded that sample transfer is readily achievable
ifliners with a small I.D. are used.
It should be emphasised that even if liners are used with a small inside diameter,
situations can occur where the column flow is insufficient to obtain good transfer
yields. For the 10 m x 50 J.lm LD. column, the inlet pressure resulting in maximum
0 20 40 60 80 100 120
Splitless time (sec) Figure 3.6 Area response for n-nonane dissolved in n-hexane as a function of the splitless time for different liner inside diameters. GC Column: CP-SilS, L = 6.5 m, l.D. = 67 J.tffi, dr= O.lJ.tm. Experimental conditions: Liner diameter: 0.8 mm ('t'), 1.2 mm (A.) and 1.7 mm (e), Pi= 6.5 bar, Tinj = 275°C, Toven = 50°C (3 min)~ 10°C/min ~ 80°C.
52 Chopter 3
separation efficiency is about 13 bar. Under these conditions, the recovery of all
analytes was below maximum, even if longer splitless times were applied. It seems
that under these conditions, the optimum column flow for maximum separation
efficiency is too low to yield maximum recovery. In order to maximise the recovery,
the inlet pressure has to be increased to about 18 bar. Under these conditions, the
average linear carrier gas velocity, and hence the column flow rate under outlet
conditions, is increased by a factor of (18/13i (see the Hagen-Poiseuille equation
( equation 2.17)) while the ditfusion of the analytes in the liner is reduced by a factor
of 13/18, resulting in higher transfer yields.
The increased inlet pressure does not necessarily hinder the potential to perform
high-speed separations with these 50 J.UU LD. columns. It should be stressed that by
increasing the inlet pressure above the optima} inlet pressure corresponding to the
conditions for maximum separation efficiency, the analysis speed is improved as
long as the measured carrier gas velocity overrules the required column length
increment [36]. The gain in analysis time is about 10%. For large values of P, the
minimum analysis time is obtained for a carrier gas velocity, Uo,opa, which can be
calculated according to:
Uo,opt2 = .J2 Uo,optl (3.2)
where Uo,optl is the average carrier gas velocity for maximum separation efficiency. In
experiments, the inlet pressure required for minimum time operation should be
increased to approximately 16 bar. The inlet pressure resulting in maximum transfer
yields is only slightly higher. For this reason, it is reasonable to conclude that it is
still possible to successfully perform high-speed separations while at the same time
obtaining optimum transfer yields under these experimental conditions.
3.6.2 Liner capacity for splitless injectioos
The internat volume, and hence the sample loadability (volume capacity) ofthe liner,
decreases proportionally to the square of the liner inside diameter. Since small
injector liners have to be used in order to obtain good transfer yields, the sample
capacity of these liners is very low. Fortunately, the sample volume is reduced
proportional to the column inlet pressure. In order to avoid losses of sample by
Non-splitting injection techniques for narrow-bore capillary GC 53
overflow of the liner, it is important to establish the (maximum) sample capacity of a
liner. A simple method to do so, is the injection of increasing volumes of n-nonane
dissolved in hexane. In this series of experiments a septurn purge of approximately
10 mllmin was applied. If the sample volume injected exceeds the capacity of the
liner, the sample will be lost through the septurn purge exit and the absolute peak
area at increasing sample volume will become constant. The experimentally observed
upper limit of the sample capacity for a hot splitless injection using a liner with an
inside diameter of 0.8 mm was approximately 1 }ll liquid sample (Experimental
conditions: 67 }liD I.D. GC column, Tinj = 275°C, Pi= 6.5 bar, position ofthe needie
inside the liner = 25 mm).
For hot splitless injections the maximum sample volume is determined by the volume
of sample vapour created by flash evaporation of the liquid sample. For a cold
splitless injection, the situation is fundamentally different as here there is a
possibility to control the speed of sample evaporation. Larger sample volumes can be
injected because solvent evaporation now is carried out slowly at a low injector
temperature. Therefore, the sample amount that can be injected now is no longer
limited by the vapour capacity of the liner, but is now limited by the amount of liquid
solvent that can be retained in the liner. To evaluate the liner capacity of a cold
splitless injection, the response of C14 was evaluated as a function of the volume of
sample injected (GC column: 100 }liD l.D., position ofthe needie in the liner: 40 mm,
Pi 6 bar). It was found that the response behaves linearly up to a sample volume of
3 }llliquid. For higher sample volumes, the area response starts to flatten and the
standard deviation ofpeak areas increases sharply. Overtoading ofthe liner is due to
the fact that the flooded zone in the liner is so long that the liquid reaches the bottorn
of the liner and leaves the injection port via the split exit. Only part of the solute is
transferred to the column and quantitative transfer of the analytes to the column is no
longer possible.
So far only an empty liner was used. In order to increase the volume of liquid that
can be retained by the liner, a packed liner can be used. In the present work, the liner
was packed with Supelcoport 60/80 packing material (Supelco, Bellafonte, PA,
USA) between two plugs of glass wool to keep it in place. The length of the bed was
approximately 4 cm. The bed was positioned in the part of the liner which is
54 Chapter 3
adequately heated. The maximum sample volume that can he injected was
detennined visually. No column was installed, a helium flow conesponding to the
splitflow of 200 mi/min was applied and the injector was set at a low temperature
{35°C). For too large sample amounts, liquid dropiets were observed leaving the
injector port. For hexane as the solvent the maximum injection volume was about 25
J.ll (length ofthe packed bed: 4 cm).
3.6.3 Focusing effects on narrow-bore columns
One of the fundamental laws in chromatography states that the initia! solute band
width must he small compared to the chromatographic peak broadening. Because
during splitless injections, sample transfer takes place over a long period, refocusing
effects are required to reconcentrate the initia! input band. Full exploitation of these
mechanisms i.e. solvent effect and cold trapping, requires carefut optimisation of the
injection conditions. Failure to do so results in distorted peaks. Important with
respect to the solvent effect is the length of the flooded zone. This parameter depends
on several conditions such as the boiling point of the solvent, the chromatographic
conditions, the sample size, etc. Efficient refocusing is more difficult to obtain the
Jonger the flooded zone in the capillary inlet. The flooded zone for hexane in the
column inlet is roughly 20 to 25 cm per microlitre sample for capillaries with an I.D.
of 250 to 320 J.lm coated with a non-polar stationary phase [37]. For narrow-bore
columns, the length of the flooded zone is obviously much longer. Therefore,
obtaining a good solvent effect is more difficult the smaller the inside diameter of the
column.
lt was found that fora hot splitless injection, sample quantities of only about 0.01 J.ll
might already cause peak distortion when 50 or 100 J.lm LD. columns were used. An
example of a hot splitless injection of 0.5 J.ll on a 50 J.lm LD. column at different
oven temperatures is given in Figure 3.7. At an oven temperature of 40°C (Figure
3.7A), the alkanes n-C8 up to n-C 10 are well focused by the solvent effect.
Apparently, the initial conditions are optima! with regard to the solvent effect. Also
for the high boiling compounds good peak shapes are observed. For semi-volatile
compounds, however, severe peak distortion is observed. Apparently, these
compounds, partly reconcentrated by cold trapping, are spread over a largelengthof
Non-splitting injection techniques for narrow-bore. capillary GC 55
the stationary phase by the relatively large amount of recondensed solvent. If the
oven temperature is increased to 70°C (Figure 3.7B), n-C 11 is efficiently focused by
solvent trapping, but peak distortion is still observed for n-C12 up to n-C17•
Additionally, n-C8, n-C9 and n-C10 now elute under the broadened solvent peak. A
possibility to reduce the amount of recondensing solvent is to increase the column
temperature to about l5°C or less below the boiling point of the solvent under
A
3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0
Time (min)
B T.,.".. = 70°C
c,s c,6 c,7 c c c20 cu 18 19
3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0
Time(min)
Figure 3.7 Hot splitless injection of 0.5 f.ll of a homologous series of alkanes from n-C8 to n-C20 dissolved in hexane. GC column: DB-I, L = 10 m, I.D. = 50 IJID, dr= O.l71Jm. Experimental conditions: Tinj 325°C, Toven = 40°C (Figure A) and 70°C (Figure B), isothermal during 3 minutes and than ballistically heated to 320°C, Pi 18 bar.
56 Chapter 3
column inlet conditions [29]. This means that under the experimental conditions the
oven temperature should be increased to 130 or 140°C. At these oven temperatures,
however, all compounds up to C15 elute under the solvent peak. Ill shaped tailing
peaks are still observed for the other compounds.
Probieros similar to those described above for hot splitless injections also occur in
the cold splitless injection mode. For cold splitless injections with 50 p.m LD.
columns severe peak distortion for the semi-volatile compounds was already
observed for sample volumes of a few tenths of a microlitre. For cold splitless
injections with 100 p.m I.D. columns, however, the sample volume can be increased
to 0.5 or 1 p.l before peak distortion is observed. Peak focusing is much better for
cold splitless injections than for injections of the same sample volume in the hot
splitless mode. A plausible explanation is that for cold splitless injections the heating
rate of the injector is rather slow - in our experiments, about 1 minute is required to
increase the injector temperature from 40°C up to the final temperature of 399°C - so
that part of the solvent is already transferred to the column before the analytes enter
the column. In this way, the effects of an excess of recondensed solvent in the
column inlet are reduced and peak distortion can be avoided. It was found that for
sample volumes surpassing 0.5 p.l, good focusing effects were only obtained if the
injector temperature is kept low for a short period after the injection. During this
time, part of the solvent is transferred onto the column before the analytes are
conveyed.
Two examples of cold splitless injections are given in the Figures 3.8 and 3.9. In
Figure 3.8, a fast separation of a citrus oil is illustrated. In Figure 3.9, a fast
separation of a diesel oil sample is shown. The latter separation is completed within
approximately 15 min. In order to obtain the same separation efficiency with a
conventional 320 p.m column an analysis time of at least 30 to 40 minutes is
required.
Non-spZitting injection techniques for narrow-bore capillary GC 57
0 2 4 6 8 10 12
Time (min)
Figure 3.8 Cold splitless injection of 0.5 J!l sample of a citrus oil dissolved in hexane. GC column: HP-1, L 10 m, I.D. = 100 J.lm, dr = 0.17 J.lm. Experimental conditions: Tinj = 40°C ballistically heated to 399°C (5 min), splitless time= 45 s, Toven = 35°C (3 min) and than ballistically heated to 320°C, Pi = 6 bar.
pristane
0 5 10 15
Time (min)
Figure 3.9 Cold splitless injection of 0.5 J.ll sample of diesel oil dissolved in hexane. GC column: HP-I, L = 10 m, LD. = 100 Jlm, dr = 0.17 J.lm. Experimental conditions: Tinj = 40°C ballistically heated to 399°C (5 min), splitless time 45 s, Toven = 35°C (3 min).....;. 20°C/min .....;. 320°C (5 min), Pi= 6 bar.
58 Chapter 3
3.6.4 Focusing effects of retention gaps in cold splitless and on-column
injections
To overcome the problem of peak distortion in the cold splitless mode due to
excessive solvent recondensation of at increased sample volumes, a retention gap can
be used. The use of a retention gap with a larger I.D. has the advantage that the
length of the flooded zone will be reduced and peak distortion at higher sample
volumes is less likely to occur than with usage of narrow-bore columns.
Additionally, ifthe flowing sample liquid spreads the solute material in the uncoated
capillary, the solute bands will be reconcentrated at the beginning of the coated
column by phase ratio focusing. In our experiments, a retention gap of about 1.5 rn
and 320 J.lm LD. was used. The conneetion of the retention gap to the analytica!
column was made by the special union described earlier. With this pre-column it was
found that after the optimisation of the instromental parameters a sample volume of
about 3 to 4 J.ll (maximum sample capacity of an empty liner) could be introduced
without any peak distortion.
Also for the on-column injections, a retention gap with a larger I.D. was used to
allow direct sample introduetion onto the column. In cold splitless injections, the
sample amount that is transferred to the column is limited by the sample capacity of
the liner (3 to 4 J.Ll). For on-column injections, larger sample volumes could be
introduced. With a retention gap of 1.5 m and a 320 J.Lm I.D., good peak focusing
was observed for all compounds up to sample volumes of about 5 to 6 ~-tl. An
example of a 4 ~-tl on-column injection of a series of alkanes dissolved in hexane is
shown in Figure 3.1 0. As can be observed, good focusing is obtained for all
compounds. For sample volumes of 6 ~-tl and higher, again peak distortion takes place
for semi-volatile compounds. Apparently at this volume the flooded zone can no
Jonger be accommodated by the retention gap and peak distortion occurs.
An application of an on-column injection is demonstrated in Figure 3.11, depicting
the separation of a mixture of polymer additives. The retention time is relatively high
since the fmal oven temperature was only 330°C to avoid excessive column
bleeding.
Non-splitting injection techniques for na"ow-bore capillary GC 59
c,2 c,6
c2o c24 c2s c32 c36
'--
0 3 6 9 12 15 18
Time(min) Figure 3.10 On-column ÎI\Îection of 4 ftl of a homologous series of alkanes from n-C8 to n-C20 dissolved in hexane. GC column: HP-1, L = 10 m, LD. = 100 ftm, dr = 0.17 ftm. Experimental conditions: Toven = 35°C (2 minutes) ballistically heated to 320°C (5 min), Pi= 6 bar.
\
0 5
Figure 3.11
10 15 20
Time (min)
On-column injection of 4 J.ll of a polymer additives sample dissolved in hexane. GC column: HP-1, L = 10 m, I.D. = 100 J.lm, dr = 0.17 J.lm, Toven = 35°C (2 minutes) ballistically heated to 330°C (10 min), Pi= 6 bar.
60 Chapter 3
In recent years, the PTV received considerable attention as a means of introducing
large sample volumes, up to 100 111 and even more. Large volume sampling using
PTV injectors is based on selective elimination of the sample solvent ftom the liner
of the PTV injector while simultaneously trapping the analytes with a low volatility
on a bed of packing material in the cold liner. When solvent elimination is completed
and the split exit is closed, the liner temperature is increased and the components are
transferred to the column. The liner has to be packed to prevent the liquid sample
ftom being pusbed to the base of the injector which could result in losses via the split
exit or in flooding of the column inlet. As discussed already in section 3.6.2, the
maximum sample capacity for the packed liner is about 25 111. Sample volumes
exceeding this volume have to be introduced in several steps or, more elegantly, in a
speed controlled manner.
In Figure 3.12, an example of a 20 J.1llarge volume PTV "at once" injection is shown.
Also here, as for on-column injections, the volatile compounds elute under the
solvent. Minor tailing was observed for C10 to C20 while good focusing is observed
for the higher boiling compounds. The slight peak tailing for the semi-volatile
compounds is due to the excess of solvent transferred to the column. On the one
hand, it is important to apply a long solvent elimination time to minimise the amount
of solvent that is transferred to the column. Opposedly, a minimum volume of
solvent is required in the liner in order to prevent the volatile compounds ftom being
vented together with the solvent. If more solvent is vented via the split, the
recoveries of the volatiles will decrease.
To enable injections of sample volumes larger than 1 111 in the splitless mode (either
or not in combination with solvent elimination) or in the on-column mode, a
retention gap has be used to enable good focusing effects. The most critical step here
is the conneetion between the normal-bore retention gap and the narrow-bore
separation column. Several unions have been evaluated with regard to their dead
volumes. It was found that with each of these unions it was impossible to avoid
tailing, especially for the semi-volatile compounds that were not fully cold trapped
on the analytica} column. The only way to overcome this problem is to use an
additional cold trap system.
Non-splitting injection techniques for narrow-bore capillary GC
0 3 6 9
Figure 3.12
12
Time (min)
61
Large volume at once injection (Vinj = 20 J.tl) of a homologous series of alkanes from n-C8 to n-C36 dissolved in hexane with solvent venting. GC column: HP-1, L = 10 m, LD.= 100 Jlm, dr == 0.17 Jlm. Experimental conditions: Tinj 40°C (90 s) ballistically heated to 400°C, solvent vent time= 60 s, splitless time= 15 sec, Toven 35°C (3 min) ballistically heated to 330°C, Pi = 6 bar.
3.6.5 Discrimination
The major concern of any sample introduetion technique is accuracy and precision in
quantitative analysis. For hot and cold splitless injections, the recovery of the
analytes depends on three steps. First of all, for hot splitless injections, most samples
already start to evaporate inside the needie when it is inserted into the hot injector
resulting in unreliable de liverance of the sample into the injector. To avoid selective
evaporation and to minimise discrimination, the injection has be carried out as
quickly as possible thereby preventing a "hot needle" situation. Selective evaporation
can also be avoided by injecting the sample in a cold injector. The next step of
sample transfer is that the analytes in the liner have to be vaporised to be transferred
to the column. Since the boiling point of the analytes is increased because of the
elevated inlet pressures in narrow-bore GC, the vaporisation step will become more
critica! and discrimination will become even more pronounced. The last step is the
transfer of the vaporised analyte. It was demonstrared before that the required
splitless time can be within acceptable limits if narrow-bore liners are used and the
62 Chapter 3
column flow is sufficiently high.
In Table 3.1, the recovery and the relative standard deviation ofthe peak areas fora
homologons series of alkanes dissolved in hexane in a hot splitless injection for 50
and 100 J.lm I.D. columns are shown. When 50 J.lm LD. columns are used, the
reeoverles start to decrease at roughly C24• C36 was not detected at all. It was found
that the recoveries for high boiling compounds can be increased by working at higher
injector temperatures. The standard deviations for peak areas increased sharply for
the high boiling compounds. For 100 J.lm LD. columns the recoveryforthese solutes
was significantly higher but still severe discrimination was observed.
As mentioned above, discrimination is an important drawback of the use of hot
splitless injections in capillary GC. lf the splitless transfer is carefully optimised, in
cold splitless injection discrimination by selective evaporation from the syringe and
sample back ejection due to explosive evaporation is avoided. In Table 3.2, the
Table 3.1 Recovery and standard deviation of peak areas for a homologous series of alkanes dissolved in hexane (Tinj 320°C) fora hot splitless injection (Vinj = 0.5 J.tl) in combination with a 50 J.lffi LD. and a 100 J.lm LD. column.
Cx 50 Jlffi I.D. column 100 J.lffi LD. column Recovery Rel. Stand. Dev (%) Recovery Rel. Stand. Dev (%) (compared to C16) for peak areas (n"" 11) (compared to c!6) for peak areas (n='6)
8 0.79 6.5 0.74 5.4 10 0.87 5.9 0.99 6.5 12 0.95 5.1 0.93 5.6 14 0.98 6.4 0.94 5.2 16 1.00 4.9 1.00 5.4 18 0.97 5.4 0.97 6.2 20 0.88 6.1 0.97 5.5 22 0.81 10.8 0.87 11.0 24 0.71 14.7 0.86 17.8 28 0.40 20.6 0.64 20.5 32 0.15 28.2 0.53 16.4 36 0.34 17.7
Non-splitting injection techniques for narrow-bore capillary GC 63
recovery, the relative standard deviations for peak areas, the minimum detectable
amounts and the minimum detectable concentration are shown for a cold splitless
injection in combination with a 100 J..tm LD. column. The minimum detectable
amount is defined as:
(3.3)
where QW is the minimum detectable amount for a mass flow sensitive detector, RN
the noise level, sm the sensitivity and O'tot the standard deviation of the
chromatographic peak of the overall band width of the peak. The sensitivity for a
mass flow sensitive detector can be calculated according to:
Sm=A/Q (3.4)
Table 3.2 Recovery, standard deviation, mmunum detectable amount and mmunum detectable concentration for a homologous series of alkanes dissolved in hexane for a cold splitless injection (Vmj = 0.5 J.!l) in combination with a 100 Jlm I.D. column.
Cx Recovery Rel. Stand. Dev (%) MDA MDC (compared to C16) for peak areas (n=6) (pg) (ppb)
8 0.91 2.4 3.0 5.9 10 1.09 4.0 4.3 8.5 12 0.91 3.9 4.5 9.0 14 0.98 2.7 4.7 9.4 16 l.OO 3.2 4.3 8.7 18 l.OO 2.1 4.6 9.2 20 0.99 3.1 5.1 10.3 22 0.99 3.8 5.4 10.8 24 1.02 3.4 5.5 11.1 28 0.91 3.8 7.1 14.1 32 0.94 2.5 7.8 15.6 36 0.94 3.3 8.8 17.6
64 Chapter 3
where A is the area response and Q the mass of sample amount of a compound
injected onto the column. Before the calculations were performed, the linearity of
the detector response was evaluated. It was found that the detector response behaves
linearly over the entire range from the detection limit up to the maximum sample
capacity of these narrow-bore columns. The minimum detectable concentration was
calculated according to:
m ~ 2 RN O'tot C0 = '\/21t -- --sm Vinj
(3.5)
where VinJ is the injected sample volume. The relative standard deviation for the area
response for cold splitless injections is about 3%. Only for C10 and C12 higher
standard deviations were observed. This can be explained by the fact that these peaks
are slightly broadened thereby resulting in small inlegration errors. When camparing
the results for the recoveries for cold splitless injection with final temperatures of
300°C and 400°C respectively, it was found that the recovery for all compounds
were comparable but the relative standard deviations were reduced from about 10 to
3% upon increasing the final temperature from 300°C to 400°C. For this reason, all
further experiments were carried out with a PTV final temperature of 400°C. The
minimum detectable amount was approximately 5 pg for all compounds,
corresponding to a minimum detectable concentration of 10 ppb at an injection
volume of 0.5 J!l. The detection limits for the high boiling compounds are slightly
higher because of the increased chromatographic peak broadening.
The results for recoveries, the standard deviation, the minimum detectable amount
and concentration for on-column injections are shown in Table 3.3 (column I.D.
100 Jlm, Vinj = 5 J.tl). The relative standard deviation is about 3% for the semi-volatile
compounds and decreases to below 1% for the higher boiling compounds. The
standard deviation of the peak areas for the semi-volatile compounds is slightly
higher due to integration errors because oftailing of these compounds. This tailing is
caused by the large amount of solvent injected onto the column. The minimum
detectable amount is in the low pg range, resulting in a minimum detectable
concentration of 1 to 2 ppb.
Non-splitting injection techniques for narrow-bore capillary GC 65
Table 3.3 Recovery, standard deviation, mmtmum detectable amount and minimum detectable concentration for a homologous series of alkanes dissolved in hexane for an on-column injection (Vinj = 5 f.ll) onto a 100 f.!m l.O. column.
Cx Recovery Rel. Stand. Dev (%) MOA MDC (compared to C16) for peak areas (n=6) (pg) (ppb)
8 10 12 1.03 3.7 6.5 1.3 14 0.91 3.1 6.4 1.4 16 1.00 2.4 7.4 1.8 18 1.04 2.6 8.8 1.8 20 1.04 1.9 8.1 1.6 22 1.03 0.9 7.9 1.6 24 1.07 0.8 8.2 1.6 28 0.94 1.0 9.2 1.8 32 1.01 0.9 9.2 1.8 36 1.03 0.9 8.9 1.8
In Figure 3.13, the recoveries for the homologous series of alkanes for the different
injection techniques are compared. The recoveries for hot splitless injections are
significantly lower compared to the recovery for cold splitless and on-column
injections. This could be due to either differences in transfer from the needie into the
liner (due to selective evaporation) or in the transfer from the injector liner to the
column (due to low liner flow rates). It has been demonstrated that complete transfer
yields can be obtained if small LD. tiners are used, even for 50 Jlm LD. columns.
Apparently, the discrimination seen for hot splitless injections is solely due to
selective evaporation of the sample from the needle. High boiling solvents can be
used to minimise evaporation of the sample inside the syringe needie thereby
reducing discrimination [38]. At the inlet pressures used, the boiling point of the
solvent (hexane) is increased to 139°C for the 100 )lm LD. columns and even to
193°C for the 50 f.!m LD. columns. Because the recovery is higher for the hot
66 Chapter 3
100
80
60
40
20
oL,--.--.--.--.--.--.--.--.--.--.~~
8 10 12 14 16 18 20 22 24 28 32 36
ex Figure 3.13 Comparison of the recovery for a homologous series of alkanes for hot splitless injections with 50 Jlm (.a.) and 100 Jlm ('f') I.D. columns (see Table 3.1), cold splitless injection (•) (see Table 3.2) and on-column injections (+) (see Table 3.3) with 100 Jlm I.D. columns.
splitless injections in combination with the 100 J.1m I.D. columns, it seems that the
effect of the increased boiling point is only of minor importallee compared to the
transfer step of the high boiling solutes from the needie to the liner so that higher
discrimination is observed for the smaller bore columns.
3.6.6 Thermal degradation
Splitless injection can give rise to degradation of thermally unstable components
because of the long residence times of the compounds in the hot liner. As will be
demonstrated in Figure 3.10, severe degradation is observed for endrio injected in
the hot splitless injection mode. To evaluate thermal degradation of endrin, the peak
areas of the degradation products, endrio aldehyde and endrio ketone, the peak areas
were measured as a function ofthe splitless time (Figure 3.14). From this figure it is
clear that the degradation of endrio is appreciable. Degradation of endrio in splitless
injections using conventional GC columns is generally Iimited to 10 to 20%. Because
the I.D. of the liner was 1.5 mm, high degradation is observed due to the long
residence time of the components in the hot injector. For a splitless time of 15 sec,
Non-splitring injection techniques for narrow-bore capillary GC 67
the peak area for endrin is higher than for its degradation produels while at a splitless
time of 1 minute, the peak of the degradation product endrin aldehyde dominates.
The Donike mixture is another test mix that can be used to evaluate degradation of
thermolabile analytes. This mixture contains some trimethylsilylesters of fatty acids
and some linear alkanes which where added as reference solutes. The silyl esters of
the mix can degrade due to hydrolytic activity of the liner. Silanol groups are
generally thought to be responsible for this phenomenon [39]. The results of the
recoveries (compared to C16) found for this mix are illustrated in Figure 3.15. In this
figure, the recoveries of the compounds in a split injection, a hot splitless injection in
a non-deactivated liner and in a hot splitless injection using a liner deactivated with
polymethylhydrosiloxane are shown. lt is clear that the degradation, especially for
the high boiling trimethylsilylesters, is almost fully eliminaled by deactivation of the
liner. It should be kept in mind however, that the degree of degradation might also
depend on the concentradon of the analytes. Especially for high boiling compounds,
the smaller the sample amount, the higher the level of degradation will be, regardless
of the injection method used. For example, for C26TMS the recovery is reduced to
0 2 3
Splitless Time (min)
Figure 3.14 Peak area of endrin and endrin degradation products, endrin aldehyde and endrin ketone as a function of the splitless time. Peak areas of endrin (e), endrin aldehyde (~) and endrin ketone (+).
68 Chopter 3
l.O
= 0.8
~ ~ 0.6 u c.. u
.2: ii 0.4 -~
0.2
Figure 3.15 Relative area response (reference C16) for the compounds in the Donike mix. Split injection (e), hot splitless injection with a deactivated liner (.6.) and hot splitless with a nondeactivated liner (T).
about 80% for a sample amount of about 15 ng. For a sample amount of a few ng
only, the recovery is reduced to only 25%. Most likely, thermal degradation can be
further reduced by slower heating ofthe injector. Unfortunately, this option was not
available with the instrument used. Thermal degradation of the Donike mix could
only be fully eliminated by using the on-column injection technique.
3.7 CONCLUSIONS
The combination of narrow-bore columns with non-splitting injection techniques was
discussed in detail. First of all, it was demonstrated that despite the low column
flows, the splitless time required to obtain quantitative transfer yields in a splitless
injection on a 50 or 100 J.Lm LD. column can be surprisingly low if liners with a small
LD. are used. For hot splitless injections, peak focusing is limited, strong
discrimination is frequently observed and severe degradation of thermolabile
analytes can occur. With regard to these parameters much better results were
Non-spZitting injection techniques for narrow-bore capillary GC 69
obtained with cold splitless and on-column injections. The applicability of 50 !J.m
LD. columns in combination with cold splitless or on-column injections is severely
limited. Because the length of the flooded zone is extremely long, it is almost
impossible to avoid peak distortion, even for small sample volumes of 0.1 !J.l injected
in the cold splitless mode. For cold splitless injections with 100 !J.m columns,
discrimination is absent up to C36 and good peak shapes can he obtained for sample
amounts up to 1 !J.l. In order to maximise the chromatographic information content,
the initial oven temperature should he sufficiently low to obtain good peak focusing.
Additionally, the possibilities of using normal-bore retention gaps in combination
with narrow-bore separation columns were demonstrated. For cold splitless
injections, large sample volumes (a few f..ll) can he injected when aretention gap is
used. The sample volume can he even further increased by performing solvent
elimination before closing the split. For on-column injections, the use of a normal
bore retention gap allows direct sample introduetion onto the column. Good peak
focusing is obtained as long as the entire flooded zone can he accommodated by the
retention gap. With aretention gap of 1.5 m and an LD. of 320 f..lm, sample volumes
up to 6 !J.l could he injected before peak distortion started to occur. Even if using
very low dead volume unions to conneet the retention gap and the separation
column, some peak tailing will always he observed, as caused by low column flow.
The only way to overcome this problem is to use an additional cold trap system at the
column inlet.
3.8 REFERENCES
1. D.H. Desty, A. Goldup, in "Gas Chromatography", R.P.W.Scott (Ed), Butterworths, London, 1960,p. 162.
2. D.H. Desty, Adv. Chromatogr., 1 (1965) 199. 3. G. Gaspar, P. Arpino, G. Guiochon, J. Chromatogr. Sci., 15 (1977) 256. 4. G. Gaspar, J. Olivo, G. Guiochon, Chromatographia, 11 (1978) 321. 5. G. Gaspar, R. Annino, C. Vidal-Madjar, G. Guiochon, Anal. Chem., 50 (1978)
1512. 6. R. Annino, J. Leone, J. Chromatogr. Sci., 20 (1982) 19. 7. R.L. Wade, S.P. Cram, Anal. Chem., 44 (1972) 131.
70 Chapter 3
8. J.B. Angell, S.C. Terry, P.W. Barth, Scientific American, 248 (1983) 44. 9. S. Sadat, S.C. Terry, American Laboratory, 16 (1984) 90.
10. H. Wohl*n, Anal. Chem., 56 {1984) 87A. 11. G. Lee, C. Ray, R. Siemers, R. Moore, Am. Lab., 10 (1985) 124. 12. A. van Es, J. Janssen, R. Bally, C.A. Cramers, J. Rijks, J. High Resolut
Chromatogr. Chromatogr. Comm., 19 (1987) 273. 13. R. Bally, Ph.D. Thesis, Eindhoven Univerisity of Technology, the Netherlands,
1987. 14. H.A. Claessens, M.J.J. Hetem, P.A. Leclercq, C.A. Cramers, J. High Resolut.
Chromatogr. Chromatogr. Comm., 11 (1986) 176. 15. A. van Es, J. Janssen, C.A. Cramers, J. Rijks, J. High Resolut. Chromatogr.
Chromatogr. Comm., 11 (1988) 201. 16. H.M.J. Sijders, H.-G. Janssen, R.M.G. Straatman, C.A. Cramers, in "Proc. 15th Int.
Symposium on Capillary Chromatography", P. Sandra (Ed.), Rivadel Garda, Italy, Hüthig Verlag, Heidelberg, 1993, p. 391.
17. Z. Liu, J.B. Phillips, J. MicrocoL Sep., 1 (1989) 249. 18. A. Peters, M. Klemp, L. Puig, C. Rankin, S. Sacks, Analyst, 116 (1991) 187. 19. J.P.E.M. Rijks, H.M.J. Snijders, A.J. Bombeeck., H.E.M. Leuken, J.A. Rijks, in
"Proc. l3th Int. Symposium on Capillary Chromatography", P. Sandra (Ed.), Riva del Garda, ltaly, Hüthig, Heidelberg, 1991, p. 35.
20. H.M.J. Snijders, J.P.E.M. Rijks, A.J. Bombeeck, J.A. Rijks, "Proc. l4th Int Symposium on Capillary Chromatography", P. Sandra, M.L. Lee (Eds.), Baltimore, Maryland, USA, Hüthig Verlag, Heidelberg, 1992, p. 46.
21. J.P.E.M. Rijks, H.M.J. Snijders, A.J. Bombeeck, J.A. Rijks, ''Proc. 14th Int Symposium on Capillary Chromatography", P. Sandra, M.L. Lee (Eds.), Baltimore, Maryland, USA, Hüthig Verlag, Heidelberg, 1992, p. 54.
22. C.P.M. Schutjes, Ph.D. Thesis, Eindhoven University of Technology, the Netherlands, 1983.
23. F.I. Onuska, J. Chromatogr., 289 (1984) 207. 24. L.M.P. Damascenco, J.N. Cardoso, R.B. Coelho, J. High Resol. Chromatogr., 15
(1992) 256. 25. M. Attaran Rezaii, L. Lattanzi, Chromatographia, 38 (1994) 235. 26. T. Gorecki, J. Pawliszyn, Anal. Chem., 67 (1995) 3265. 27. T. Gorecki, J. Pawliszyn, J. High Resol. Chromatogr., 18 (1995) 161. 28. P.A. Leclercq, C.A. Cramers, J. High Resolut. Chromatogr. Chromatogr. Comm., 8
(1985) 764. 29. K. Grob, in "Classica! split and splitless injection in capillary gas chromatography",
Hüthig Verlag, Heidelberg, 1986, p. 132. 30. K. Grob, in "Classical split and splitless injection in capillary gas chromatography",
Hüthig Verlag, Heidelberg, 1986, p. 124.
Non-splitting injection techniques for narrow-bore capillary GC 71
31. W. Vogt, K. Jacob, H.W. Obwexer, J. Chromatogr .• 174 (1979) 437. 32. W. Vogt, K. Jacob, A.-B. Ohnesorge, H.W. Obwexer, J. Chromatogr., 186 (1979)
197. 33. G. Schomburg in "Sample Introduetion in Capillary Gas Chromatography", Vol 1,
P. Sandra (Ed.), Hüthig Verlag, Heidelberg, 1985. 34. F. Poy, L. Cobelli in "Sample Introduetion in Capillary Gas Chromatography", Vol
1, P. Sandra (Ed.), Hüthig Verlag, Heidelberg, 1985. 35. C.A. Saravelle, F. Munari, S. Trestianu, J. Chromatogr., 239 (1982) 241. 36. T.H.M. Noij, Ph.D. Thesis, Eindhoven Univerisity of Technology, the Netherlands,
1988. 37. K. Grob Jr., J. Chromatogr., 213 (1981) 3. 38. K. Grob, in "Classical split and splitless injection in capillary gas chromatography",
Hüthig Verlag, Heidelberg, 1986, p. 222. 39. M. Donike, Chromatographia, 6 (1973) 190.
72
Chapter 4
High-speed narrow-bore capillary gas
chromatography coupled to
electron capture detection
SUMMARY The combination of high-speed narrow-bore capillary gas chromatography with
electron capture detection is evaluated The make-up gas flow rate is a key
parameter in the successful coupling of narrow-bore columns to the ECD detector.
The make-up flow should he sufficiently high to eliminate peak tailing caused by the
large deleetion cell volume. The sensitivities at these elevated make-up flow rates
(400 to 1000 mi/min), measured forsome pesticides like HCB and dieldrin, were
very good. Defection limits for these compounds of 0.1 pg were obtained, resulting
in minimum deleetabie concentrations of approximately 0.2 ppb for a splitless
injection of 0.5 pl sample. The performance of the system is illustrated by several
high-speed analyses of environmentally relevant samples of PCBs and pesticides.
74 Chapter4
4.1 INTRODUCTION
The evolution of high-resolution chromatography bas brought along stringent
requirements for the detection devices. Especially for high-speed narrow-bore
separations, rapid, !ow-volume, sensitive, and selective methods of detection are
required.
One of the most important features of the detector is the sensitivity. Especially for
detectors used in combination with narrow-bore columns, the need for high
sensitivities will be more pronounced since the sample capacity of these columns is
limited. The ratio between the sample capacity and the minimum detectable amount
determines the dynamic range ofthe column/detector combination. In order to obtain
an acceptable working range, the detector should be very sensitive.
The high resolution offered by capillary columns also means that more complex
mixtures can be separated, thus necessitating the use of detectors that respond
selectively to compounds of specific interest. Detectors have been developed that
respond selectively or specifically to certain classes of compounds such as N, P, S or
Cl containing molecules. These compounds can be detected selectively in the
presence of other species not containing these elements. Molecular mass, ion
mobility, ionisation potential, atomie emission, and optical absorbance are used as
the basis for selective detection.
A primary requirement of capillary GC detectors is that the extra column band
broadening is minimised to retain the integrity of the chromatographic separations.
Carefut attention should be paid to interfacing detectors to capillary columns. Short
and low volume transfer Iines from the column to the detector aid in reducing
dispersive effects. With many detectors, the band broadening occurring in the
transfer process can be nearly eliminated by inserting the column directly in the
detector. If direct insertion is possible, the major contribution to post-column band
broadening becomes the volume ofthe detection zone. This latter parameter depends
on the geometry and the flow through the detector. To reduce the residence time of
the analyte in the detector cell and to avoid additional band broadening, a flow of
make-up gas can be used [1]. It is important, however, to investigate the influence of
the make-up gas flow rate on the sensitivity, especially for concentration sensitive
High~speed narrow~bore capillary GC coupled to ECD 75
detectors.
Apart from the additional peak broadening due to geometrie factors, slow
( electronic) response characteristics of the detector and the data acquisition system
will also be translated into high time constants, eventually resulting in distorted
peaks. For high~speed capillary GC, a considerable increase in the sampling
frequency of the data acquisition system is required for accurate registration of the
chromatographic separation. The rapid elution of narrow chromatographic zones
requires short response times of the detector to accurately track the concentration
profile of the sample as it elutes from the column. If the data acquisition is
insufficiently fast, tailing peaks will be observed resulting in delayed retention times
and unreliable quantitation. For accurate registration of a chromatographic peak,
Leclercq stated that about 20 data points per peak are required [2]. As a rule of
thumb, the overall time constant should be at least ten times smaller than the
standard deviation of the peak. If this condition is met, the plate height is reduced by
less than 0.5% while the retention time shift is less than 10% of the standard
deviation of the peak [3]. A disadvantage of fast detection systems is that they
intrinsically generate more noise. In order to improve the detection limits, the noise
can be reduced by noise filtering techniques. Several techniques can be
distinguished. The first one is the analogue filter which is an electronic filter that is
very often already present in the detector amplifier. An RC circuit is deliberately
included in the amplifier to reduce the noise level. The chromatographic peak width
is an important factor in the filtering process since the optimum conditions are
achieved when the time constant of the detector electronics is adjusted to the
chromatographic peak width. The higher the time constant, the lower the noise level
will be. As a consequence of the slow response time, however, peak areas and
retention times of chromatographic peaks will become less accurate. In practical
situations, a campromise has to be found. The detector response time has to be
reduced at the expense of an increased noise level.
Another way to perform noise reduction is obtained by the use of digital filtering
techniques. This can be performed in the time domain as well as in the frequency
domain. Examples operatingin the time domain are the moving average filter [4] and
the Savitsky-Golay filter [5]. Another example is Fourier transfarm where filtering
76 Chapter4
takes place in the frequency domain [4,6-9]. A back transformation to the time
domain provides a filtered signal. An important advantage of filtering in the
frequency domain is that the filter shape can be optimised according to the frequency
spectrum of the noise and the shape and width of the peak.
A series of detectors has been developed and evaluated for high-speed gas
chromatography. In the next paragraph, a brief overview of the various detectors
used in combination with narrow-bore columns will be given.
4.2 OVERVIEW OF DETECTION DEVICESIN COMBINATION WITH
IDGH-SPEED NARROW-BORE COLUMNS
A nearly universa! response to organic compounds, high sensitivity, simplicity of
operation, low dead volume and fast signal response combined with an exceptional
linear response range have contributed to the flame ionisation detector (FID) being
the most popular detector currently in use. For most applications, addition of make
up gas is recommended in order to optimise the detector performance and to rapidly
sweep the detector volume, thus minimising band spreading. The detector response is
a function of the interdependency of the carrier gas flow, make-up flow and the
cambustion flows. Early experiments in characterising the FID showed that there is a
maximum sensitivity at a particular ratio of the make-up gas and the hydrogen flow
[10]. The optimum ratio of the gas flows depends on the particular make-up gas
used. Schutjes et al. [11] evaluated the influence of the addition of make-up gas on
the sensitivity of the detector. It was found that with several make-up gasses like
C(h, N2, Ar and air, the sensitivity was increased. Unfortunately, the increased
sensitivity was in all instances accompanied by an almost equal increase of the
detector noise level so there was no substantial impravement of the detector
performance.
The thermal conductivity detector (TCD), is a universa}, non-destructive,
concentration sensitive detector. Hence, the addition of make-up gas in order to
decrease the volumetrie time constant is unfavourable for the detection limits. To
obtain good detection limits, the detector volume has to be reduced. Van Es et al.
High-speed narrow-bore capillary GC coupled to ECD 77
[12] reported on a J..LTCD with a detection cell volume of only 1.5 nl and a thermal
time constant of 200 J..LSec. He demonstrated that with regard to the detection limits,
the application of a J..LTCD with narrow-bore columns becomes increasingly
advantageous as compared to the FID. An alternative route to reduce the volumetrie
time constant is obtained by increasing the detector flow by the addition of make-up
gas. As this approach results in dilution, the addition of make-up gas is unfavourable
for the detection limits. Altematively, the detector flow can he increased by
operating the TCD under reduced pressures. Van Es et al. [ 12] reported that the
increase in sensitivity was found to he inversely proportional to the detector pressure
while the noise remained constant at reduced pressures. Operation at reduced
detector pressures therefore enables the application of relatively large TCD detection
volumes for high-speed separations.
The photoionisation detector (PID) is another interesting detector because it is a non
destructive detector which can he used in tandem with other detectors. The smallest
detector volume commercially available is 40 !Jl [13], which implies that make-up
gas or reduced pressures should he applied in order to reduce the post-column band
broadening. Levine et al. [14] have demonstrated the applicability of a PID as
detector for very fast chromatographic separations. Van Es et al. [12] found that in
contrast to a TCD, the sensitivity did not increase at reduced pressures. For this
reason, the effect of pressure on sensitivity and detection limits is equivalent to the
use of make-up gas. Due to dilution, a loss of detectability occurs. The option of
reducing the detection cell volume by reducing its dimensions is limited since the
sensitivity is proportional to the path length ofthe cell [12].
The electron capture detector will be discussed in more detail in the next section. A
complete discussion of mass spectrometric detection devices for high-speed narrow
bore separations will be presented in the chapter 5.
78 Chapter4
4.3 ELECTRON CAPTURE DETECTION IN CAPILLARY
CHROMATOGRAPHY
The structure-selective, electron-capture detector is the second most widely used
ionisation detector. In the past the electron capture detector (ECD) has proven to be a
very sensitive, selective and reliable detector. For this reason, the ECD is now one of
the most frequently used detectors for low level detection of specific target substances
in complex samples and becomes especially important for analytica! problems
encountered in environmental and biomedical studies.
In many instances the ECD can be used for direct detection of molecules that have
electronegative functional groups and, therefore, capture electrons without requiring
prior chemica! modification. Examples of these organic compounds include
halogenated and nitroaromatic hydrocarbons. By chemica! derivatisation procedures,
the applicability of the ECD for trace analysis can be expanded further to include
classes of compounds that normally do not respond strongly [15].
The ECD is known to be a relatively complex detector whose operational
characteristics, including those associated with sensitivity, chromatographic resolution,
and quantitative responses, are known to vary significantly with alterations in most of
the experimental parameters set by the operator. lt is neither possible nor necessary to
provide here a comprehensive description and discussion of all the aspects of the ECD,
since those can be found in review articles [16-22]. Despite the importance of the
detector in environmental analysis, until now only limited attention has been paid to
the coupling of high-speed GC to ECD detection. Keet al. [23] used the ECD for fast
gas chromatographic air quality monitoring. The ECD had a cell volume of 90 j..tl and
the make-up gas used was argon/methane ( 5%) at a flow rate of 120 ml/min. The
detection limits obtained were in the range from 1 to 100 pg, corresponding to a
minimum detectable concentration (MDC) of 0.1 to 10 ppb. Schutjes et al. [24]
evaluated the performance of the ECD under reduced pressures in combination with
a normal-bore column (310 J..tm I.D.). As reported by Cramers and co-workers
[25,26], vacuum outlet conditions may appreciably increase the speed of analysis in
capillary GC using conventional inside diameter columns. An additional advantage
of applying vacuum outlet conditions was that the effective cell volume of the ECD,
High-speed narrow-bore capillary GC coupled to ECD 79
which is a well known souree of extra peak broadening, is reduced. Only moderate
vacuum could he applied (down to 0.4 bar) because at lower pressures leakage of air
into the detector cell occurred. Furthermore, a make-up flow had to he supplied, also
under reduced pressure conditions, in order to maintain the sensitivity.
In the following paragraphs, the combination of high-speed narrow-bore capillary GC
with electron capture detection is studied. The influence of the ECD make-up flow on
detector band broadening and sensitivity is investigated. Under optimum make-up flow
conditions, the detection limits and the working range are established. The speed and
sensitivity attainable in high-speed GC with ECD detection is illustrated with various
industrial and environmental applications including the analysis of pesticides and
PCBs.
4.4 INSTRUMENTATION
A Carlo Erba 4160 Fractovap gas chromatograph equipped with an ECD 800 (Fisons,
Milan, Italy) was used. The split injector was adapted to allow operation at high inlet
pressures of the helium carrier gas. For this purpose a Tescom 44-100 high pressure
regulator (Tescom Inc., Minnesota, USA) was installed. The inlet pressure was 12 bar.
Both split and splitless injections were performed. For the split injection high split
flows were used in order to minimise band broadening caused by the injection. Low
concentrations of a number of high boiling pesticides and PCBs were injected in the
splitless mode in order to meet the required detection limits. Because the column flow
of 50 Jlm columns is very low (0.4 to 0.5 ml/min), a liner with a small inside diameter
(1.5 mm) was used to speed up sample transfer. The injector was operated at a
temperature of 285°C.
The column was a CP-Sil 5 CB Column (Chrompack, Middelburg, the Netherlands)
with a length of5 m, 50 Jlm inside diameter and a film thickness of0.2Jlm.
The ECD contains a 63Ni beta particles emitting radioactive souree of 379 MBq (10
mCi) and is operated in the constant current mode. The reference current was 1 nA and
the pulse amplitude applied was 50 V. The ECD detector temperature was held at
320°C. According to the manufacturer specifications, the sensing volume of the
80 Chapter 4
detector is 450 J.ll. The pressure controller for the make-up flow in the GC was by
passed and another pressure controller was installed to allow make-up flow rates up to
a few thousand mllmin. The pressure inside the ECD cell was measured using a
pressure controller (Wallace & Tieman- Chlorator GmbH, Günzburg/Do., Germany)
installed just before the make-up gas (N2) inlet of the ECD cell. The flow in the ECD
cell was measured by connecting the ECD outlet to a bubble flow meter. The pressure
and temperature corrected flow in the detector was calculated according to:
_ Patm x Fatm x Tdet Fdet -
Tatm x Pdet (4.1)
where Pdet is the pressure in the ECD cell, F det the detector flow, T det the temperature of
the detector, Patm atmospheric pressure, Fatm the measured flow and Tatm ambient
temperature (298 K). Data acquisition was performed using a VG Xchrom data
acquisition system {VG Data Systems, Cheshire, England).
4.5 THEORY
Narrow-bore columns offer a large number of theoretica! plates per unit time.
Assuming that the only sourees of band broadening are the column and the detector,
the total band width CStot can be expressed as:
(4.2)
where CSebrom is the chromatographic standard deviation and crdet the standard deviation
of the detector. The standard deviation due to the detection volume can be described
by:
CSdet = K' x Fdet
(4.3)
where V det is the detector cell volume and 1C the profile factor. K' equals .Jï2 in case of
plug flow in the detector whereas 1C equals unity in case of exponential flow. With the
use of an ECD for high-resalution chromatography, some loss of chromatographic
High-speed narrow-bore capillary GC coupled to ECD 81
resolution will occur owing to the significant "mixing volume" [27-29] ofthis detector.
The magnitude ofthis effect will depend on the physical design ofthe detector.
To minimise detector band broadening due to the detector volume, the cell volume of
the ECD should he as small as possible. A second beneficia! factor inherent to the
reduction of the cell volume is that the flow pattem through the ECD becomes more
"plug-like" rather than "well mixed". With this flow pattem, the residence times of
analytes in the cell are significantly reduced and the contribution of the detector to
band broadening is decreased. However, the minimum cell volume is restricted to a
certain minimum because the high energy P electrous generated at the anode are not
allowed to move directly to the cathode. A gas of moderate molecular weight is
passed through the detector (the so-called quench gas) in order to efficiently
attenuate the beta radiation, thereby creating a population of positive ions and
secondary electrons throughout the active volume of the detector. These electrons,
which possess thermal energy, are responsible for the actual electron capturing
mechanism. Some distance between the anode and the cathode is required to ensure
sufficient collisions with gas molecules. To limit peak distortion due to the detector
cell volume as wellas to provide sufficient molecules for collisions, make-up gas has
to he added.
Under optima! operating conditions the 5 m column used in the experimental work
should yield approximately 115000 plates for a compound with a retention factor of
two. At an average linear velocity of 50 cm/s, this gives a chromatographic standard
deviation ( crchrorn) of 86.6 ms. lf we allow a detector band broadening equal to 10% of
the total chromatographic standard deviation, the detector broadening ( crdet) should be
less than 27.7 ms. lf it is further assumed that plug flow conditions prevail in the
detector, the minimum make-up flow required fora detector with a volume of 450 J.tl is
approximately 300 mllmin. Under exponential flow conditions in the detector cell, the
minimum make-up flow rate should he increased up to about 1000 mi/min.
Similar to the situation for a thermal conductivity detector, the electron capture
detector is generally assumed to he a concentration sensitive detector. Because higher
make-up flow rates are required to minimise detector band broadening, one expects the
minimum detectable amount (MDA) to increase drastically at higher make-up flow
rates.
82
-~ uo u
Figure 4.1
450
400
350
300
250
200
150
100
50 0 100 200 300 400 500 600 700
Column Diameter (J.tm)
Chapter4
Detection limits for an ECD as a function of the column inside diameter allowing 10% detector band broadening under exponential flow (A) and plug flow condition (B). Experimental conditions: C12, Toven = 373K, k = 5.76, ~ = 200, stationary phase: SE-30, carrier gas: He, Po= 1 bar, N = 100000.
It is shown in chapter 2 that the column inside diameter has a strong influence on the
minimum detectable amounts. In Figure 4.1, theoretica! calculations for the detection
limits of an ECD as a function of the column inside diameter are presented. To
calculate this graph, the chromatographic standard deviation ( crchrom) under optimal
separation conditions (N = 100000) was first calculated as a function of the column
inside diameter. Next the maximum allowable detector band broadening was arbitrarily
set to 10% of the chromatographic band broadening. The detector band broadening
was than calculated according to:
2 = 0.1 x 2 cr det cr chrom (4.4)
From equation (4.3) the required make-up gas flow rate could be calculated for
different flow pattems inside the detector. In Figure 4.2, the make-up flow rates for the
two extreme flow pattems are plotted as a function the column inside diameter. For
High-speed narrow-bore capillary GC coupled to ECD 83
wide-bore columns the required make-up flow is very low but it increases drastically if
the inside diameter is reduced. Assuming that the ECD behaves as a concentration
sensitive detector, the minimum detectable concentration can be calculated according
to:
(4.5)
where cg is the minimum detectable concentration for a concentration sensitive
detector,~ the noise level, se the sensitivity calculated fora concentration sensitive
detector, 0'101 the overall band width, Fdet the detector flow rate and Vmj the injected
sample volume. Forthese calculations the sensitivity and the noise level of the ECD
were assumed to be constant, i.e. independent of the make-up flow rate. The value for
the sensitivity and the noise level were 1015 Hz.ml/g, and 20 Hz respectively, as was
found in literature [30]. For Vinj = 0.5 fll, the detection limits calculated vary from 85
-c: 2000 ] A 8 1600 ........,
~ [.1. 1200 §" I
~ 800 8 Cl 400 u ~
100 200 300 400 500 600 700
Column Diameter (f.lm)
Figure 4.2 Make-up flow required to reduce detector band broadening to 10% under exponential flow (A) and plug flow conditions (B) as a function ofthe column inside diameter. Experimental
conditions: C12, Toven = 373K, k = 5.76, ~ = 200, stationary phase: SE-30, carrier gas: He, Po
= 1 bar, N = 100000.
84 Chapter4
ppb for 50 Jlm columns to 150 ppb for wide-bore (530 Jlm) columns under plug flow
conditions and increase up to 300 to 350 ppb for exponentlal flow conditions in the
detector. It is obvious from this theoretica} calculation that the MDA is favoured by the
reduction of the column inside diameter although high make-up flows are required to
minimise band broadening when narrow-bore columns are used. In the calculations
presented above, it was assumed that the noise level and the sensitivity were
independent of the make-up flow rate and the ECD behaves as a concentration
sensitive detector. Whether this is really the case must be verified experimentally.
4.6 RESULTS AND DISCUSSION
The addition of a make-up gas flow prior to the detector is a widely used method to
reduce the effective volume of a chromatographic detector. For mass flow sensitive
detectors, the addition of make-up does not affect the detection limits of the
chromatographic system. For concentration sensitive detectors, however, the detection
limits are unfavourably affected by the addition of make-up gas. Therefore, the
magnitude of the make-up flow rate should always be carefully optimised. Too high
values should be avoided because of the adverse effects on sensitivity, whereas too
low values result in band broadening and loss of resolution. In ECD detectors the
situation is even more complicated as here the make-up gas actively participates in the
detection mechanism. The make-up gas acts as a queuehing gas that couverts high
energy ~ particles emitted by the radioactive foil into thermal energy electrous that are
eventually responsible for the electron capturing process. Although the actual
mechanism of queuehing is not yet fully understood, it is known that also here there is
an optimal flow rate of queuehing gas. In the following paragraphs the influence of the
quench gas flow rate on the detector band broadening and detection limits will be
addressed subsequently.
High-speed narrow-bore capillary GC coupled to ECD 85
4.6.1 Detection band broadening
To avoid an excessive detector contribution to overall band width, high make-up flow
rates are required as is evident from Figure 4.3. This figure shows the tailing factors of
chloroform, chlorohexane and bromobenzene in a fast GC separation as a function of
the make-up flow rate. The tailing factor is the ratio of the peak width on the right si de
over the peak width on the left side at 10% of the maximum peak height. From this
tigure it is evident that indeed very high make-up flow rates are required to minimise
peak tailing. This is especially true for peaks with retention times smaller than 1
minute. The tailing factor for the later eluting compounds is smaller due to the larger
chromatographic band broadening. Even at very high make-up flow rates, peak tailing
does not completely disappear. Most likely this residual tailing is caused by the fact
that manual injection is used.
Typical make-up flow rates with an · ECD used in combination with normal bore
columns vary from 20 to 40 ml/min. Because higher make-up flow rates are required
with narrow-bore columns, the influence of this on the sensitivity and detection limits
was evaluated experimentally. This is described in the following section.
2.5 ,-------------------.
chlorobenzene
bromobenzene
500 1000 1500 2000
Make-up Flow (mi/min) Figure4.3 Tailing factor of chloroform (e), chlorobenzene (•) and bromobenzene (+)as a function ofthe ECD make-up flow.
86 Chapter4
4.6.2 Sensitivity, deteetion limits and dynamic range
In contrast to the situation with a thermal conductivity detector (TCD) where the
make-up gas is an inert diluting gas, the make-up gas actively participates in the
detection mechanism in the case of an ECD. For the TCD the increased flow of the
make-up gas does not affect the sensitivity of this concentration sensitive detector.
Because the make-up gas in the case of the ECD has an active contribution to detection
itself, it beoomes necessary to investigate the influence of the make-up gas flow rate on
the sensitivity and the noise level. In the past, the influence of the make-up gas flow on
the detector behaviour bas been studied, but only for flows in the range of 10 to 100
mllmin [31-35]. In most instances, a maximum was observed at a make-up gas flow of
about 30 mllmin. At higher flow rates, the response generally decreased exponentially
with the flow rate. Here, the sensitivity for different compounds will be evaluated at
elevated make-up flow rates.
For concentration sensitive detectors, the sensitivity can be calculated according to the
equation:
C A x Fdet s = Q
(4.6)
where ge is the sensitivity of the concentration sensitive ECD (Hz.mllg), A the area
response (mV.sec) and Q (g) the sample amount introduced onto the column. Figure
4.4 shows a plot of the measured sensitivities of some pesticides vs. the make-up flow
rate. At detector flows below 150 mllmin a decrease of sensitivity with increasing flow
rate was observed. At higher flow rates, however, the sensitivity starts to increase
drastically with increasing flow rate for all compounds tested. Why the sensitivity
increases upon increasing the make-up flow rate from 150 to 1100 mi/min is yet
unknown. Increased sensitivities at higher make-up flows were also observed by Cram
et al. [35]. At ECD make-up flow rates exceeding 1100 mllmin, sensitivity starts to
decrease again. For some compounds, such as 1-chloroheptane and 1,6-
dibromohexane, the sensitivity already decreases at a make-up flow rate of 300
mllmin. This difference between the various solutes most likely is caused by the
reaction times required for the electron capturing reactions. When the effluent of a
capillary GC column is passed through the detector, the following ionic processes
High-speed narrow-bore capillary GC coupled to ECD 87
[36-37] can occur:
rate coefficient
~+make-up gas~ p+ + e· + ~· s (4.7a)
e"+A~ A" keA (4.7b)
e·+ B ~ s· kes (4.7c)
e· + p+ ~ neutrals Re (4.7d)
A" + p + ~ neutrals Rt (4.7e)
These reactions represent: (4.7a) ionisation of the detector quench gas by beta
radiation to form positive i ons and secondary electrons; ( 4. 7b) either associative or
dissociative electron capture (EC) by the analyte A to form negative ions, A"; (4.7c)
EC by gas impurities, B; (4.7d) ion-electron recombination; and (4.7e) ion-ion
recombination. The ECD response to solute A is caused by the EC reaction ( 4. 7b ). At
low flow rates of the carrier gas, this reaction takes place on a time scale that is fast
relative to the time required for gas flow through the detection cell. The lower
Figure 4.4
-'"-'
too.----------------------------------,
ECD Flow (mi/min)
Sensitivity of the ECD for hexachlorobenzene ( • ), heptachlor (•) and dieldrin ( +) as a function ofthe make-up flow rate (assuming concentration sensitive behaviour).
88 Chapter4
sensitivity for these compounds at higher flow rates is probably due to the fact that
the reaction rate constant keA is relatively low and the residence time of the
compound in the detector cell is not sufficiently long to obtain maximum response.
In the calculations discussed above it is assumed that the ECD behaves as a
concentration sensitive detector. ECDs can however, under specific conditions of
carrier gas type, flow, pressure, temperature, voltage, etc., show a behaviour that
holds an intermediate position somewhere between either purely mass flow and
purely concentration-dependent behaviour [38,39]. Consirlering that the ECD detector
is operating as a mass flow sensitive detector, the sensitivity can be calculated
according to the equation:
A sm = - (4.8) Q
where smis the sensitivity fora detector with mass flow sensitive behaviour (Hz.s/g).
If this definition was adopted, it was found that the sensitivity was almost unaffected
by the make-up flow for flow rates in the range from 400 ml/min up to 1600 ml/min
(Figure 4.5).
From the results shown in the Figures 4.4 and 4.5 it appears that the sensitivity of the
ECD depends on the make-up gas flow rate although also regions of constant
sensitivity can occur. Good sensitivity can be obtained at elevated flow rates. At make
up gas flows between 400 and 1100 ml/min, the ECD appears to exhibit mass flow
sensitive behaviour. The explanation for this observation is not known until now. In
addition to the make-up gas flow rate, other instrumental parameters, i.e. the detector
temperature, purity of the gases, the detector regime (frequency and width of pulses)
[30, 32, 40], etc., can all affect the response of the detector. It is possible that these
parameters become more important at elevated make-up gas flows. At these elevated
flows also secondary effects can become important. For example, the detector
temperature can be influenced by the use ofhigh make-up gas flow rates [41-44]. Also
the presence of oxygen or other trace impurities in the make-up and carrier gas passing
through the detector can change the detector response characteristics [39]. It is also
known that the response depends on the flow pattem in the detector cell [44]. This
could be different at higher flow rates.
High-speed na"ow-bore capillary GC coupled to ECD 89
ECD Flow (mi/min) Figure 4.5 Sensitivity of the ECD for hexachlorobenzene (e) heptachlor (•) and dieldrin ("") as a function ofthe make-up flow rate (assuming mass flow sensitive behaviour).
The ultimate detection limits in high-speed GC-ECD are not only a function of the
sensitivity, but also of the noise level. Hence it is also interesting to investigate the
influence of the make-up flow on this parameter. A gradual decrease in the noise level
and base frequency was observed when the make-up flow rate was increased to 50
ml/min. Both base frequency and noise amplitude were virtually constant in the make
up gas flow range of 50 to 1000 ml/min. From the base frequency experiments it
appears that the number of electrons availabie for capturing reactions is almost
constant in the range from 50 to 1000 mi/min of make-up gas. At flow rates exceeding
1000 mi/min, both noise leveland base frequency increased sharply.
Although there is no exact explanation for the sensitivity behaviour of the ECD
detector, it is evident from the results shown in the Figures 4.3, 4.4 and 4.5 that the
ECD detector is compatible with 50 J.tm I.D. narrow-bore capillary GC columns. The
high make-up gas flow rates required to eliminate detector band broadening have no
adverse effects on the detection limits. Good detection limits can be achieved despite
the high make-up flow rates.
For narrow-bore columns, the detection limits forsome pesticides (hexachlorobenzene,
heptachlor and dieldrin) were approximately 100 fg in the range of make-up gas flows
90 Chapter4
of 400 to 1000 mllmin. The minimum detectable concentradon with high~speed GC is
0.2 ppb at an injection volume of 0.5 }.11 (splitless injection}. The detector response was
found to exhibit linear behaviour up to a few hundred pg at these elevated make~up gas
flowrates.
4.6.3 Applications
In Figure 4.6, a fast separation of a test mixture is presented. The separation of the 8
compounds is accomplished in less than 20 seconds.
In Figure 4.7 the separation of the PCB standard Arochlor 1242 is shown. The
injection was performed in the split mode.
Figure 4.8 shows the separation of PCBs extracted from transfarmer oil. The clean-up
ofthe sample was done according to the procedure publisbed by Sandra et al. [45].
0 2 4 6 8 10 12 14 16 18
Time (s)
Figure4.6 High-speed chromatagram of a test mixture. GC column: CP-Sil 5, L = 5 m, LD. = 50 11m, dr = 0.2 f.1m. Experimental conditions: ll0°C, Pi = 20 bar, split flow is extremely high (> 1000 mVmin), sample introduction: headspace 20 }ll. ECD make-up flow = 900 mVmin. Components in order of elution: chloroform, 1-iodopropane, 1,1,2-trichloroethane, 1-iodobutne, 1,1 ,2-trichloro-propane, tribromomethane, 1,1 ,2,2-tetrachloroethane, diiodom .. thane.
High-speed narrow-bore capillary GC coupled to ECD
7 8 9
Figure 4.7
10 11
Time(min)
91
High-speed chromatogram of a PCB mixture (Arochlor 1242). GC column: CP-Sil5, L = 5 m, LD. = 50 r.tm, dr = 0.2 r.tm. Experimental conditions: 150°C -+ 20°C/min -+ 280°C, Pi= 12 bar, split flow= 400 mUmin. ECD make-up flow= 400 mUmin.
6 7
Figure 4.8
8 9
Time (min)
High-speed chromatogram of a PCB extract (Arochlor 1260) oftransformer oil. GC column: CP-Sil 5, L = 5 m, LD. = 50 r.tm, dr = 0.2 r.tm. Experimental conditions: 50°C (3 min) ballistically heated to 280°C, Pi= 12 bar, splitless time= 3 min, Vinj = 0.3 111. ECD make-up flow = 400 mUmin.
92 Chapter4
Here the injection mode used was splitless. The approximate concentratien of the
PCBs in the oil was 1 ppm. The separation of this complex mixture was achieved in
approximately I 0 minutes. Obtaining similar resolution on a normal-bore column
would take approximately 45 minutes.
In Figure 4.9, the separation ofan SFE extract ofthe PCB standard reference material
1939 of the N.B.S. (Gaithersburg, USA) is shown. The injection mode was splitless.
Before the elution of the PCBs starts, a significant amount of co-extracted compounds
was observed.
0 5
Figure 4.9
10 15
Time (min)
High-speed chromatogram of an SPE-extract of PCBs from sediment (N.B.S. Standard Reference material 1939). GC column: CP-Sil 5, L = 5 m, LD. = 50 JJ.m, dr = 0.2 J.lm. Experimental conditions: 40°C (4 min)~ 20°C/min ~ 275°C, pi= 12 bar, splitless time= 3 min, V inj = 0.3 J.1l. ECD make-up flow = 400 ml/min.
In Figure 4.10 a splitless injection of some pesticides is shown. The concentratien of
the individual pesticides ranged between 10 and 100 ppb. From the Figures 4.8 to 4.10
it is clear that the combination of narrow-bore columns, splitless injection and electron
capture detection results in excellent concentratien detection limits. Unfortunately, due
to the fairly long residence times of the sample in the hot injector liner, splitless
High-speed narrow-bore capillary GC coupled to ECD
Q.l
ill ä ~
"t:! ·Ë ::r: :§
~ ~
j ~ .9
.d
~ Q. Q.l .d
4 5 6
Figure 4.10
·ê ·ê :9 ] t :ä
"t:! 7il = ·c ] Q.l
7
~ ~ ..:.: = ·c
"t:!
5
8
Time (min)
93
High-speed chromatogram of a pesticide test mixture. GC column: CP-Sil 5, L = 5 m, LD. = 50 !Jm, dr= 0.2 IJm. Experimental conditions: 50°C (3 min) ballistically heated to 280°C, Pi= 12 bar, splitless time= 3 min, Vinj = 0.3 IJL ECD make-up flow rate = 400 ml/min.
injection can give rise to degradation of thermally unstable components. In Figure
4.1 0, a relatively high concentration of degradation products of endrin, endrin
aldehyde and endrin ketone is observed. A more in-depth discussion of the
possibilities and limitations of splitless injections in narrow-bore capillary GC was
presented in chapter 3.
4.7 CONCLUSIONS
The combination of 50 J.l.m LD. columns with electron capture detection enables high
speed analysis. Although high make-up flow rates are required in order to minimise
peak tailing, very good sensitivity could he obtained at these elevated flow rates.
Detection limits of 0.1 pg were obtained, corresponding to a minimum detectable
concentration of 0.2 ppb for an injection volume of 0.5 J.l.l. The detector exhibits linear
behaviour up to a few hundred picogram. The GC-ECD system described in this work
provided reliable high-speed analysis of trace quantities for industrial and
94 Chapter4
environmental applications.
4.8 REFERENCES
I. F.J. Yang, S.P. Cram, J. High Resolut. Chromatogr. Chromatogr. Comm., 5 (1982)
454.
2. P.A. Leclercq, in "Quantitative Analysis by Gas Chromatography", J. Novak: (Ed.),
Marcel Dekker Inc., New York, 1987, chapter 8.
3. J.C. Sternberg, in "Advances in Chromatography", J.C. Giddings, RA. Keller
(Eds.),Vol. 2, Marcel Dekker Inc., New York, 1966.
4. R. Annino, Advances in Chromatography, 15 (1977) 33.
5. A. Savitsky, MJ.E. Golay, Anal. Chem., 36 (1964) 1627.
6. S. Frans, Chromatography Review, 14 (1988) 5.
7. RB. Lam, RC. Wieboldt, T.L. Isenhour, Anal. Chem., 53 (1981) 889A.
8. G. Horlick, Anal. Chem., 44 (1972) 943.
9. T.A. Maldecker, J.E. Davis, L.B. Bogers, Anal. Chem., 46 (1974) 637.
10. AJ.C. Nicholson, J. Chem. Soc., Faraday Trans. I, 79 (1982) 2183.
11. Schutjes, Ph.D. Thesis, Eindhoven University of Technology, the Netherlands,l983,
138.
12. A.J.J. van Es, C.A. Cramers, lA. Rijks, J. High Resolut. Chromatogr. Chromatogr.
Comm., 12 (1988) 862.
13. J.N.Driscoll, in "Detectors in Capillary Chromatochraphy", H.H. Hili, D.G. McMinn
(Eds.), Chemical Analysis Series, Vol. 121, J. Wiley & Sons Inc., New York, 1992, p.
65.
14. J.N.Driscoll, in "Detectors in Capillary Chromatochraphy", H.H. Hill, D.G. McMinn
(Eds.), Chemical Analysis Series, Vol. 121, J. Wiley & Sons Inc., New York, 1992, p.
69.
15. D.W. Later, M.L. Lee, B.W. Wilson, Anal. Chem., 54 (1982) 117.
16. W.A. Aue, S. Kapila, J. Chromatogr. Sci., 11 (1973) 225.
17. E.D. Pellizzari, J. Chromatogr., 98 (1974) 323.
18. J.E. Lovelock, J. Chromatogr., 99 (1974) 3.
19. J.E. Lovelock, A.J. Watsson, J. Chromatogr., 158 (1978) 123.
20. C.F. Poole, J. High Resolut. Chromatogr. Chromatogr. Comm., 5 (1982) 454.
21. B. Dressler, in "Selective Gas Chromatographic Detectors", Elsevier, New York,
1986.
High-speed narrow-bore capillary GC coupled to ECD 95
22. A. Zlatkis, C.F. Poole (Eds.), in "Electron Capture. Theory and Practice in Chromatography", J. Chromatogr. Libr., Vol20, Elsevier, Amsterdam, 1981.
23. H. Ke, S.P. Levine, RF. Mouradian, R Berkley, Am. Ind. Hyg. Assoc. J., 53(2)
(1992) 130-137.
24. C.P.M. Schutjes, E.A. Vermeer, G.J. Scherpenzeel, RW. Bally, C.A. Cramers, J. Chromatogr., 289 (1984) 157.
25. C.A. Cramers, G.J. Scherpenzeel, P.A. Lec1ercq, J. Chromatogr., 203 (1981) 207.
26. P.A. Leclercq, G.J. Scherpenzeel, E.A.A. Vermeer, C.A. Cramers, J. Chromatogr.,
241 (1982) 61.
27. D.G. Peters, J.M. Hayes, G.M. Hieftje, in "Chemical Separations and Measurements",
Saunders, Philadelphia, PA, 1974.
28. G. Wells, R Simon, J. High Resolut. Chromatogr. Chromatogr. Comm., 6 (1983) 427.
29. G. Wells, J. High Resolut. Chromatogr. Chromatogr. Comm., 6 (1983) 651.
30. T.H.M. Noij, Ph.D. Thesis, Eindhoven University of Technology, the Netherlands,
1988, 21.
31. P. Devaux, G. Guiochon, J. Gas Chromatogr., 5 (1967), 341.
32. P. Devaux, G. Guiochon, J. Chromatogr. Sci., 7 (1969) 561.
33. J.J. Franken, H.L. Vader, Chromatographia 5 (1973) 22.
34. P. Rotock.i, B. Drozdowicz, J. Chromatogr., 446 (1988) 329.
35. F.J. Yang, S.P. Cram, J. High Resolut. Chromatogr. Chromatogr. Comm., 2 (1979)
487.
36. P.L. Gobby, E.P. Grimsrud, S.W. Warden, Anal. Chem., 52 (1980) 473.
37. W.E. Wentworth, E.C.M.Chen, in "Electron Capture Detector. Theory and Practice in
Chromatography", A. Zlatkis, C.F. Poole (Eds.), J. Chromatogr. Libr., Vol 20,
Elsevier, Amsterdam, 1981.
38. RE. Kaiser, RL Rieder, in "Electron Capture Detector. Theory and Practice in Chromatography", A. Zlatkis, C.F. Poole (Eds.), J. Chromatogr. Libr., Vol 20,
Elsevier, Amsterdam, 1981, p. 120.
39. J. Sevcik, in "Detectors in Gas Chromatography", J. Chromatogr. Libr., Vol. 4,
Elsevier, Amsterdam, p. 72.
40. R.J. Maggs, P.L. Joynes, A.J. Davies, J.E. Lovelock, Anal. Chem., 43 (1971) 1966.
41. M. Scolnick, J. Chromatogr. Sci., 7 (1969) 300.
42. K. Peltonen, LC-GC Int., 3 (1990) 52.
43. C.-Y. Chen, Y.-C. Ling, Chromatographia, 33 (1992) 272.
44. M. Cigánek, M. Dressler, V. Lang, J. Chromatogr., 668 (1994) 441.
45. P. Sandra, F. David, G. Redant, B. Denoulet, J. High Resolut. Chromatogr.
Chromatogr. Comm., 11 (1988) 840.
96
Chapter 5
High-speed narrow-bore capillary gas
chrontatography coupled to various mass
spectrometric detection methods
SUMMARY
The on-line hyphenation of chromatography to mass speetrometry has become one of
the most efficient and versafile systems for the study and identification of organic
compounds in complex sample matrices. In this chapter, several mass spectrometers,
including the ion trap, a sector instrument, the rejlectron time-of-jlight MS and the
orthogonal acceleration time-of-jlight MS, are evaluated as defection devices for
Jast capillary gas chromatographic separations. The performance of these mass
analysers is are compared mutually and to other mass spectrometers described in
literature. Special emphasis is put on the acquisition rates, defection limits, mass
spectrometric resolution and quality of mass spectra. It wil! be demonstrated that
depending on the type of mass spectrometer, a campromise has to be made between
these parameters. In our opinion, the orthogonal acceleration time-of-jlight MS
currently offers the best possibilities as defection device for high-speed separations
although several modifications are still required toforther imprave the performance
of the instrument.
98 Chapter 5
S.l INTRODUCTION
A chromatagram provides information regarding the sample complexity (number of
peaks), the quantity (peak height or peak area) and at least in principle, also on the
identity (retention parameter) of the components in a mixture. Identification based
solely on retention is generally considered very suspect, even for simple mixtures.
Spectroscopie techniques provide a rich souree of qualitative information from
which the identity of substances may be elucidated with a reasonable degree of
certainty. Spectroscopie techniques, however, have two practical limitations. It is
often difficult to extract quantitative information from their signals and, for reliable
information, they require pure samples. Chromatographic and spectroscopie
techniques thus provide complementary information about the identity of the
components and their concentration in a sample.
The on-line combination of gas chromatography with mass speetrometry is without
any doubt the most powerfut hyphenated technique for the separation of unknown
samples with subsequent identification and quantitation of the constituents. Mass
speetrometry is noteworthy among modem structural elucidation tools because the
information gathered is of a chemica} nature and most substances produce unique or
distinctive fragmentation patterns allowing identification of most substances from
their spectra. The coupling of a chromatograph with a two-dimensional (mass, ion
abundance) detector that can be highly selective provides a vast amount of
information about the sample.
Continuing improvements and developments in separation sciences put higher
demands to the instrumentation. Especially in combination with narrow-bore
capillaries, the detector requirements are being pushed to their limits. First of all, the
sample amount that can be introduced onto the column is very low. For this reason,
the detection device has to be very sensitive. Additionally, narrow-bore columns
produce very narrow peaks. Consequently, the detector must have a very low time
constant [1]. In the particular case ofmass spectrometry, examination ofthe scanning
capabilities of the mass spectrometers is necessary to assess the feasibility of
obtaining adequate mass speetral data. The scan speed should be sufficiently high to
minimise changes of the sample concentration in the ion souree during the time a
High-speed narrow-bore capillary GC coupled to various mass spectrometers 99
spectrum is acquired. In this way distortion of relative peak intensities in a given
spectrum is minimised. Another reason why high scan repetition rates are important
is the requirement to be able to reconstruct the chromatogram accurately from the
consecutively recorded spectra [2-5]. The higher the frequency of spectrum
acquisition, the greater the number of data points that is available to define the
chromatographic profile. At least 5-10 data points (spectra) per peak are required for
accurate reconstruction ofthe peak profile. In contrast to the situation in the full scan
mode, in the selected ion monitoring mode only a few mass channels are monitored.
Due to its high repetition rate this approach can provide a large number of points per
chromatographic peak. Unfortunately, it is useful solely for analyses in which only
the ion currents of a few pre-selected masses are of interest ("target compound
analysis").
The utility of full scan mass speetral analysis requires that the rate of speetral
generation is sufficient so as not to lose information along the chromatographic axis.
Unfortunately, however, the scan speed is limited by the physicallaws goveming the
mass analysis process, and the response time of the detector processing components
[1,6]. Evidently, the faster a mass spectrometer (MS) is scanning, the less time is
available for assessment of each mass peak in the spectrum. Hence the sensitivity
and the accuracy of ion current measurements will he reduced at higher scanning
rates of the MS. Moreover, as for all spectroscopie techniques, high sensitivity can
only be obtained at the expense of mass spectrometric resolving power (selectivity)
and vice versa. As a consequence, depending on the type of the MS used, a
compromise between the scan speed, mass resolution and sensitivity has to be made.
In this chapter several mass spectrometers, including the ion trap, a sector
instrument, the reflectron time-of-flight and the orthogonal acceleration time-of
flight mass analyser are evaluated as detectors for high-speed narrow-bore
separations. Special emphasis will be putto the scan speed, detection limits, quality
of spectra and mass spectrometric resolution. The performance of the GC/MS
systems will be demonstraled by various applications. The performance of these
different mass spectrometers is compared mutually as well as with other mass
spectrometers described in literature. In this way, an overview is given of the present
state-of-the-art ofGC/MS systems for fast gas chromatography.
100 Chapter 5
5.2 ION TRAP MASS ANALYSER
5.2.1 Introduetion
The basic principles of operation of the ion trap are perhaps best considered in
relation with those of the rather more familiar quadrupale mass filter. The
quadrupale mass analyser consists of four rod electrades which are connected to
power supplies such that opposite pairs of electrades are coupled together. Ions
emitted from the ion souree pass into the mass analyser. Depending on the mass-to
charge ratio and the instromental parameters of the MS, ions have either a stabie
trajectory and pass through the analyser to the detector while other ions follow
unstable trajectories and collide with the electrodes. By rotating the rod structure of
the quadrupale instrument, one pair of electrades will form a surface rather like a
ring donut while the two other rods form separated convex surfaces. Therefore the
ion trap is a three electrode device. A section of the ion trap is shown in Figure 5.1.
Generally, the ion trap is contigured such that the two end-cap electrades are
connected together and held at ground potential while non·zero potentials are applied
only to the ring electrode. As with a quadrupale analyser, ions within the electric
Figure 5.1
lens ----- UuJ filament
~[?=J elw-~ =0 0
top end-cap electrode
ring electrode
~\\~ bottomend-cap ~· __ . 1 \_____j electrode
electron multiplier
Schematic ofthe ion trap.
High-speed narrow-bore capillary GC coupled to various mass spectrometers 101
field will have either stabie trajectories and remain trapped or will he unstable and he
ejected from the trap. All i ons in the ion trap with m/z values greater than a minimum
value will have stabie trajectodes in the ion cell. As the RF voltage is scanned
upward, ions of increasing m/z values will become unstable and are ejected through
the end-caps, where they are detected. This means that the ions are trapped in the ion
trap until they are ejected according totheir mass-to-charge ratio. As a consequence,
most of the i ons ( 50 %) detected so that higher sensitivities can be obtained. In this
respect, quadrupole instruments suffer from an important disadvantage being the
limited sensitivity by measuring ion currents only during a fraction of the scan time
(scan time/mass range). Ion trap mass spectrometers appear highly promising for
interfacing with narrow-bore columns as these instruments are inherently more
sensitive than quadrupole instruments and therefore more compatible with the minute
sample quantities eluting from a narrow-bore column.
5.2.2 Experimental
The GC-MS system used consisted of a Varian 3400 gas chromatograph (Varian,
Walnut Creek, CA, USA) and anion trapMS (Varian Saturn II MS). The GC was
equipped with an 8100 autosampler and a 1077 split/splitless injector. To allow
operation at high inlet pressures, a Tescom 44-1100 high pressure regulator (Tescom
Inc., Minnesota, USA) and a custom-made digital pressure indicator were installed in
the carrier gas line. The system was operated in the pressure controlled mode. The
injector and the GC/MS transfer line temperatures were operated at 275°C.
Instrumental control and data processing were performed using the V arian Satum
software which ran on a Compaq PC (386-20e).
The fundamental principles and developments in ion trap MS have been reviewed by
Todd in 1991 [7]. After ionisation, the ions are trapped in the ion trap until they are
ejected according to their mass-to-charge ratios. In the earlier versions of the ion trap
spatial charge interactions caused by high ion concentrations in the trap resulted in
reduced linearity and loss in mass resolution, especially at high concentrations of the
compounds. A first way to reduce space charge perturbation is to use a controlled
scan function [8]: the Automatic Gain Control (AGC) in the electron impact (EI)
mode and the Automatic Reaction Control (ARC) in the chemical ionisation (Cl)
102 Chapter5
mode (Figure 5.2). With these scan functions, the number of ionsin the trap is kept
under controL An estimate of the number of i ons fonned in the trap is provided by a
short prescan that consists of a 0.1 msec ionisation period. The duration of the
variabie ionisation time (VIT) in the EI mode or the reaction gas ionisation period in
the CI mode used for the actual scan is now automatically detennined on the basis of
the ion concentradon in proportion to the user-selected target value (TV). In doing so
not only the space-charge perturbation is minimised but also the signal-to-noise ratio
is increased approximately 50 times. The analytical scan in the EI mode (AGC-scan)
and the CImode (ARC-scan) are shown in Figure 5.2A and 5.2B, respectively.
Another impravement to the original ion trap design was introduced by Weber-
A analytica! scan prescan < >~------------------~
B prescan analytica! scan E )
"0
ë .g 8. '.;:J = = § 0 0 ~~ ·-= 'i ~ <I)
<I) ·a ·a ·a .5! .Sl ~ OI)
Figure 5.2 Controlled scan functions with the ion trap: Automatic Gaio Control (A) for the EI mode and Automatic Reaction Control (B) for the Cl mode.
High-speed narrow-bore capillary GC coupled to various mass spectrometers 103
Grabau [9-10] by making use of an axial modulation voltage (7V). An extra voltage
was applied to the end-caps ofthe ion trap. In this way, the ion cloud in the souree is
enlarged, resulting in reduced charge density. As a consequence, the interaction
between ions and molecules is reduced resulting in better spectra and higher mass
resolution. Despite these various improvements, ion trap mass spectra sametimes still
can be different from quadrupale spectra, especially when larger sample amounts
elute from the column. This is due to the long trapping times (in comparison with ion
souree residence times in "normal'' mass spectrometers), which cause various ion
molecular and non-unimolecular reactions.
A3 m x 50 J.liD I.D. DB-1 column (Fisons, J&W Scientific, Folsom, USA) with 0.17
11m film thickness was used to determine the detection limits in the EI mode. A
shorter DB-1 column (1 m x 50 J.lffi) with a film thickness of0.05 J.lm was used in the
Cl mode. In all experiments an injection volume of 1 !J.l and a split ratio of about
11500 was used. The column temperature was 90°C.
The scan range was 40 to 180 Da in the EI mode and 50 to 200 Da in the Cl mode.
Eight scans per second were obtained. Because EI is a "hard" ionisation process, ions
will have a higher energy so that fragmentation during ion storage is highly probable.
As a consequence, the relative abundances of ions with a high mass-to-charge ratio
will be affected injurious compared to the relative abundance of ions with a lower
ratio. To reduce this effect, the analytica! scan in the EI mode is subdivided in four
segments (see Figure 5.2A). In this way the storage time of the ions in the trap is
shorter and fragmentation or ion-molecule interactions are less likely to occur.
Normally, the four segments correspond to the following four portions of the mass
range: 10 to 99 Da, 100 to 249 Da, 250 to 399 Da, and 400 to 650 Da. For our
experiments, scanning the mass range from 40 to 180 Da in the EI mode, the number
of scans was maximised by selecting the mass of the end of the frrst scan segment
higher than the highest mass of interest (200 Da). Than, only one analytica! scan
segment is needed to complete the scan. With the Saturn software the maximum
number of scans is limited to 10 per second. Forsome applications, the entire mass
range (up to 650 Da) was scanned. Than, the four segments of the AGC scan
function were maintained. In this situation, the number of scans is reduced to about 3
per second.
104 Chapter5
The influence of the helium flow rate on the mass resolution was evaluated by
studying the width of the ion peaks from the calibration gas FC~43. The effect of the
helium flow on the sensitivity was evaluated by injecting a mixture of alkanes from
n~C13 to n~C 16 diluted in hexane at different flow rates. The separations were
performed with the 1 m x 50 11m DB-1 column at a temperature of l25°C.
5.2.3 Results and discussion
5.2.3.1 Instromental parameters
Optimisation of the instrument control parameters was performed on the basis of
measured peak areas and noise levels at various filament emission current (FEC) and
TV settings. In the EI mode, both the signa} and noise increased with increasing
FEC. The best signal-to-noise ratio was obtained at a FEC of 30 11A and an AGC-TV
of30000.
In the Cl mode, the FEC was set at 10 11A. The peak heights were virtually
independent of the target values while the lowest noise level was observed for high
ARC-target values (60000).
5.2.3.2 Influence of the helium background on mass resolution and sensitivity
As opposed to the situation in quadrupole mass spectrometry, with ion trap detectors
maximum mass resolution is obtained at relatively high helium background
pressures. According to Stafford et al. [11], carrier gas ions tend to stabilise the
trajectories of the sample ions in the ion trap, which results in an increased mass
resolution and an improved sensitivity.
For 50 J.lm I.D. GC columns, typical (optimum) gas flows are approximately 0.5
mllmin (measured under ambient conditions). To test whether this relatively low
helium flow suffices for stabilisation of the trapped i ons, a series of experiments was
performed at different helium flows. Under each of these conditions the mass
resolution and the response were measured. The experimental results of these studies
are shown in the Figures 5.3 and 5.4.
High-speed narrow-bore capillary GC coupled to various mass spectrometers 105
r--------~-----------,
15001
l 614
IOOO·j: ~~ soo,~ ·~--~,__~:_j
J~~~~~-~ 264
' mj
0 I 2 J 4 5
Column flow (mllmin)
Figure 5.3 Mass spectrometric resolution obtained with the ion trap for ions of the calibration gas FC-43 as a function ofthe column flow rate.
r··· : ' ·I
0 2 3 4 5 6 7 8
Column flow (mllmin)
Figure 5.4 Response (area counts) ofn-C 13 (•)and n..C 16 (•)obtained with the ion trap as a function ofthe column flow rate.
106 Chapter 5
The flow rates under atmospheric conditions (F ""'""m) are calculated from:
_ _1t_ 2 (·273.15+Tatm J r~) Fout atm - de U ' 4 273.15 + Toven dPatm
(5.1)
where d, is the column inside diameter, TI and p; are the average linear velocity and
the inlet pressure of the carrier gas, respectively. T0 ,,0 is the ternperature of the
column and T"m is the temperature of the Iabaratory environment.
A minimum ion peak width ( corresponding to maximurn mass resolution) was
observed at a flow rate of 0.5 milmin which exactly equals the minimum of the H-u
curve for the 50 flm column used in the present study. At lower flow rates, the widths
ofthe ion peaks increased rapidly most likely because there are notsuftkient heliurn
i ons for stabilisation of the trajeetori es of the analyte i ons. At higher flow rates, and
hence higher pressores in the trap, the ion trajectories are disturbed by collisions
between ions and neutral molecules. This causes the mass resolution to decrease
again.
To evaluate the influence of the helium background pressure on the sensitivity, the
peak areas of a nurnber of alkanes was rneasured as a function of the column flow
rate. At flow rates below I rnl/rnin, the response increased sharply. The exact reason
for this observation is yet unknown. The deercase in response observed at higher
flow rates is most likely due to less efficient trapping at higher callision rates with
helium molecules. It was also found that the noise levels rernain constant for helium
flows of up to 3 mi/min, but increased sharply at higher flow rates. It is clear that at
higher flow rates, ion trajectories are distorbed by collisions between ions and
neutral molecules, resulting in reduced rnass resolution and signal-to-noise ratios.
Unfortunately, we were not able to rneasure the exact influence of the column flow
on the pressure in the souree and to correlate the mass resolution and signal-to-noise
ratio with the ion souree pressure. It is, however, evident that higher column flows
result in higher ion trap pressures.
lligh-speed narrow-bore capil/ary GC coupled to various mass spectrometers 107
5.2.3.3 Defection limits and dynamic range
The deleetion limits were calculated aceording to equation (2.27) and are tabulated
for several compounds in bath the EI and Cl mode in Table 5.1. As can be seen from
this table, the deleetion limits are approximately I pg in the EI mode. Under
chemica! ionisation conditions with C~ as the reaction gas the deleetion limits are
approximately 5 times higher. For sensitive deleetion under Cl conditions, the
concentration of the reagent gas should be carefully optimised. If the concentration
ofthe reagent gas is too low, reduced sensitivity will be observed.
Quantitication is greatly simplified if the detection system exhibits a linear response
over a wide range of concentrations. To evaluate the linearity of the GC-ion trap
combination, various samples with increasing concentrations were injected. At
injected amounts (on the column) exceeding 2 to 3 ng, ill-shaped, Jeading peaks were
observed. Th is peak shape is clearly caused by overloading of the chromatographic
column. Therefore it can be concluded that the working range, i.e. the ratio of the
amount where overloading of the column starts to occur over the minimum
detectable amount, of the narrow-bore GC-MS combination was approximately
2.5x l 03 in the EI mode and is about 5 times smaller in the Cl mode. Due to the high
sensitivity of the ion trap a reasanabie working range is preserved despite the low
sample capacity ofthe narrow-bore columns.
Table 5.1. Deteetion limits (pg) and woricing range in the EI and Cl mode for the Varian ion trap.
Analyte El Mode Cl mode (CH.)
decane 0.7 5.0
2,2,3,3-tetramethylhexane 0.8 4.8
l-chlorooctane 1.4 4.5 2-nonanone 1.5 5.3 undecane 1.4 6.6
2,4-dimethylphenol 1.6 6.9
tertiary butylbenzene 1.5
cymenc 1.4
108 Chapter 5
5.2.3.4 Quality ofmass spectra
In the first generation of ion trap detectors, introduced a decade ago, the mass spectra
strongly depended on the amount of solute introduced onto the column. Large ·
differences in the relative ion abundances were even observed in subsequently
acquired mass spectra from one peak. The influence of the operational conditions on
the mass spectra obtained with an ion trap mass spectrometer has been the subject of
a number of research papers [12-15). The main reason for the dislortion of the
spectra obtained with the earlier instruments was the occurrence of space charging
during the relatively long residence times of the i ons in the ion trap. In this respect,
the use of narrow-bore columns in combination with ion trap mass spectrometric
deleetion might be advantageous because of the minute sample quantities that elute
from the column. Because the concentration of solute molecules in the souree is
smaller, ion-molecule interactions, resulting in distorted spectra, are less likely to
occur.
For a wide variety of components including esters, phenols, ketones, chlorinated
al kanes, etc., the variation of the FEC and the TV in our experiments were found to
cause only smal! differences in the mass spectra. For most of the solutes, library
searchable spectra were obtained. For a limited number of substances, reproducible
spectra but significantly different from quadrupole spectra, were observed. Due to
the relatively long storage time of the ions in the ion source, chemica! interactions
between ions and molecules, or extensive ion fragmentation can occur. In this way
new ions are formed which resulted in spectra that differ somewhat from quadrupale
and magnetic sector spectra. Although several modifications were introduced, like
the axial modulation voltage and controlled scan functions, and despite the use of
only very smal! sample quantities in narrow-bore GC, the occurrence of chemica!
interactions could not be completely avoided.
5.2.3.5 Applications
The combination of narrow-bore columns with ion trap mass spectramettic deleetion
is an attractive hyphenated technique for a large number of high-speed applications.
In Figure 5.5 an example of a high-speed separation of the PCB mixture Arochlor
1248 is illustrated. The separation ofthis complex mixture is achieved in less than 4
High-speed narrow-bore capillary GC coupled to various mass spectrometers 109
4 4 4
3 4 4
3 4 4
4 4 4
5 5
3 44 4 6
200 250 300 350 400 450 500 550 Scan nurnber 1:14 1:32 1:51 2:09 2:28 2:46 3:05 3:23 min
Figure 5.5 High-speed chromatogram (TIC) of a PCB mixture (Arochlor 1248) in CHC13 with ion trap detection. GC column: DB-I, L = 7 m, LD.= 50 f.llTI, dr= 0.05 f!m. GC conditions: 150°C ~ 25°C/min ~ 300°C, Pï = 15 bar. MS conditions: EI (70 eV), scan speed= 0.35 s/scan, mass range = 50-650 Da.
minutes. A similar separation on a conventional 320 f..1JI1 column would take
approximately 30 to 40 minutes. With mass chromatography, the clusters of PCBs
containing equal numbers of chlorine atoms can be readily identitied. In the tigure
the number of chlorines of the biphenyl moiety is shown above the major peaks.
An example of a combination of a high-speed GC analysis and chemica! ionisation
with C~ asthereagent gas is presented in Figure 5.6. This tigure shows the analysis
of a naphtha reference mixture (no. 4-8265, Supelco, Bellafonte, USA). The sample
contains a mixture of alkanes ranging from C-3 to C-12 including a series of
aromatic constituents.
In Figure 5. 7 an example of the high-speed separation of a liquid crystal mixture is
presented. The analysis of this mixture with a 0.32 mm LD. column (generating the
same plate number) takes approximately 1 hour. The larger peaks correspond to
approximately 0.5 ng. For the later eluting components peak overloading and
broadening already starts to occur, illustrating the low sample capacity of narrow-
110
Figure 5.6
200 400 0:26 0:50
600 800 1000 Scan number 1:15 1:40 2:05 min
Chapter 5
High-speed chromatagram (TIC) of naphtha standard reference mixture (no. 4-8265) with ion trap detection. GC column: DB-I, L = 7 m, I.D. = 50 J.t.m, dr= 0.05 J.t.m. GC conditions: 35°C (0.5 min)--)- 40°C/min--)- 200°C, Pi = 20 bar. MS conditions: Cl (Cf!t), scan speed= 0.125 slscan, mass range 50-230 Da.
Figure 5.7
*
100 200 300 400 500 600 0:36 1:11 1:46 2:20 2:55 3:30
Scannumber min
High-speed chromatagram (TIC) of a mixture of liquid crystals with ion trap detection. GC column: DB-1, L = 1 m, LD.= 50 f.tm, dr= 0.05 J.I.ID. GC conditions: l40°C--)- 40°C/min-)-3000C, Pi= 5 bar. MS conditions: EI (70 eV), scan speed= 0.35 slscan, mass range= 50-650 Da.
High-speed narrow-bore capillary GC coupled to various mass spectrometers 111
bore columns. In Figure 5.8A, the mass spectrum ofpeak 5 is presented. Figure 5.8B
shows the library spectrum of this component (measured with a quadrupole
instrument). As can beseen from this figure, the experimental mass spectrum closely
resembles the library spectrum.
A 100
138
,..-._ 80 èf?. -.È'
""' 60 110 s= 0 Cl) 0
d ..... Cl)
40 i> ·~
ë 20
55 81
304 0
50 100 150 200 250 300
mlz
B 100 138
- 80 èf?. '-"
~ ·;;; 60
~ ...... Cl)
40 i> ·.g 110 -~ 83 20 55
1.11 .l .I 304 0
50 100 150 200 250 300
mlz
Figure 5.8 Comparison of the experimental mass spectrum (A) of peak 5 in Figure 5.6 and the library mass spectrum (B) ofthis component.
112 Chapter 5
5.2.4 Conclusions
The combination of 50 j.tm I.D. columns with ion trap mass spectrometric detection
enables fast and sensitive gas chromatographic separations in the range of minutes
with detection limits in the low pg range. The dètector exhibits linear behaviour over
the entire working range. The spectra were not always fully comparable to library
spectra due to self-chemical ionisation or fragmentation during ion storage. To
obtain good sensitivity and mass resolution, helium carrier gas is required to stabilise
the ion trajectories. Although the optimal column flow for maximum separation
efficiency of these narrow-bore columns is relatively low (about 0.5 mi/min), the
performance of the ion trap in terms of mass resolution and sensitivity under these
conditions is very good.
High-speed narrow-bore capillary GC coupled to various mass spectrometers 113
5.3 FAST SCANNING DOUBLE FOCUSING SECTOR INSTRUMENT
5.3.1 Introduetion
Magnetic sector instruments have several advantages over quadrupole and ion trap
devices, including higher mass resolution and greater sensitivity. Quadrupole
instruments are limited in scan speed ofthe low energy ofions [6] required to obtain
adequate mass separation and good sensitivity. Magnetic instruments use ions of
higher energy and are not limited in this respect.
In this section, the use of 50 J..l.m inside diameter GC columns in combination with a
modem fast scanning sector instrument is reported. We evaluated the scan speed in
the full scan mode, and in the selected ion monitoring mode by using both voltage
and magnetic field switching. Conventional magnetic sector instruments are limited
to 3 to 5 scans per second. However, with the instrument used, higher scan speeds
were anticipated, because the instrument had a smaller, faster scanning, fully
laminated, low inductance magnet. At elevated scan speeds the speetral quality was
evaluated. Special emphasis was also paid to the detection limits in the full scan and
in the SIM mode. One advantage of sector instruments is the ability of higher
selectivity by using the high resolving power ofthe mass spectrometer. For complex
mixtures, detection levels are limited by interferences from other compounds in the
sample ("chemical noise"). A comparison was made between a mass speetral
resolving power of 2000 and unit mass resolution, such as available with quadrupale
and ion trap instruments. The performance of high-speed GC/MS will again he
illustrated by the analysis of various samples from industrial and environmental
origin.
5.3.2 Experimental
The system used consistedof a Fisons GC 8000 (Fisons Instruments, Milan, Italy) in
combination with a MasSpec double focusing magnetic mass spectrometer (Fisons
Instruments, Manchester, England). The mass spectrometer bas an EBE tri-sector
geometry (Figure 5.9). Before focusing, the ion beam from the ion souree bas both
angular and energy divergence [16]. The energy dispersion, resulting from the
difference in the positions at which various ions are formed in the ion chamber and
114 Chapter 5
detector
GC souree
Figure 5.9
Schematic representation ofthe VG MasSpec.
from the kinetic motion of the i ons, is corrected by the combined energy focusing of
the sectors. Ions are also focused spatially at the detector. A decrease in mass
dispersion is achieved by the magnetic sector. Because ofthis "double focusing" it is
possible to obtain higher mass spectrometric resolution.
For all experiments a 5 m x 50 J.lm I.D. DB-1 column (J&W Scientific, Folsom,
USA) with a 0.17 !liD film thick:ness was used. A Tescom 44-1100 high pressure
regulator (Tescom Inc., Minnesota, USA) was used to enable high inlet pressures for
the helium carrier gas. The chromatographic system was operated in the constant
pressure control mode. High pressures are required in order to obtain maximum
separation efficiency. The optimum inlet pressure for this 5 m x 50 f.tm LD. column
was 10 bar over pressure. The inlet pressure was monitored by a custom-made digital
pressure indicator. Because high split ratios were used to obtain small input band
widths, it was difficult to get quantitatively reproducible results with manual
injection. To overcome this problem, a Carlo Erba A200S autosampler (Carlo Erba,
Milan, ltaly) was used. The SSL 71 split/splitless injector was operated at 250°C. The
sample volume was 0.5 f.ll and a split ratio of about 111200 was used. The column
High-speed narrow-bore capillary GC coupled to various mass spectrometers 115
ternperature was 1 00°C and the transfer line frorn the gas chrornatograph to the MS
was rnaintained at 250°C.
In the full scan mode, the scan range was 40 to 200 Da operating the MS at a rnass
resolving power of 400 (10% valley). In the SIM mode, the detection lirnits of 1,4-
dichlorobenzene were evaluated by monitoring the ions of rn/z 111.00, 145.97 and
147.97 at rnass spectrornetric resolving powers of 300 and 2000, respectively. Data
processing was performed using the VG Opus software.
5.3.3 Results
In a scanning rnass spectrometer, ionsof different rnasses are detected sequentially.
For quadrupole instrurnents, the scan speed is lirnited because of the relatively low
velocity of the i ons in the rnass analyser. The velocity of the i ons has to be srnall in
order to allow enough oscillations in the rnass analyser for separation. If the
quadrupole is scanning too fast, the speetral quality is lost and the resolution is
reduced drastically. At too high scan rates, no ion can pass through the rnass
analyser.
Normally, with conventional sector instrurnents, the ion velocity is typically more
than 100 tirnes higher than in a quadrupole. The scan speed of rnagnetic instrurnents
depends on the ability to rapidly change the rnagnetic field without causing Eddy
currents which distort the field shape and slow down the scan. This problern can be
overcorne by the use of srnall, fast scanning, fully larninated, low inductance magnets
with high power current supplies. Also in the SIM mode, the use of a srnall magnet
cornbined with the high dynarnic accuracy of a multi-point Hall probe (to control the
magnet field) enables the rnass spectrometer to jurnp frorn peak top to peak top at
medium resolution (R=2000) in approxirnately 25 rnsec. Reducing the switch time
allows more dweil time (rneasuring time) per selected ion. In this way a higher
sensitivity can be obtained because the time spent monitoring the ionsper unit time
is higher or higher sampling frequencies can be applied without losing sensitivity.
Additionally, for a sector instrument, the rnass resolving power (10% valley) R =
mi ~m is constant over the entire rnass range. F or a quadrupole, the peak width is
constant which rneans that the resolution depends on the transrnitted rn/z value. For
quadrupole instrurnents, a peak width value of 0.5 arnu is quite cornrnon, hence the
116 Chapter 5
resolution is R = 2 m. In the lower mass range (40-300 Da}, which is of interest for
high speed GC separations, the peak width for a sector instrument operating at
resolution 300 is smaller than the peak width obtained with a quadrupole instrument.
In this respect, higher selectivity is obtained and the use of a sector instrument is
favoured. Extra selectivity could be obtained by operating the mass spectrometer at
higher resolving power. Higher scan speeds in combination with lower detection
limits and higher selectivity make the combination of high-speed narrow-bore
capillaries with sector instruments highly promising.
5.3.3.1 Scan speed
The most common form of magnetic scan is the quadratic one, either upward or
downward in mass. It has the advantage of producing mass speetral peaks of constant
width. With the MasSpec, the magnet is scanning from high to low masses. The
cycle time for completing one speetral scan depends on the scan time and the reset
time. The scan time is a function of the scan speed of the magnet and the mass range
selected. A small reset time for the magnet is essential because otherwise the
measuring time, and thus sensitivity, will be reduced. By the use of a fast switching
magnet, higher scan speedscan be obtained. In Table 5.2, an overview of the scan
time, reset time and the resulting scan speed as a function of different mass ranges
Table 5.2 Scan speed in the full scan mode for different mass ranges and mass resolving powers (at 10% valley) for the sector instrument.
Mass Range (Da)
R=300
R=2000
fscan (msec) t.eset ( msec) scan speed (scans/s)
tscan ( msec) 1:reset ( msec) scan speed (scans/s)
50-200
30 20 20
120 20
7.14
50-500
50 50 10
300 50
2.86
High-speed narrow-bore capillary GC coupled to various mass spectrometers 117
and resolving powers is presented. The scan speed varled from 10 to 20 scans per
second, depending on the mass range for a resolving power of 300, and decreased to
about 3 to 7 scans per second fora medium resolving power of2000.
It should be noted that at higher scan speeds, the dynamic resolution (when actually
scanning a given mass range at a certain scan speed) is different from the static
resolution (as determined by adjusting the slit width when tuned on a reference
peak). This is illustrated in Table 5.3. From this table it is clear that the dynamic
resolution is limited to 300 for high acquisition rates in the full scan mode. In the
SIM mode or at low scan speeds in the full scan mode, the dynamic resolution is
hardly affected if the static resolution is below 5000. For higher values of the static
resolution, the dynamic resolution decreases rapidly.
Often, the compounds of interest are present at low levels and hence selected ion
monitoring (SIM) techniques are preferred in order to obtain better sensitivity.
Because the detection limit is dependent on the dwell time (measuring time) per ion
mass, it is important that the switch time for jumping from one m/z-value to another
is minimised. Fora sector instrument it can be derived [16] that the mass to charge
ratio (m/z) monitored by the detector, at a fixed radius r, is proportional to the square
of the strength of the magnetic field (B) and inversely proportional to the
accelerating voltage (U):
Table 5.3 Comparison of static and dynamic resolving power at different scanning modes and scan speeds for the sector instrument (Data were taken from the instrument specifications document).
Static resolving power Dynamic resolving power Full scan Full Scan SIM
3 scans/sec 20 scans/sec 20 cycles/sec
300 300 300 300 2000 2000 300 2000
1 0000( ultimate) 3000 300 >5000
118 Chapter 5
(5.2)
From this equation it can he seen that there are two ways for jumping from one ion
mass to another. In the first situation the magnet current is switched while the
accelerating voltage is kept constant. The other way is to keep the magnetic field
strength constant and vary the accelerating voltage. For reasons of detectability, it is
preferabie to vary the magnetic field. For conventional sector instruments, relatively
long switch times are required when jumping the magnet. This restrietion is of no
consequence for conventional GC because of the broader peaks. It becomes,
however, more critica} for highMspeed separations. The scan speed should be
compatible with the separation speed to match the chromatographic resolution. The
measurement time per ion should be as high as possible in order to maximise
sensitivity. Because the MasSpec has a small magnet, the switch time can be reduced
to 15 msec. For voltage sweeping, a switch time of only 5 to 10 msec is required.
Unfortunately, the ions may be monitored only over a restricted mass range, limited
to a 2:1 mass ratio, because of differences in the efficiency of ion transmission [ 16].
In Table 5.4 the switch times for voltage and magnetic SIM for going from one mass
to another are shown. Than, the dweil time per ion per cycle can be calculated as a
function of the number of selected i ons and the number of cycles per time unit. The
dwell time per ion per cycle is illustrated in Figure 5.10. Forthese calculations, it is
Table 5.4 Switch times in the SIM mode from peak top to peak top in magnetic SIM and voltage SIM, as a function of the mass resolution for the sector instrument (Data were taken from the instrument specifications document).
Mass resolving power (at 10% valley)
300 2000
Switch Time (msec) in magnetic SIM
15 25
Switch Time (msec) in Voltage SIM
5 5
High-speed narrow-bore capillary GC coupled to various mass spectrometers
A
B
c
Figure 5.10
60
40;
20
2
40
30
20
10
0 2 3
4
4
5 6 7 8 9 10
Number of selected ions
5 6 7 8 9 JO
Number of selected ions
w.-------------------------.
15
10
5
2
0 2 3 4 5 6 7 8 9 10 Number of selected ions
119
Dweil time per selected ion per scan as a function of the number of selected ions at (A) 5 cycles per second, (B) 10 cycles per second and (C) 20 cycles per second switching voltages and jumping the magnetic field in SIM at low and medium resolution MS; voltage SIM (l), magnetic SIM at low resolution (2) and magnetic SIM at medium resolution (3).
120 Chopter 5
assumed that the magnet reset time (time required for switching from the last peak
top of a cycle to the first peak top of the next cycle) is twice the switch time. From
Figure 5.10, it can be concluded that the sensitivity, which is proportional to the
dweil time per ion per cycle, decreases drastically when the number of selected i ons
is increased. For 5 cycles per second (Figure 5.10A), it is possible to perform SIM by
jumping the magnet or by switching voltages while good sensitivity is obtained.
Only for high resolution SIM by jumping the magnet, the number of i ons recorded is
limited to approximately seven. Even with 10 cycles per second, it is possible to
perform magnetic SIM with a few ions (Figure 5.10B). If the number of cycles is
increased up to 20 (Figure 5.10C), it is no longer possible to perform SIM by
jumping the magnet In the same tigure it can be seen that only a few ions can be
recorded by switching voltages.
5.3.3.2 Detection limits and dynamic range
The detection limits were again calculated according to equation (2.27). To evaluate
the sensitivity, a series of mixtures with increasing sample concentrations was
injected. The sensitivity was calculated from the slope of a plot of the area response
vs. the amount injected onto the column.
In the full scan mode, the detection limits were in the range of I to 4 pg (see Table
5.5). The detection limits in the SIM mode depend on the selected mass resolving
power, the sample complexity and the time an ion is recorded per unit time. The
Table 5.5 Detection limits of the combined high-speed GC-MS set-up in the full scan mode for the sector instrument.
Compound
decane 2,2,3 ,3-tetramethylhexane 1-chlorooctane undecane 2-dodecanone
Detection limit (pg)
1.3 1.7
2.5 1.8 3.4
High-speed narrow-bore capillary GC coupled to various mass spectrometers 121
Table 5.6 Switch time and dweil time for the selected i ons of 1 ,4-dichlorobenzene for the determination ofthe detection limits at low (300) and medium (2000) mass spectrometric resolution with the sector instrument.
rn/z low resolution (R = 300) medium resolution (R = 2000) (amu) switch time dweil time switch time dweil time
(msec) (msec) (msec) (msec)
147.97 30 10 40 10 145.97 20 10 30 10 143.00 30 20 111.00 20 10 30 10
cycle of dwell and switch times for the selected ions are presenled in Table 5.6.
When performing SIM at higher resolution it is necessary to monitor at least one
reference mass of the calibration gas. For this reason the calibration gas valve is
opened during the chromatographic run. By the use of a lock-mass check channel it
is possible to ensure the elimination of drift and hysteresis effects, allowing the mass
spectrometer to switch accurately to the peak tops by checking the reference mass
and adjusting mass calibration during the acquisition. In Table 5.7, the detection
limits for the most abundant i ons from 1 ,4-dichlorobenzene are shown. As can be
seen from this table, the detection limits of the system are improved when the
resolving power was decreased. When performing higher resolution MS, the souree
and collector slits are adjusted to accept a narrower ion beam. In this way ion
transmission is reduced and the sensitivity is decreased. The difference in detection
limits for selected masses is influenced by the relative abundance of the selected i ons
in the spectrum and the background noise level. For example, consirlering the
isotopic pattem of two chlorine atoms, the theoretica} abundance ratio of the i ons at
mlz 147.97 and 145.97 should be close to 113. At low resolving power, however, the
ratio of the detection limits was found to be about 2.5 and increased even up to 4 at
high resolution. This difference is caused by the difference in ( chemical) noise level
for these masses.
122 Chapter5
Table 5.7. Detection limits in the SIM mode for some selected ions and the sum of the selected ion currents of 1,4-dichlorobenzene obtained for resolution 300 and 2000 jumping the magnet for the sector instrument.
mlz
111.00 145.97 147.97 sum of selected ion currents
Detection limit (pg) R=300
6.2 2.0 5.6 2.6
Detection limit R=2000
28.5 10.0 39.2 13.4
For clean samples the detection limits are lower and hence better, when operating the
mass spectrometer at low resolution. Under these conditions, the sensitivity is
maximised because the mass window is broader and almost all ions can reach the
detector. At low resolution lower detection limits are only obtained if the sample is
not complex. SIM data acquired via quadrupole GC/MS systems can be misteading
due to interferences from other eluting compounds (chemica! noise) within the
sample that can mask the compounds of interest. For more complex mixtures and in
instances where the concentration of the compounds of interest is particularly low,
chemical background signals become a significant problem that hinder both
identification and quantitation. These problems have in the past been overcome by
pre-concentrating or isolating these compounds prior to GC/MS. This, however,
often involves complicated and lengthy chemical and chromatographic separation
techniques which can often produce misteading results. By increasing the resolution
in the SIM mode, extra selectivity can be obtained. Disturbing components can be
eliminated based on their masses. This is illustrated by the analysis of PCBs spiked
in a waste oil sample. For this experiment, the untreated sample was injected and
analysed in the magnetic SIM mode at low and medium mass spectrometric
resolution. Table 5.8 gives a list of the ions monitored together with their elemental
composition, switch and dweil times in one cycle. In two instances the ions
High-speed narrow-bore capillary GC coupled to various mass spectrometers 123
Table 5.8 Elemental composition, switch time and dwell time for the selected ions in the analyses of PCBsin an oil at low (300) and medium (2000) mass spectrometric resolution with the sector instrument.
m!z elemental low resolution (R 300) medium resolution (R = 2000) (amu) composition switch time dweil time switch time dweil time
(msec) (msec) (msec) (msec)
222.00 C12Hs35Ch 40 10 50 10 255.96 C12H/5Ch 15 10 25 10 291.92 Ct2H/5CI/7Cl 15 10 25 10 325.88 C12H735Cll7Cl 15 10 25 10 359.99 C12Hl5Cll7Ch 15 10 25 10 218.99 Reference mass 25 20
containing one 37 Cl isotope have been monitored and even the for the highest
selected mass ions incorporating two 37Cl isotopes have been selected because of
their more intense ion currents.
In Figure 5.1la and 5.11b, the results of monitoring the ions m/z 222.00, 255.96,
291.92 and 325.88 at low resolving power (R = 300) and medium resolving power (R
= 2000) respectively, are shown. The responses obtained at low resolution exhibit
severe interference. Improvements in the quality of data are seen in all traces by
operatingat higher resolution. In the m/z 325.88 trace the rising baseline towards the
end of the run in the low resolution mode suggests the presence of column bleed.
Interferences limit the detectability of compounds in the lower concentration range at
low mass speetrometrio resolution. By the use of higher resolution MS, an
appreciable decrease in interferences is observed. When the ion intensity of the
different mass chromatograms is examined, it can be seen that the intensity of the ion
mass currents registered with medium resolution MS is decreased by at least a factor
of ten. This can be explained by the small mass window by which many of the i ons
of the same nominal mass were discarded. Although the sensitivity for the
compounds of interest is reduced, the interferences from ions at the same nomina!
mlz value are reduced even more, thereby enhancing the signal-to-noise ratio.
124
3 4
Figure 5.lla
5
Chapter 5
6
Retention time (min)
High-speed separation of a waste oil spiked with PCBs (Arochlor 1242) recorded with a sector instrument: GC colwnn: DB-I, L = 5 m, LD. = 50 Jlffi, dr= 0.17 Jlm. GC conditions: Teven= 50"C ballistically heated to 280"C, Pi = 11 bar. MS conditions: magnetic SIM, ion current of 222.00, 255.96,291.92 and 325.88 at low resolution MS (R = 300).
High-speed narrow-bore capillary GC coupled to various mass spectrometers
3 .--------.-----------------------------------. m/z=222.00
2
1
0 ~~~~~~~~==~~~~~~~~~
4 .--------------------------------------------. m!z= 255.96
3
2
1
0 ~------------~----------~----------~----~ 8
6
4
2
0
1.5
1.2
0.9
0.6
0.3
Jvl
mlz 291.92
~~ f\.
mlz= 325.88
0.0 l::::;==========::========~~::....:_~:._::::.._:~==::::::J 3 4 5
Figure 5.11b
6
Retention time (min)
125
High-speed separation of a waste oil spiked with PCBs (Arochlor 1242) recorded with a sector instrument: GC column: DB-1, L = 5 m, LD. 50 J.lm, dr 0.17 J.liD. GC conditions: Toven = 50°C ballistically heated to 280°C, Pi = 11 bar. MS conditions: magnetic SIM, ion current of 222.00, 255.96, 291.92 and 325.88 at mediwn resolution MS (R = 2000).
126 Chapter 5
To evaluate the linearity of the system, various samples with increasing
concentration were injected. At injected amounts (on the column) exceeding a few
ng, overtoading of the separation column was observed. It was found that the mass
spectrometer behaves perfectly linearly over the entire range from the detection
limits of approximately 1 pg in the full scan mode, or 5 to 50 fg in the SIM mode up
toa fewng.
The quality of the mass spectra in the full scan mode was very good over the entire
range range. Library search gave good results. Even in the low pg range it was
possible to identify the compounds.
5.3.3.3 Applications
In Figure 5.12 a fast separation of a reference alkylate standard mixture (no. 4-8267,
Supelco, Bellafonte, USA) is shown. The separation is carried out in less than 90
seconds. To obtain the same separation efficiency with a 320 11m column, 10 to 15
minutes are required. Compared with reference chromatograrns, the most volatile
compounds were not observed because of evaporation of these compounds from the
0 15 30 45 60 15 90
Time (s)
Figure 5.12 High speed separation of a reference alkylate standard recorded with a sector instrument. GC column: DB-1, L = 5 m, LD. = 50 J.lm, dr 0.17 J.lffi. GC conditions: T0ven = 40°C ~ 40°C/min ~ 100°C, Pi = 11 bar. MS conditions: El, scan speed = 12.2 scans/s, mass range 60-200 Da, mass resolution = 300.
High-speed narrow-bore capillary GC coupled to various mass spectrometers 127
sample vial.
Another interesting application, shown in Figure 5.13, is the separation ofthe EPA
610 P AH mixture with splitless injection. The separation was performed within 12
minutes. A comparison of a measured mass spectrum and a library spectrum is
shown in Figure 5.14. Libracy search gave good peak matching and all compounds
could readily be identified.
*
3 s 7 9 11 13
Time (min)
Figure 5.13 High speed separation of EPA 610 P AH standard recorded with a sector instrument. GC column: DB-1, L 5 m, LD. = 50 J.l.m, df = 0.17 J.l.ffi. GC conditions: 50°C (2 min) ballistically heated to 300°C, splitless time = 2 min, V inj = 0.3 J.l.l, Pi = 11 atm. MS conditions: EI, scan speed 9.55 scans/s, mass range= 50-500 Da, mass resolution = 300.
5.3.4 Conclusions
It has been demonstrated that the EBE sector instrument operating in the full scan
mode is compatible with chromatographic separations in the minute range. In the full
scan mode, the maximum scan speed varied from 10 to 20 scans per second,
depending on the mass range for low resolving power and decreased to about 3 to 7
scans per second for a medium resolving power. 5 to 10 cycles per second could be
128 Chapter 5
A 100
- 80 <f, -.€ .., 60 .I
0 40 > ·:g -e 20
0 50 75 100 125 150 175 200
mlz
B 100
- 80 <f, -.€ on 60 t::
-~ 0 40 > .... lii ]
20
0 11 .11 lil
50 75 100 125 150 175 200
mlz Figure 5.14 Comparison of the experimental mass spectrum (A) of fluoranthene (peak indicated by * in Figure 5.13) and the library mass spectrum (B).
obtained with magnetic SIM. Faster multiple ion detection could only be applied
using fast electric switching. The detection limits obtained in the full scan mode
were in the low pg range. In the SIM mode, detection limits as low as 5 to 50 fg were
obtained, depending on the sample complexity and speetral resolution. It was
demonstrated that for complex mixtures the best signal-to-noise ratio was obtained
=:-:· v_perating the MS in the magnetic SIM mode at medium resolving power.
High-speed narrow-bore capillary GC coupled to various mass spectrometers 129
5.4 TIME-OF-FLIGHT MASS ANALYSER
5.4.1 Introduetion
With the mass spectrometers described in previous sections, the scan rate was limited
to 10 to 20 spectra per second. This is sufficient for chromatographic separations in
the range of a few minutes. To perform chromatographic separations in the range of
seconds, higher scan speeds are required.
Higher sampling frequencies can be obtained by application of a time-of-flight
(TOF) mass analyser. This type of mass spectrometer operates in a pulsed mode
rather than in a continuous mode. The ions formed are sent as a discrete ion pulse
into the field-free region of the flight tube. All ions are accelerated to the same
kinetic energy and hence, each ion has a characteristic velocity that depends on its
mass. As a result, ions of different mass-to-charge ratios are spatially separated as
they travel in the flight tube. In principle, a time-of-flight mass spectrum can be
recorded within 100 to 150 f.!Sec.
In this section, two types of time-of-flight (TOF) mass analysers are discussed, i.e.
the reflectron time-of-flight [17-18] and the orthogonal acceleration time-of-flight
(oa-TOF) [19-22]. The basic principlesofthese two mass analysers are described in
some detail. The influence of the experimental set-up on the performance (scan rate,
sensitivity, mass resolution, quality of spectra and mass accuracy) are described and
illustrated by some fast separations.
5.4.2 Instrnmentation
In both GC/MS systems a HP 5890 gas chromatograph was applied (Hewlett
Packard, Palo Alto, CA, USA). The gas chromatograph had a split/splitless injection
port. To allow operation at high inlet pressures, a Tescom 44-1100 high pressure
regulator (Tescom Inc., MN, USA) and a custom-built digital pressure indicator were
installed in the carrier gas line. The original splitter valve was replaced to allow for
higher splitflows. Helium was used as carrier gas.
For the experiments with the reflectron TOF instrument, a 2.7 m x 50 J.lm LD. DB-1
column (J&W Scientific, Folsom, USA) with a 0.05 J.lm film of stationary phase was
used. The inlet pressure for maximum separation efficiency was about 8 bar. The
130 Chopter 5
injector and GCIMS transfer line were held at 280°C and 250°C, respectively. For
the oa-TOF experiments, a longer DB-1 column of 10 m x 50 J.Lm I.D. with a 0.05
J.Lm film was used, operated at an inlet pressure of 20 bar. The GC injector and the
custom-built transfer-line were operated at 300°C.
The experiments with the reflectron time-of-flight mass analyser were performed in
co-operation with the laboratory ofProf.dr. H. WoUnik ofthe University ofGiessen
(Germany}. The experimentalset-up is shown in Figure 5.15. The reflectron time-of
flight mass spectrometer was equipped with an electron impact storage ion-souree
where ions are stored in the negative space-charge formed by the emission current of
a cylindrical catbode (5 mA}. The stored ions were extracted by pulsing one of the
electrodes in the ionisation region. The data acquisition was started with a pre-set
turbo pump
ionpackets
Figure 5.15
turbo pump
grid-free ion reflector
Schematic diagram ofthe GC-reflectron time-of-flight MS.
High-speed narrow-bore capillary GC coupled to various mass spectrometers 131
delay. After passing through a lens and a pair of deflectors the ions entered a grid
free ion mirror and were focused onto a discrete-dynode secondary electron
multiplier (SEM) ETP AF-820 (ETP Ltd., Australia). The detector signal was DC
coupled to the pre-amplifier of the transient recorder. Data acquisition was done by
an on-line averaging transient recorder PI 9825 (Precision Instruments Inc.,
Knoxville, TN, USA) installed in a HP Veetra 386/25 PC as a double plug-in board
directly interfaced to the AT bus. This board was capable of adding a new 8-bit
digitisation of the input signal to a 16-bit sum every 5 nsec and storing it into 262 K
wordsof summation memory. The digitisation rate, the number of data points, the
number of summations and the "net" scan rate (number of averaged spectra per
second) were controlled by custom-built software. Each averaged spectrum was
transferred to the host computer's memory. After the run all spectra were storedon
hard disk for retrieval and further processing. With custom-built software
(programmed in C), the abundances of the peaks in each mass spectrum above a
threshold value were accumulated in order to reconstruct the chromatogram.
The experiments with the orthogonal acceleration time-of-flight mass spectrometer
were performed in the laboratory of Prof.dr. M. Guilhaus of the university of New
South Wales (Australia). A schematic representation of the experimental set-up is
shown in Figure 5.16. The ion souree of the oa-TOF was a VG 70-70 electron
ionisationlchemical ionisation souree (VG Analytical Ltd., Manchester, UK)
operated with an extraction voltage of approximately -60 V. In the orthogonal
acceleration region of the mass analyser, the i ons were gated orthogonally by a push
out pulse of 88 V. The ions travelled along the drift tube over approximately 1.5 m
and were detected by a micro channel plate (25 mm diameter Galileo-type 3025N
metal anode, Galileo Electro-Optics, Sturbridge, MA, USA). The signal was
monitored with a LeCroy 9450 digital oscilloscope (LeCroy, Chesnut Ridge, NY,
USA) or recorded on-line by an integrated transient recorder PI 9825 (Precision
Instruments Inc., Knoxville, TN, USA) installed in an Intel 486 computer. A detailed
description ofthe experimentalset-up ofthe oa-TOF is given elsewhere [19,22].
132
GC
Figure 5.16
EI/Cl souree
beamaxis
orthogonal accelerator
inner drift-tube ~
derec:~V pbne~
Schematic diagram of the GC-oa-TOF MS.
Chapter 5
electron multiplier continuous ion beam detector
The digitisation rate, the number of data points, the number of transient
accumulations and the number of scans were controlled by custom-built software. In
high-speed GC/TOF a vast amount of data is generated. For this reason, it is
necessary to reduce the size of the datafiles. In the work with the oa-TOF, data
rednetion was performed by custom-made software in Matlab. First of all, the
coherent noise, presumably caused by the unavoidable digitisation of analogue
signals generated on the data acquisition board, had to be subtracted :from the raw
time-of-flight mass spectra. This is important in order to detect small peaks and to
determine the flight-time accurately. The coherent noise was approximated by
averaging 50 blank spectra. The influence of base line subtraction for peak detection
is illustrated in Figure 5.17. In Figures 5.17A and 5.17B, a small part ofthe mass
spectrum of 1 ,3-dibromopropane is shown. Line 1 is the intensity of the mass
High-speed narrow-bore capillary GC coupled to various mass spectrometers
A
2
3
B
2
3
Figure 5.17
27
2220 2235 2250 2265 2280
Channel Number
93 95 105 107
4100 4150 4200 4250 4300 4350 4400 4450 4500
Channel Number
133
Part ofthe mass spectrum of 1,3-dibromopropane with m/z = 27 (A) and m/z = 93, 95, 105 and 107 (B) recorded wit the oa-TOF-MS. Line 1: intensity in the recorded TOF spectrum. Line 2: approximation for the coherent noise. Line 3: baseline corrected spectrum.
channels in the raw mass spectrum. Line 2 is the approximation of the coherent
noise. Line 3 is obtained by subtracting line 2 from line 1.
From this figure, it is clear that the signal-to-noise ratio is increased, especially for
ions with a low abundance. In each baseline corrected scan, a peak is recognised as a
mass peak if the intensity of a mass channel exceeds a threshold value. The total ion
current at any time is the sum of intensities of all successive channels with an
134 Chopter 5
intensity higher than the threshold value. The centred channel position for a
particular ion is calculated using the equation:
Centred Channel Position = L: Channel Intensity
L: Channel Position x Channel (5.3)
The centred channel positions and channel intensities are stored in the reduced
datafile. This file was converted so the GC/MS results could be evaluated with
Shrader software.
5.4.3 Results
5.4.3.1 Scan speed
The time-of-flight mass spectrometer is inherently much faster than conventional
scanning instruments. In principle, the scan time of a time-of-flight MS equals the
flight time of the heaviest ion of interest. For the reflectron TOF-MS this was about
130 JlSec for anion with a mass of 550 Da. The time was even reduced to 50 JlSec
with the oa-TOF-MS. This means that per time unit more than a few thousand
spectra could be recorded. In practice however, these instruments were significantly
slower due to the computer time required for processing and storing the acquired
spectra. This transfer caused an appreciable dead time of approximately 10 ms during
which no spectra could be taken. This puts an upper limit of 100 scansis to the scan
rate. Moreover, in order to obtain acceptable signal-to-noise ratios, very often a
number oftransients (raw mass speetral scans) had to be accumulated prior to storing
a full spectrum to disk. The increase ofthe signal-to-noise ratio is proportional to the
square root of the number of accumulations, assuming there is only white noise. The
eperating time to record one time-of-flight mass spectrum is the sum of:
- number of transients to accumulate multiplied by the flight-time for the heaviest
ion of interest
- time for accumulating a preselected number of transients
- transfer time neerled for sending a spectrum to the host computer's memory.
Unfortunately, with the instruments used, a relatively long time was needed to
transfer a final spectrum to the computer's memory.
High-speed narrow-bore capillary GC coupled to various mass spectrometers 135
In Table 5.9 a number of possible combinations of scan acquisition rates and
transient accumulation factors are given for the reflectron TOF-MS used. The table
shows the resulting number of spectra/second, the corresponding number of
accumulated scans and the number of single scans (transients) collected per second.
Two extreme situations can be distinguished. First of all, in the case no mass speetral
accumulation is applied, every single transient recorded can be stored on disk as a
full spectrum. In this situation, the scan rate is maximised. 61 scans could be
recorded per time unit. On the other hand however, one could also opt for
accumulating a number of raw scans before storing the resulting spectrum on disk.
The maximum number of accumulations (which can be further increased by
modifying the software) was 128. In this situation, the signal-to-noise ratio was
improved at the expense of a much slower speetral acquisition rate of only 9
scans/sec. In practice a compromise has to bemadebetween optimum response and
maximum speetral acquisition rate. Lower overall acquisition rates (high number of
scans accumulated) result in a higher sensitivity.
This situation becomes even more pronounced for the oa-TOF MS. Here, the
transient spectra could be recorded even faster than with the reflectron TOF. The
time required to register a transient time-of-flight mass spectrum with the oa-TOF
Table 5.9 Possible combination scan acquisition rates, transient accumulation factors and the corresponding single transients recorded per second for the reflectron TOF MS.
Number of transients Resulting Number of single accumulated spectra/second transients/second
128 9 1152 64 16 1024 32 27 864 16 36 576 8 51 408 4 55 220
61 61
136 Chapter5
was in the range of 40 to 50 IJ.Sec for ions of mass 350 to 550 Da, respectively. In
Table 5.10, a number of possible combinations of the number of transients per
second, the number of accumulations, the resulting net scan speed and actual
acquisition time per time unit for the oa-TOF are shown. Unfortunately, also here,
the relatively long time required to transfer the speetral data from the data acquisition
board's memory to the memory of the host computer drastically reduced the
obtainable maximum scan speed. The upper limit for the scan rate was only 23 scans
per second. In our experiments it was found that the transfer time required was much
longer than that expected from the specifications of the transient recorder supplied
by the manufacturer. From this table, it can be seen that when the number of
accumulations was very low, the computer was busy most of the time with
transferring data. When each transient recorded was slored as a speetral scan, the
computer was occupied with data transfer for 99.9% of the available time. Only
during 0.1% ofthe total time, a TOF spectrum was recorded. As a consequence, the
sensitivity was very low. With this instrument, up to 32768 transients could be
accumulated. By accumulating more transients, not only the signal-to-noise ratio is
improved, but also the total acquisition time is increased. Because the number of
transients recorded per second is increased, the sensitivity is improved. As can be
seen from Table 5.10, it was possible to acquire more than 17000 transients per
second time for a high number of accumulations. In this situation, about 71% of the
total time was used for data registration. The remainder ofthe time the computer was
occupied by data accumulation and transfer. By increasing the number of transients
from 1 to 32768, the net scan rate reduced form about 23 to 0.5 scans per second. As
mentioned earlier, a compromise has to be made between optimum response and
speetral acquisition rate. Care should be taken, however, to use acquisition rates that
are high enough to accurately reconstruct the chromatogram. In the future, faster
transient boards will enable faster data handling and higher scanning rates.
High-speed na"ow-bore capillary GC coupled to various mass spectrometers 137
Table 5.10 Possible combination of number of transients recorded per second, number of transients accumulated per scan, acquisition rate and acquisition time per time unit with the oa-TOF.
Numberof Number oftransients Resulting scan Acquisition time transients accumulated per scan speed persecond per second (transients/scan) per second (msec/1 second)
(transients/second) (scans/second)
23 22.80 0.9 90 4 22.62 3.7
179 8 22.36 7.4 357 16 22.30 14.6 710 32 22.19 29.1
1415 64 22.11 57.9 2783 128 21.74 114.0 5545 256 21.66 227.1 9114 512 17.80 373.4
12308 1024 12.02 504.3 14520 2048 7.09 594.6 16015 4096 3.91 656.2 16875 8192 2.06 690.0 17367 16384 1.06 711.0 17837 32768 0.53 717.3
5.4.3.2 Mass resolution and quality of mass spectra
For reflectron TOF mass analysers, the most successful method to achieve mass
independent focusing at the detector is obtained by using an ion mirror. Mass
resolutions up to 1500 full width half maximum (FWHM) can be obtained with
reflectron TOF.
With the oa-TOF, higher mass spectrometric resolution can be obtained. In the
earlier versions of the oa-TOF, deflection plates axially decelerated the ions that
were gated orthogonally. The axial deceleration imposed by the steering plates
degraded the mass resolution because of transfer of axial velocity dispersion into the
direction of the flight tube and the inhomogeneons electric field created by the
steering plates. The geometry of the orthogonal time-of-flight mass analyser used
138 Chapter5
here, conserved the axial velocity of the ions which are deflected orthogonally.
Guilhaus et al. reported that by using only pure orthogonal acceleration, resolving
powers up to 4000 (FWHM) fora 1.5 m flight tube could be obtained [22], without
making use of a reflectron. For our experiments the instromental set-up was slightly
modified to increase its sensitivity. These modifications resulted in a decreased
resolving power of about 700 for mass 40 and 1000 for mass 200.
Mass calibration and mass accuracy for the oa-TOF were evaluated by recording
spectra of 1 ,3-dibromopropane. This compound yields a number of fragment i ons
spread over the mass range of interest (up to 200 Da) for (high-speed) GC. A
splitless injection of 1 ,3-dibromopropane was performed and 1000 scans were
recorded. One scan was . used to perform mass calibration. An excellent linear
relationship was found between the square root of the mass and the flight time. A
correlation coefficient of 0.9999 was obtained. For the determination of the mass
Table 5.11 Mass repeatability and mass accuracy measured over 1000 scans for 1 ,3-dibromopropane for the oa-TOF.
standard deviation • l1le mass accuracy
(amu) (ppm) (ppm)
27.023 44 101 39.023 36 2 41.039 19 14 92.933 14 32 94.931 20 2
106.949 27 17 108.947 22 4 120.965 43 37 122.963 9 9 199.883 18 25 201.881 15 33 203.879 29 31
·Mass accuracy = (experimentally determined mass- exact mass ofthe ion)/exact mass of the ion.
High-speed narrow-bore capillary GC coupled to various mass spectrometers 139
repeatability and mass accuracy, the average of each mass over the 1000 scans was
calculated and compared to the exact mass. The results are tabulated in Table 5.11.
The relative standard deviation of these masses is less than 45 ppm. The calculated
mass is also very accurate. For all masses the deviation was smaller than 40 ppm,
except for mlz 27.023 which had a deviation of approximately 100 ppm. This higher
deviation can be explained because the centred channel position could not be
calculated accurately due to the low number of datapoints available over the mass
peak and its low intensity.
To compare the spectra obtained by TOF mass spectrometric detection with library
spectra, the flight times in the raw TOF spectra have to be converted to mass units. In
Figure 5.18 a TOF spectrum of nonane recorded with the reflectron TOF is shown. In
43 57
02 41
71 85
39 99 11 ll I I .u
0 1000 2000 3000 4000 5000 6000 Channel
Number I j • j j • j j j j I j j ' j j j j j j I ' ' ' ' ' j ' ' I I' "" '"'I """'"I""'" .. I.," "" !" .. ' ""! .. "' " .. ! ....
30 40 50 60 70 80 90 100 110 120 mlz
Figure 5.18 Raw TOF spectrum of nonane recorded with the reflectron TOF MS.
140 Chapter 5
the upper axis, the time-of-flight time is indicated. In the ax1s below, the
corresponding mass scale is given. The reconstructed spectra were not always
comparable with library spectra. This can be explained by the fact that the ions are
stored in the ion souree until an extraction pulse is applied to pulse the ions into the
field-free region to record a new transient. During this storage time, similar to the
situation in an ion trap, ion-molecule interactions and/or ion fragmentation can
occur, leading to distorted spectra. Additionally, the electron beam continuously
ionises all gaseous molecules in the ion souree and keeps the ions in the "potential
well" formed by the space charge of the electron beam. Only a limited number of
ions is stored in the source. When no compound is eluting from the column, the ion
souree is tilled with ions of residual gases. During the elution of a compound, i ons
are produced and the potential well in the souree is tilled with these ions. Once this
well is filled, overflow and ion exchange starts. Herewith, preferentlal storage of the
heaviest i ons occurs. As a consequence, the abundance of the heaviest i ons is higher
than that of the ions with a lower mass. Because of the competition reactions during
ion storage and the preferential storage of the heaviest ion, the absolute peak
intensity for masses with a lower mass-to-charge ratio is reduced. Also the relative
intensity of the i ons of the residual gas is reduced. If a plot is made of the intensity of
the ions of the residual gases as a function of the time of analysis, the intensity will
be complementary to the total ion current ofthe chromatogram.
In the oa-TOF-MS the ions were continuously extracted from the ion souree and a
continuous ion beam was created. In this way, ion-molecule reactions are avoided so
that self-chemical ionisation is eliminated. In the oa-TOF-spectrum, the abundance
of the highest masses is higher than of the i ons with a low mass. Assuming that the
charge of all i ons is equal to one, the energy of the i ons in the continuous ion beam
will be the same. As a consequence, the number of ions in the acceleration region
with higher mass is somewhat higher due their lower travelling velocities. For the
reconstructed mass spectra, intensity correction was performed. The reconstructed
spectra obtained in this way show perfect agreement with library spectra.
High-speed narrow-bore capillary GC coupled to various mass spectrometers 141
5.4.3.3 Applications
A fast separation of a ten component mixture obtained using a narrow-bore column
coupled to reflectron TOF-MS is illustrated in Figure 5.19. The speetral acquisition
rate was 35 spectra/s and each spectrum was obtained by accumulating 16 transients.
Each dot in the tigure represents one mass spectrum. From this tigure it is evident
that very high mass speetral acquisition rates are required in order to obtain a correct
chromatagram free of resolution loss due to a too low scan rate.
Another nice fast separation recorded with the reflectron TOF-MS is demonstrated in
Figure 5.20. Here 16 spectra are recorded per time unit and each spectrum was
obtained by accumulating 64 transients.
The detection limits obtained with this experimental set-up were in the range of 10 to
20pg.
In Figure 5.21, a fast separation of a 14 compound mixture, recorded with the oa
TOF-MS is shown. All compounds elute between 25 and 40 seconds. Each of the
6 7 8 9 10 11 12
Time (s)
Figure 5.19 High-speed chromatagram (TIC) of a ten-component mixture recorded with the reflectron TOF MS. GC column: DB-1, L = 2.7 m, I.D. = 50 J.lm, dr= 0.05 J.lm. GC conditions: Toven= 75°C, Pi= 8 bar. MS conditions: EI (70 eV), scan speed= 35 spectra/s, mass range= 35-200 Da, 1 spectrum is obtained by accumulating 16 scans. Componentsin order of elution: nhexane, cyclohexane, n-heptane, methylcyclohexane, toluene, n-octane, chlorobenzene, ethylbenzene, o-xylene, n-nonane.
142
0 20
Figure 5.20
40 60 80
Time (s)
ChapterS
High-speed chromatograms (TIC) of car fuel recorded with the reflectron TOF-MS. GC column: DB-1, L = 7 m, LD. = 50 J.lffi, dr = 0.05 J.lffi. GC conditions: 25°C ballistically beated to 200°C, Pi 12.5 bar. MS conditions: EI (70 eV), scan speed= 16 spectrals, mass range= 35-200 Da, eacb spectrum is obtained by accumulating 64 transients.
compounds could be easily identified with library searcb. Because the detection limit
of the mass analyser was approximately tens of picograms and, on the other band, tbe
sample capacity ofthe narrow-bore column was limited toa few nanograms only, the
working range was very small. At the end of the chromatogram, overloaded peaks
are observed. During this work it became apparent that the limited dynamic range of
the transient recorder, in conjunction with the significant baseline noise, made it
impossible to observe low ion intensities. A more appropriate detection system
would be based on the counting of individual ion pulses with a time-to-digital
converter (TDC) [23].
The retention times can be further reduced by operating the GC system at higher
temperatures, higher pressures or by using shorter separation columns. Under these
conditions, base line separation of the peaks will no longer be obtained but for
overlapping peaks, the peak profile for each compound can be reconstructed and the
compoundscan be identified by deconvolution [24].
High-speed narrow-bore capillary GC coupled to various mass spectrometers
25 30
Figure 5.21
35 40
Time (s)
143
High-speed chromatagram (TIC) of a 14 component mixture recorded with the oa-TOF MS.
GC column: DB-1, L = 10 m, LD.= 50 IJ.ID, df = 0.05 IJ.ID. GC conditions: T001 = l25°C, Tinj = 275°C, Pi = 20 bar. MS conditions: El, scan speed = 22 scans/s, 1 scan is obtained by
accuruulating 256 transients. Components in order of elution: chloroform, 1,1,1-
trichloroethane, cyclohexane, heptane, methylcyclohexane, toluene, 1,2-dibromopropane, 1-iodobutane, 1-bromopentane, chlorobenzene, ethylbenzene, styrene, nonane, 1,3-
dibromopropane.
5.4.4 Conclusions
It has been shown that TOF mass analysers provide scan rates that match the speed
of high-speed gas chromatographic separations and allow a highly reliable
reconstruction of TIC-chromatograms. The detection limits found were about 10 to
20 pg for the reflectron TOF, and are a few times higher for the oa-TOF. Further
investigations will focus on methods to speed up transfer of spectra so that higher
scan speeds can be obtained. In this way the scan speed and/or the number of
transients recorded per time unit can be increased. The possibility of recording a
spectrum in about 100 JlSec offered by TOF instruments is not only of interest for
high-speed GC that needs high speetral acquisition rates, it can also be useful for
conventional GC. Than, a very large number of transients can be accumulated to
obtain spectra of which the signal-to-noise ratio is significantly improved. In this
144 Chapter 5
way, the TOF mass analysers offertheuser the choice between higher acquisition
rates or higher sensitivities. For the reflectron TOF, mass resolution of 700 to 1500
could be obtained. The experimental spectra might be slightly different from library
spectra due to ion/molecule reactions or fragmentation during ion storage. On the
contrary, with the oa-TOF the mass resolution can be increased up toa few thousand
while the mass accuracy is extremely good. The spectra recorded by the oa-TOF
show good agreement with library spectra. In this respect the oa-TOF offers a better
performance over the reflectron TOF.
High-speed narrow-bore capillmy GC coupled to various mass spectrometers 145
5.5 COMPARISON OF DIFFERENT MASS ANALYSERS AS DETECTORS
FOR HIGH-SPEED NARROW-BORE CAPILL...RY GAS
CHROMATOGRAPHY
First, in this paragraph the possibilities and the limitations of other mass
spectrometers which have not been described in detail in the previous paragraphs,
wîll be discussed briefly. In this way a complete overview is obtained of all mass
analysers as detectors for fast capillary GC. In the last part of this section, the
performance ofthe various mass spectrometers investigated are compared. From this
information the MS ( currently) most suitable for high-speed GC is selected. A few
modifications will be suggested in order to further improve the performance of this
MS with respect to sensitivity, selectivity, scanning rate and/or mass speetral
integrity.
5.5.1 Otber mass analysers
In GC/MS, the scanning speed of the mass spectrometer has to be increased
proportiona!ly to the increased speed of analysis, not on!y for accurate recording of
the chromalogram but also to keep mass discriminatîon effects within acceptable
lîmits. In recent work we used a quadrupole instrument at a scan speed of I 0 scansis
to analyse compounds with a molecular weight be!ow 200 Da, and a speed of 4
scansis scanning the complete mass range up to 650 Da [25]. In 1985, Leclercq et al.
[26] reported already about acquisîtion rates up to 20 scansis for quadrupole
instruments coupled to 50 J.lm LD. columns. Unfortunately, results about sensitivity
and speetral quality at these high acquisition rates were not reported. Disadvantages
of these high scan speeds include the loss of mass spectrometric resolution, mass
accuracy and sensitivity. Recently, Grimm et al. [27] reported about fast scanning
quadrupole instruments. By decreasing the mass filter step to l amu and selecting
only a very narrow mass range (50 -150 Da), acquisition rates of about 250 scans per
second were obtained. lt is important to îndicate here that sînce the mass spectrum is
being undersampled, the resultant total ion current is not quantîtative. Additionally,
the mass resolution was seriously reduced but spectra were sufficiently distinct for
compound identification. In the selected ion monitoring mode, the quadrupole MS
146 Chapter 5
can provide up to I 00 ion intensity readings per second. In the SIM-mode,
unfortunately, only limited mass speetral information is obtained. Hence, this
technique can not be used for the unarnbiguous identification of unknown sample
constituents. F or the partic u! ar quadrupale tested the deleetion limits in the full scan
mode were found to be in the range of a few lens of pg and were reduced to a few pg
in the SIM-mode.
A mass analyser that has high potentials for hyphenation to narrow-bore columns is
the focal plane mass spectrometer (e.g. of the Martaueh-Herzag type) [28-30] which
is a non-scanning EB mass spectrometer with a sensitive detector. A spectrograph of
this geometry disperses the i ons of different masses spatially along the focal plane of
the magnetic sector. Previously, measurements of these ions were made by using a
photographic plate along the focal plane. In this way high resolution could be
obtained but the pbotoplate lacked sensitivity. This problem can be evereome by
using an array detector with high sensitivity and spatial resolution. This detector is
known as the microchannel plate detector (MCP). The detector possesses the high
gain of an electron multiplier and measures the intensities of i ons of different masses
simultaneously. Measuring the full mass range, the MCP detector provides a mass
resolution of 300-500 (with I 0% valley defmition). If the mass range of interest is
reduced to I Da, the mass spectrometric resolution can be increased up to 10000.
This MS has the ability to integrate the signa! over a wide range of times (l ms to
30 s).
Another non-scanning mass spectrometer is the Fourier Iranstorm mass
spectrometer. In Fourier transfarm mass speetrometry (FTMS) [31-35], ions are
trapped in a cylindrical or cubic cel! in a high-vacuum (<I 0." Torr) eh amber centred
in a homogeneaus magnetic field. The ions move in a cyclic motion. By applying a
short broad-band RF signa], the ions become excited and move into a higher
coherent orbit. The packet of ions that is formed induces a smal! alternating current
on the receiver plates. A Fourier transformation is then used to deconvolute this
signa! and delermine the m/z values for the ions present in the cel!. This instrument
is capable of taking up to I 00 spectra!second. Moreover the resolution that can be
obtained with this instrument is very high. The mass resolution is linearly related to
the duration of signa! acquisition. Using a 4 MHz analogue-to-digital converterand a
High-speed narrow-bore capi/lary GC coupled to various mass spectrometers 147
7 T magnet, mass resolution 10000 can be obtained at rn/z = 500 digitising the
transmilled signa! for about 60 ms. This resolution can be finther improved by
narrow-band data acquisition. With the acquisition of only a narrow mass range
(about 10 Da), mass resolution greater than 1.5 106 can be obtained [36]. In this
respect, FTMS is highly interesting. Unfortunately, the FTMS instrument currently
cannot handle higher souree pressures and is therefore less suited for coupling with
GC. Also its sensitivity is currently incompatible with GC. In addition, a large
computer is needed to perform the many complex calculations. The development of
FTMS for the analysis of complex mixtures is still at an early stage.
5.5.2 Comparison of several mass spectrometers as detectors for high-speed
gas chromatographic separations
Various mass spectrometric detectors have been investigated to evaluate their
compatibility and performance in combination with high-speed GC. The principle of
mass selection and the performance of the ion trap, EBE sector instrument, the
reflectron time-of-flight and the orthogonal acceleration time-of-flight mass
spectrometer have been described in detail. The characteristics of the quadrupole
instrument, the Mattauch-Herzog EB spectrometer with a microchannel plate array
detector and the Fourier Transform mass spectrometer are briefly summarised in the
previous section. In Table 5.12, an overview ofthe performances of all these mass
spectrometers conceming scanning rates, deleetion limits, sensitivity, mass
resolution and quality of spectra is given. Finally, in this section the MS most suited
for the registration of high-speed separations is selected and suggestions for further
improvement ofthe performance ofthis instrument are given.
From Table 5.12, it can be seen that the acquisition rate for the scanning mass
spectrometers (quadrupole, ion trap and sector instrument) is far too low for
separations in the range of seconds. The non-scanning mass spectrometers offer
higher acquisition rates. With TOF mass analysers, complete spectra can be recorded
within 100 j.isec. Unfortunately, however, up to now this does not mean that
acquisition rates of I 0000 spectra per second can be obtained. The limiting factor is
the computer time required for data transfer. Data handling takes too much time
compared to mass speetral acquisition. The computer is not sufficiently fast to
-Table 5.13 ... 00
Overview of the scanning rate, sensitivity, speetral quality and mass resolution for several mass spectrometers.
Instrument Scanning rate Sensitivity Speetral Quality Mass RessoJution
Quadrupole FS 4-10 scansis FS 11)..20pg + llllÎt
SIM up to 50 Jlls SIM I pg
Ion trap FS 3-8 scansis El I pg ± 300-1000 Cl (CH.) 5 pg
Sector lustrument FS, R=300 1 0-20 scansis FS, R=300 I pg + 300-2000 (EBE) FS,R=2000 3-7 scansis SIM, R=300 2-lOfg
SIM, R=300 30-50 Jlls SIM, R=2000 10-40 fg SIM, R=2000 20-50 Ills
Reflectron TOF FS 9-66 scansis FS 10-20 pg ± 700-1500
oa-TOF FS 0.5-23 scansis FS 50pg + 1500-4000
MH-MCP FS up to 50 scansis SIM JO fg n.r. FS: 300-500 (EB) mass range 1 amu:
!0000
FTMS FS up to 100 scans/sec n.r. n.r. m!z=SOO: 10000 (acq time: 60 msec) m!z 500: 1.5 1 o' (acq time; 2 s, mass range I 0 aruu)
Legend: oa-TOF: orthogonal acceleration time-of-fligbt, MH-MCP: Mattauch-Herzog microchaonel plate, FTMS: Fourier transfarm mass ~ spectrometer, FS: fuU scan, SIM: selected ion monitoring, R: mass spectrometric resolutionj Ws: Jon intensities per secondt EI: 'G
electron impact, Cl: chemica! ionisation~ acq time: acquisition time, n.r.: oot reported in literature. ~
"'
High-speed narrow-bore capillary GC coupled to various mass spectrometers 149
operate at these high acquisition rates. In the future, faster computers will certainly
become available so that higher acquisition rates can be anticipated.
Good quality spectra are obtained with quadrupole instruments, sector MS machines
and the oa-TOF MS. The spectra obtained with the ion trap are not always library
searchable because of the occurrence of ion-molecule reactions or fragmentation
during the storage period of the ions. The same situation also occurred in the
reflectron TOF MS used here because also this instrument used ion trap storage.
These processes depend on a lot of instrumental parameters like storage time,
concentration of ions andlor molecules in the trap, ion trap temperature, etc. By the
use of some modifications such as the controlled scan functions and the axial
modulation technique, the occurrence of these ion processes can be reduced but not
completely avoided. On the other hand, the detection limits can be improved by the
use of ion trap storage. The detection limits of the ion trap are much better than that
of quadrupole instruments. In a quadrupole instrument, only a small fraction of the
ions is measured. In an ion trap, ions remain trapped in the ion trap until a change in
operating voltages causes the trapped ions of a particular mass-to-charge ratio to
adopt unstable trajectories. In this way, ions of successively increasing m/z values
are ejected from the souree and detected. During scanning, 50% of the ions are
detected so that better detection limits are obtained.
Storage of ions is not necessary to obtain good sensitivity. With the EBE sector
instrument tested, the detection limits were also in the low pg range in the full scan
mode although a significant fraction of the ions is lost during transmission in the
instrument. Here the sensitivity is considerably higher because the extraction of ions
from the ion souree is much more efficient due to the higher extraction voltages that
are applied. As already mentioned before, quadrupole instruments are limited in scan
speed because the ions are of low energy [6]. The ions have to traverse the mass
analyser slowly to obtain good mass separation and mass accuracy. For magnetic
instruments, mass separation is not hampered by the high energy of the i ons. With
the EBE sector MS operating in the SIM mode, the detection limits were even
lowered to the fg range. This can be explained by the fact that in SIM the
measurement time per ion is longer. The detection limits with both TOF mass
150 Chapter5
analysers were rather poor. With these instruments the detection limits can be
improved in several ways. First of all, small modifications in the construction of the
ion souree could improve sensitivity. For the oa-TOF MS, the ion extraction
efficiencies can be increased by applying higher extraction voltages. With the oa
TOF MS, the extraction voltage was 60V while a few thousand kilovolts are applied
in sector instruments. Another alternative is to use a more appropriate algorithm to
identify a mass peak in the raw TOF spectrum. In our experiments, the total ion
current chromalogram was calculated by summing the intensity of all mass channels
in each spectrum above a minimum threshold value. With other algorithms, smaller
mass peaks can be detected so that better detection limits should be obtained. Faster
computers will enable hiph, _ acquisition rates as they require shorter times for data
handling and transfer. In this way, the number oftransients recorded per time unit is
increased and better detection limits are obtained.
As was demonstrated by the application of the PCB spiked oil, high mass
spectrometric resolution is important to have high selectivity. Only a few mass
analysers including the EBE sector MS, the oa-TOF MS, the MCP and the FTMS,
have the ability to obtain mass resolutions of a few thousand. With the EB-MCP
instrument, high resolution MS data can be obtained when the mass range of interest
is only a few Da. The highest resolution is obtained with the FTMS. For the FTMS,
the mass resolution is inversely proportional to the duration of signal registration.
High resolution can only be obtained (5000-10000) at relative long acquisition times
( 60 msec ). The resolution in principle can be further increased to 1000000, but than
2 seconds data registration is required per scan. Hence high resolutions can only be
obtained at the expense of acquisition rates. Additionally, the FTMS suffers from
serious disadvantages such as the limited sensitivity caused by the high vacuum
required.
To our opinion, the oa-TOF MS has the greatest potential as mass analyser for
combination with high-speed GC. A high number of spectra can be recorded per unit
time. In the future faster computers will enable acquisition rates of a few hundred
spectra per second. Additionally, the hyphenation of thls instrument with
conventional columns (250-320 J.lm LD. columns) can also be interesting. For such
systems the high scan speed is of no importance, a large number of transients can be
High-speed narrow-bore capillary GC coupled to various mass spectrometers 151
accumulated so that the signal-to-noise ratio in the resulting spectrum is increased.
By accumulating for example 10000 transients to 1 spectrum, the signal-to-noise
ratio will he increased 100 times (assuming there is only white noise). In this way,
the oa-TOF offers the user the possibility to select between high scan rates and high
sensitivities. Moreover, the technique offers high resolution and high mass accuracy.
For the moment, the instrument bas a serious disadvantage being the poor sensitivity.
As indicated before, a more appropriate algorithm for detecting small mass peaks in
the spectrum will be useful to improve sensitivity. Additionally, a new ion souree
design and improved ion transmission will increase the instrument's sensitivity.
5.6 CONCLUSIONS
With rnass-scanning instruments including the quadrupole, the ion trap and the EBE
sector machine, the acquisition rates are limited to a maximum of some 20 scans per
sec. This maximum scan rate is sufficient for chromatographic separations in the
minute range. With non-scanning techniques, such as the time-of-flight MS and the
EB-microchannel plate MS, a complete spectrum can he recorded in 1 msec or less.
The actual scan speed unfortunately is still limited because the computer time
required for data transfer and data handling is far too long. In the future faster
computers will enable scan speeds over a few hundred spectra per second.
Because the sample capacity of the narrow-bore columns used in HSGC is limited to
a few nanograms per compound, highly sensitive detection devices are required.
With most of the mass spectrometers described here, the detection limits were in the
low pg range in the full scan mode and down to even the fg range in the SIM mode.
Hence, an acceptable working range is obtained.
Apart from having a high sensitivity it is also important to have a high mass
resolution. In this way it is possible to add additional selectivity to the
chromatographic analysis procedure.
With each of the instruments considered bere, a compromise bas to he made between
sensitivity, mass resolution (selectivity) and scanning rates. Additionally, the quality
152 Chapter5
of spectra is very important. F or the moment, the oa-TOF offers the best possibilities
as a mass spectrometer for hyphenation to high-speed narrow-bore GC separations.
5. 7 REFERENCES
1. J.F. Holland, C. G. Enke, J. Allison, J.T. Stuffs, J.D. Pinkston, B. Newcome, J.T. Watson, Anal. Chem., 55, 997A (1983).
2. R. Rites, K. Biemann, Anal. Chem., 42 (1970) 855. 3. S.C. Gates, M.J. Smisko, C.L. Ashendel, N.D. Young, J.F. Holland, C.C. Sweely,
Anal. Chem., 50 (1978) 433. 4. C.C. Sweely, J.J. Vrbanac, J.D. Pinkston, Biomed. Mass Spec., 8 (1981) 436. 5. JJ. Vrbanac, C.C. Sweely, J.D. Pinkston, Biomed. Mass Spec., 10 (1983) 155. 6. C.C. Grimm, S.W. Lloyd, "Proceedings of the 42nd Annual Conference on Mass
Speetrometry and Allied Topics", Chicago, Illinois, USA, 1994, p. 492. 7. J.F.J. Todd, Mass Speetrometry Reviews, 10 (1991) 3. 8. G.C. Stafford, D.M. Taylor, S.C. Bradshaw, J.E.P. Syka, "Proceedings of the 35th
Annual Conference of the American Society for Mass Spectrometry", Denver, Colorado, USA, 1987, p. 775.
9. J.N. Louris, R.G. Cooks, J.E.P. Syka, P.E. Kelley, G.C. Stafford Jr., G. Todd, Anal. Chem., 59 (1987) 1677.
10. J.E.P. Syka, J.N. Louris, G.C. Stafford, W.E. Reynolds, U.S. Patent nr. 4 736 101,1988.
11. G.C. Stafford Jr., P.E. Kelley, J.E.P. Syka, W.E. Reynolds, J.F.J. Todd, Int. J. Mass Spectrom. Ion Processes, 60 (1984) 85.
12. C.K. Huston, J. Chromatogr., 606 (1992) 203. 13. J.S. Brodbelt, J.N. Louris, R.G. Cooks, Anal. Chem., 57 (1987) 1278. 14. G.C. Stafford Jr., P.E. Kelley, D.C. Bradford, Am. Laboratory, June 1983, 51. 15. J.W. Eichelberger, W.L. Budde, Biomed. Env. Mass Spectrom., 14 (1987) 357. 16. W. McFadden, in "Techniques ofcombined gas chromatography/mass spectrometry:
Applications in organic analysis", J. Wiley & Sons, New York, 1973, p. 36. 17. R. Kutscher, R. Grix, G. Li and H. Wollnik, Int J. Mass Spectrom. Ion Processes,
103 (1991) 117. 18. R. Grix, U. Grüner, G. Li, H. Stroh and H. Wollnik, Int J. Mass Spectrom. Ion
Processes, 93 (1989) 323. 19. H.H.J. Dawson, M. Guilhaus, Rapid Commun. Mass Spectrom., 3 (1989) 155. 20. J. Coles, M. Guilhaus, Trends in Anal. Chem., 12 (1993) 203. 21. M. Guilhaus, J. Am. Soc. Mass Spectrom., 5 (1994) 588. 22. J.N. Coles, M. Guilhaus, J. Am. Soc. Mass Spectrom., 5 (1994) 772.
High-speed narrow-bore capillary GC coupled to various mass spectrometers 153
23. M. Guilhaus, J. Mass Spectrom., 30 (1995) 1519. 24. L.R Roach, M. Guilhaus, Organic Mass Spectrom., 27 (1992) 1071. 25. P.G. Van Ysacker, H.-G. Janssen, H.M.J. Snijders, H. Wollnik, P.A. Leclercq, C.A.
Cramers, "Proc. 16th Int. Symposium on Capillary Chromatography", P. Sandra (Ed.), Rivadel Garda, Italy, Hüthig Verlag, Heidelberg, 1994, p. 785.
26 P.A. Leclercq, C.P.M. Schuges, and C.A. Cramers, in "Science of Chromatography", F. Brunner (Ed.), J. Chromatogr. Libr., Vol. 32, Elsevier, Amsterdam, 1985, p. 55.
27. C.C. Grimm, S.W. Lloyd, L. Munchausen, Int. Lab., July 1996, p. lOA. 28. P.A. Leclercq, H.M.J. Snijders, C.A. Cramers, K.H. Maurer, U. Rapp, J. High
Resolut. Chromatogr., 12 (1989) 652. 29. M.P. Sinha, G. Gutnikov, Anal. Chem., 63 (1991) 2012. 30. M.P. Sinha, G. Gutnikov, J. MicrocoL Sep., 4 (1992) 405-410. 31. R.L. Settina, J.A. Kisinger, S. Gadheri, Eur. Spectroscopy News, 58 (1985) 16. 32. M. Barber, R.S. Bardoli, G.S. Elliot, R.D. Sedgwickand, A.N. Tyler, Anal. Chem.,
54 (1982) 645A. 33. D.A. Landi, C.L. Johlman, R.S. Brown, D.A. Weil, C.L. Wilkins, Mass Spectrom.
Rev., 5 (1986) 107. 34. M.V. Buchman, M.B. Comisarow, in "Fourier Transform Mass Spectrometry,
Evolution, Innovation and Applications", Am. Chem. Soc., Washington, DC, USA, 1987, Chapter 1.
35. R.B. Cody, J.A. Kinsinger, in "Fourier Transform Mass Spectrometry, Evolution, Innovation and Applications", Am. Chem. Soc.,Washington, DC, USA, 1987, Chapter4.
36. L. Olimpieri, P. Traldi, in "Mass Speetrometry in Biololecular Sciences", R.M. Caprioli, A. Malomi, G. Sindona (Eds.), Kluwer Academie Publishers, Dordrecht, the Netherlands, 1996, p. 198.
154
Chapter 6
Comprehensive two-dimensional
gas chromatography
SUMMARY A system for comprehensive two-dimensional GC was developed for the analysis of
complex petroleum mixtures. The separation is orthogonal such that the separation
on the second column is independent from that of the first column. In this system, a
cold trap is used to accumulate substances eluting from the first column and to
reinject them onto a jast secondary column at regu/ar time intervals. The separation
on the second column is very jast so that the separation on the first column is not
distorted. An increase in the speed of the final dimension of a multidimensional gas
chromatographic system is directly translated into an increase in peak capacity and
resolving power. Improving the separating power of the second dimension allows the
first dimension column to be shortened and reduces the total time required for a
separation of a complex mixture.
156 Chapter 6
6.1 INTRODUCTION
As pointed out by Giddings [1,2], many sample mixtures require more than one
separation mechanism to resolve the componentsof interest. For complex samples,
analytica! separation techniques of high resolving power are required for reliable
analysis. Typical examples of such samples are mineral oil derived products.
Petroleum derivates, for example, are highly complex mixtures with innumerable
sample components of varying volatility, polarity and concentration. The number of
positional isomers and enantiomers increases substantially with increasing carbon
number making it impossible to completely separate these mixtures by any known
analytica! technique.
Resolving power can be increased moderately by lengthening the column, but this
approach results in increased retention times. Schutjes demonstraled in 1982 the
detailed separation of condensales from natura! gas [3]. A complete chromatogram,
up to n-C20, obtained on a 95 m x 65 j.l.m I.D. column took more than 5 hours.
Despite its large separating power (106 theoretica! plates), its peak capacity was still
clearly insufficient to separate all components. For example, only overlapping peaks
were observed for the C15/C 16 fraction ofthe condensate.
Usually, improved chromatographic resolution can be obtained at the expense of
analysis time. If the separation power of the chromatographic system proofs to be
insufficient for the separation of complex mixtures, hyphenated techniques such as
gas or liquid chromatography coupled to mass speetrometry or Fourier transfarm
infrared spectroscopy can be used for the analysis of the samples. These hyphenated
techniques are not only powerfut analytica! techniques for the separation and
identification of unknown samples, they can also be used to increase the ability to
distinguish between different components [ 4-7]. By using multi-channel detection
overlapping components can be separated by their individual patterns. In GC/MS for
example, GC can be used to partially resolve the sample while the MS further
resolves the fractions eluting from the separation column. Since many compounds
co-elute, mass spectra are usually mixed and too complex to interpret in detail.
Mathematica! deconvolution techniques can than be used to separate non-separated
Comprehensive two-dimentional gas chromatography 157
components. With this approach, the analysis time of complex mixtures can be
reduced without sacrificing any analytica! information.
lt is commonly believed that coupled separation methods in hyphenated instruments
can be made nearly independent of each other by combining methods that are as
different as possible. [8-1 0].
In GC, the first two-stage instrument was designed in 1958 by Simmons and Snyder
[11] to analyse gasoline mixtures. The invention of the Deans [12] switch in 1968
made multidimensional GC a more practical technique in the analysis of complex
mixtures. With conventional multidimensional gas chromatographs using coupled
columns, only a fraction of the primary column eluent is sampled into the secondary
column for further separation [13]. Here the word fraction basically means a time
slice or a cut. The primary column provides sample clean up and a rough separation
while the secondary column analyses the transferred fraction in detail. Only during a
small fraction of the total analysis time on the first column, the sample is transferred
to the second column. The rest of the sample is either discarded or subjected to only
single-column separation. The so-called heart-cutting methods can provide high
resolving power but only for selected sample portions.
Multidimensional separation provides the higher peak capacity andresolving power
needed for the separation of complex mixtures. Classical coupled column
chromatography is two-dimensional in time, but is not comprehensive since only a
small (time) fraction of the sample is analysed on the second column. In
comprehensive two-dimensional gas chromatography (2D-GC), the second
dimension separation is applied to the entire eluent stream emerging from the
primary column. In this way, comprehensive two-dimensional gas chromatography
can provide simultaneously high resolving power and high speed. Liu and Phillips
[14-19] demonstrated a comprehensive 2D-GC in which the second dimeosion
separation was fast in comparison to that on the first column, in this way allowing to
sample at a frequency more or less compatible with the first separation. The second
high-speed separation generates continuously a series of second dirneusion
chromatograms. In this way, the peak capacity is increased to several thousands.
In this chapter, a system for comprehensive 2D-GC is described. Some preliminary
results of the application of comprehensive 2D-GC to the analysis of complex
158 Chapter 6
petroleum samples are presented. The experimental contiguration described here has
two major advantages over the set-up used by Phillips et al. First of all, the second
narrow-bore column is serially coupled to the tirst normal-bore column by a T-piece
which allows flow splitting. In this way, the columns that have large differences in
column inside diameter can both be operated at their respective optimallinear carrier
gas velocities. Additionally, because of the split flow, the dead volume of the
coupling union is reduced. Moreover, the split reduces the probability of overtoading
of the secondary column. Another advantage of this experimental contiguration is
that by the use of the cold trap/reinjection system very narrow input bands can be
injected onto the second column. Moreover, due to the intermediate cold trapping,
the preseparation in a short time cut is nullitied prior to starting the separation on the
second column. In this way, the true potentials ofthis narrow-bore secondary column
can be fully exploited to obtain fast and high resolution gas chromatographic
separations. It will be demonstrated that the peak capacity of this method is very high
allowing several thousands of peaks to be separated in a short time. As each sample
component has two characteristic retention times, the identitication is more reliable
than that from a one-dimensional separation.
6.2 EXPERIMENTAL
Figure 6.1 shows the design of the comprehensive two-dimensional gas
chromatograph developed within the framework of this research project. The second
column is serially coupled to the tirst column by a T -piece which allows flow
splitting. The cold trap at the head of the second column is a key feature of the
design. By using a cold trap, the clusters of peaks eluting from the tirst column are
concentrated and reinjected as series of sharp concentration pulses at the head of the
second column. The short narrow-bore second column generates a series of high
speed chromatograms as the separation in the primary column proceeds. At the end
of the chromatographic separation, the series of secondary chromatograms are
plotted in an appropriate form, resulting in a comprehensive two-dimensional gas
chromatogram.
Comprehensive two-dimentional gas chromatography
Figure 6.1
Exit I
Detector
Schematic diagram of the experimental set-up.
Injector
column
pressure regulator
159
A Fisons 8000 gas chromatograph (Fisons, Milan, Italy) equipped with an SSL 71
split/splitless injector and an EL 980 FID flame ionisation detector was used. The
original electrometer was replaced to decrease the time constant to eliminate the
contri bution of the detector electtonics to peak broadening. The injector temperature
was held at 275°C and the FID was operated at 300°C. Data acquisition was
performed using a VG Xchrom data system (VG Data Systems, Cheshire, U.K.)
which allows data registration up to 800 Hz. For our application, 100 data points
were recorded per second.
The primary column was a 25 m, 250 Jlm LD. open tubular column with a 0.23 Jlm
thick CP-Sil 19 stationary phase (Chrompack, Middelburg, the Netherlands)
containing 86% dimethyl-, 7% phenyl-, 7% cyanopropylpolysiloxane. The secondary
column was a 7 m DB-1 column (J&W scientific, Folsom, CA, USA) with a 50 JliD
LD. and a film thickness of 0.17 JliD of 100% polydimethylsiloxane. The two
columns were connected by using a T -piece. In order to obtain good transfer
efficiencies of the sample eluting from the first column to the secondary column, the
160 Chapter6
end of the first column and the beginning of the secondary column were positioned
closetoeach other in a short wide-bore column (530 J.tm LD.). The 250 Jlm LD.
column was operated at an average linear carrier gas velocity of 30 cm/sec,
corresponding with a column flow of about 8 mi/min under atmospheric conditions.
The flow of the secondary column was much smaller ( only about 0.6 mi/min). A
third column of about 9 m with a 100 Jlm I.D. was connected to the T-piece and
served as a restrictor for careful regulation of the individual column flows of the
primary and secondary column. To obtain the above mentioned column flows, the
inlet pressure of the helium carrier gas was increased up to 17 bar. To enable this
high inlet pressure, the original carrier gas inlet pressure regulator was replaced by a
Tescom 44-1100 high pressure regulator (Tescom Inc., Minnesota, USA).
Cold trapping is frequently applied in capillary GC as a preconcentration step [20-
26]. The potentials of cold trap/reinjection systems to achieve input band widths
which are compatible with narrow-bore columns have been demonstrated [27,28]. A
home-built cold trap/reinjection device is mounted inside the GC-oven at the head of
the secondary column. All parts ofthe cold trap are mounted in a polyimide block to
ensure electrical and thermal isolation. The heart ofthe device consistsof a low mass
metal capillary surrounding the narrow-bore column as tightly as possible to ensure
optimal heat transfer. The electrical leads of the transformator are connected to the
metal capillary with graphite ferrules and brass unions to ensure good electrical
contact.
Only a few millimetres of the second separation column in the trap is cooled by
means of a continuous flow (250 mi/min) of cooled helium. Helium is cooled down
by guiding it through a copper coil immersed in a Dewar vessel filled with liquid
nitrogen. The use of helium as cooling gas is preferred over nitrogen because with
nitrogen a discontinue flow is sametimes obtained due to recondensation in the
cooling circuit. During the chromatographic run, the flow of cooling gas was not
interrupted during the heating step. Since the cooling was not switched of, the next
analysis can be performed after a few seconds.
The electronic heating circuit used for fast desorption of the trapped components
consists of two stages. The first stage yields a high heating rate during a very short
period (20 msec) to achieve ultra fast thermal desorption of the trapped components.
Comprehensive two-dimentional gas chromatography 161
The second stage is started when the first stage heating is switched off and consists
of a continuous heating of the trap in order to avoid recondensation of the desorbed
components at possible cold spots in the trap. The pulse sequence is controlled by a
custom-built timer box. A more detailed description of this system can be found
elsewhere [27-30].
6.3 RESULTS
In Figure 6.2, the comprehensive separation of a naphtha mix ranging from C6 to C12
is shown. This chromatogram consists of a large number of separate high-speed
chromatograms which results from accumulated portions of sample in the cold trap
that are being releasedas sharp preconcentrated pulses into the second column. The
second column generates a high-speed chromalogram every l 0 seconds while the
cold trap accumulales the next portion of samples eluting from the primary column.
Since the selectivity mechanisms differ in the two columns, sample components co
eluting on the primary column are likely to be resolved on the secondary. The series
of chromatograms recorded reflects the resolution and peak capacity developed by
the combination of the two columns.
In Figure 6.3, a contour plot is shown of a 2D-GC chromatogram. The
chromatographic peaks are scattered around rather than along a line. If the retention
times in the two dimensions are independent of each other, the two-dimensional
chromatogram is orthogonal. Orthogonal separations are important for two reasons.
First of all, orthogonal separations allow higher peak capacities than non-orthogonal
separations. If the two dimensions of separation are orthogonal, the total peak
capacity is equal to the product of the two one-dimensional chromatographic peak
capacities. Additionally, the retention times in the two dimensions are determined by
two different and independent mechanisms. Hence, the system provides two
independent measurements of molecular properties and is therefore much more
reliable in qualitative analysis. As can be seen from this figure, the peaks are indead
scattered around rather than along a line so that a very high peak capacity is obtained
in a relatively short time.
162 Chapter 6
I
I
I I ,I
i i
i
I I
I
I
f----- - . -- - A_J! 11 '11 II llt I IJ il 1~11 liLJÜ llholllt 0 2 4 6 8 10 12
Retention time (min)
4 5 6 7
Retention time (min) Figure 6.2 Series of high-speed chromatograms of the comprehensive two-dimensionsal separation of a naphta mixture ranging from C6 to C12• GC Column 1: CP-Sil19, L = 25 m, LD.= 250 IJ.m, dr= 0.23 J.lm. GC Column 2: DB-1, L = 7 m, LD.= 50 J.lm I.D., dr= 0.17 J.lm. Experimental conditions: Tinj = 275°C, Toven = 35°C (1 min) ~ l0°C/min ~ 200°C ,Tdet = 300°C, Pi= 17 bar, flow column 1 = 8 ml/min, flow column 2 = 0.6 mllmin, reinjection delay = 10 s.
Comprehensive two-dimentional gas chromatography 163
--. Cl)
"-"' ("") (".! M s
~ -M
(s) 1 ~UIU
Figure 6.3 Contour plot of the comprehensive two-dimensional separation of a naphta mixture ranging from C6 to C12• Column 1: CP-Sill9, 25 m, 250 J.lm I.D., dr= 0.23 J.lm. Column 2: DB-1, 7 m, 50 J.liD I.D., dr= 0.17 J.liD. Tinj= 275°C, Toven= 35°C (1 min)-+l0°C/min -200°C ,Tdet= 300°C Pi= 17 atm, flow column 1 8 mllmin, flow column 2 = 0.6 mllmin, reinjection delay = 10 s.
164 Chapter 6
6.4 CONCLUSIONS
For very complex mixtures such as petroleum products, comprehensive two
dimensional gas chromatography has an obvious speed advantage. Separations of
such mixtures are limited in practice by column length; increasing length to improve
resolving power results in a time penalty. Comprehensive multi-dimensional
chromatography provides a means of high peak capacity even at high resolution in a
reasonable time. The system for comprehensive 2D-GC described here has a number
of advantages over systems previously described in literature.
6.5 REFERENCES
1. J.M. Davis, J.C. Giddings, Anal. Chem., 57 (1985) 2168. 2. J.M. Davis, J.C. Giddings, Anal. Chem., 57 (1985) 2178. 3. C.P.M. Schutjes, Ph.D. Thesis, Eindhoven University of Technology, 1983, p. 85. 4. M. Meader, A. Zilian, Chemometr. Intell. Lab. Syst., 3 (1988) 205. 5. M. Meader, Anal. Chem., 59 (1987) 527. 6. M. Meader, A.D. Zuberbuhler, Anal. Chim. Acta, 181 (1986) 287. 7. L. Roach, M. Guilhaus, Org. Mass Spectrom., 27 (1992) 1071. 8. J.C. Giddings, in "Multidimensional chromatography: Techniques and applications",
H.J. Cortes (Ed.), Marcel Dekker, New York, 1990, p. l. 9. F. Erni, R.W. Prei, J. Chromatogr., 149 (1978) 561.
10. M.M. Bushley, J.W. Jorgenson, Anal. Chem., 62 (1990) 978. 11. M.C. Simmons, L.R. Snyder, Anal. Chem., 30 (1958) 32. 12. D.R. Deans, Chromatographia, 1 (1968) 18. 13. G. Schomburg, LC-GC, 5 (1987) 304. 14. Z. Liu, J.B. Phillips, J. Chromatogr., 29 (1991) 227. 15. J.B. Phillips, Z. Liu, US Patent 5, 135, 549 (1992).
16. J.B. Phillips, Z. Liu, C.J. Venkatramani, V, Jain, in "Proc. 13th Int. Symposium on Capillary Chromatography", P. Sandra (Ed.), Rivadel Garda, Italy, Hüthig Verlag, Heidelberg, 1991.
17. J.B. Phillips, L. Zhang, C.J. Venkatramani, in "Proc. 15th Int. Symposium on Capillary Chromatography", P. Sandra (Ed.), Rivadel Garda, ltaly, Hüthig Verlag, Heidelberg, 1993, p. 744.
18. C.J. Venkatramani, J.B. Phillips, in "Proc. 15th Int. Symposium on Capillary Chromatography", P. Sandra (Ed.), Riva del Garda, Italy, Hüthig Verlag, Heidelberg, 1993, p. 885.
Comprehensive two-dimentional gas chromatography 165
19. C.J. Venkatramani, J. Zu, J.B. Phillips, Anal. Chem., 69 (1996) 1486. 20. S. Jacobsson, S. Berg, J. High Resolut. Chromatogr. Chromatogr. Comm., 5 (1982)
236. 21. B.A. Ewels, R.D. Sacks, Anal. Chem., 57 (1985) 2774. 22. B.J. Hopkins, V. Pretorius, J. Chromatogr., 158 (1978) 465. 23. C.A. Jacques, S.L. Morgan, J. Chrom. Sci., 18 (1980) 679. 24. V. Pretorius, K. Lawson, J. High Resolut. Chromatogr. Chromatogr. Comm., 9
(1986) 278. 25. J.A. Settlage, W.G. Jennings, J. High Resolut. Chromatogr. Chromatogr. Comm., 3
(1980) 146. 26. B.V. Burger, Z. Munro, J. Chromatogr., 370 (1986) 449. 27. A. van Es, J. Janssen, R. Bally, C.A. Cramers, J.A. Rijks, J. High Resolut.
Chromatogr. Chromatogr. Comm., 10 (1987) 273. 28. A. van Es, Ph.D. Thesis, Eindhoven University of Technology, the Netherlands,
1990. 29. H.M.J. Snijders, J.P. Rijks, A.J. Bombeeck, J.A. Rijks, in "Proc. 13th Int.
Symposium on Capillary Chromatography", P. Sandra, M. Lee (Eds.), Baltimore, Maryland, USA, Hüthig Verlag, Heidelberg, 1992, p. 46.
30. J.P. Rijks, H.MJ. Snijders, AJ. Bombeeck, J.A. Rijks, in "Proc. 13th Int. Symposium on Capillary Chromatography", P. Sandra, M. Lee (Eds.), Baltimore, Maryland, USA, Hüthig Verlag, Heidelberg, 1992, p. 54.
166
Summary
Chromatographic separation originates from a difference in migration veloeities of
compounds in the separation column. Chromatographic band broadening results from
inequalities in the migration veloeities of the molecules of a solute in the column.
Under resolution normalised conditions, keeping the retention factor and the
separation factor constant, the analysis time is proportional to the ratio of the plate
height over the average linear carrier gas velocity. To decrease this ratio, the radial
mass transfer has to be increased. The time required for radial equilibration is a
complex function of amongst others, solute diffusion coefficients in the mobile and the
stationary phase, column dimensions, flow profile ofthe mobile phase, etc.
In chapter 2, various approaches to reduce the analysis time in capillary gas
chromatography (GC) are summarised. Because of their high permeability, open
tubular columns provide the highest column efficiencies. Thin-film columns are
favoured over thick-film columns in most applications. For thin-film open-tubular
columns, it is demonstraled that the most efficient means to reduce the analysis time is
to use columns with low inside diameters and hydrogen as carrier gas. For these
columns, the gain in separation speed by operating the column under vacuum outlet
conditions is negligible. Vacuum outlet operation can be very attractive when only low
plate numbers are required as for short or wide-bore columns. The use of coiled
columns or turbulent flow conditions in open tubular systems or columns packed with
small-size particles offers only limited practical advantages. Moreover, each of these
options require extremely high inlet pressures. Besides affecting the analysis time, the
varlation of the column dimensions also strongly affects the minimum detectable
amount, the column loadability, the minimum detectable concentration and the
requirements imposed upon the instrumental design. The instrumental requirements are
the more critica! the smaller the inside diameter. The lack of compatible
168
are the more critical the smaller the inside diameter. The lack of compatible
instrumentation has hindered the wide-spread acceptance of narrow-bore columns. In
this thesis, various injection systems and detectors are evaluated for hyphenation to
narrow-bore columns.
In literature, various injection devices able to produce very narrow input bands
compatible with narrow-bore columns have been described. With these injection
systems, the sample volume actually introduced onto the column is extremely small.
Despite the attractive minimum detectable amount offered by narrow-bore columns,
the - in daily practice more relevant minimum detectable concentration - is therefore
still too high for many practical applications. In chapter 3, the possibilities and
limitations of several non-splitting injection techniques as hot splitless, cold splitless,
on-column and large volume sampling, are studied. It is demonstrated that despite the
low column flow, the splitless time required to obtain quantitative transfer yields for a
splitless injection onto narrow-bore columns can be surprisingly low, if liners with
small inside diameters are used. For hot splitless injections, peak focusing is limited
and severe discrimination is observed. With regard to these parameters much better
results are obtained with cold splitless and on-column injections. Additionally, the
possibilities of using normal-bore retention gaps in combination with narrow-bore
separation columns are demonstrated. For cold splitless injections, large sample
volumes (up to a few J.tl) can be injected. The sample volume can even be further
increased by performing solvent elimination prior to splitless transfer of the sample
components to the column. For on-column injections, the use of a normal-bore
retention gap allows direct sample introduetion onto the column. Good peak focusing
is obtained as long as the flooded zone can be accommodated by the retention gap.
Unfortunately, the conneetion between the retention gap and the separation column
will result in some peak tailing, even if very low dead volume unions are used. The
only way to overcome this problem is to use an additional cold trap system at the
column inlet.
High demands are also imposed on the detection system when narrow-bore columns
are used. The detection device has to be very sensitive to preserve an acceptable
working range. Additionally, the chromatographic peak broadening of narrow-bore
columns is very small. As a consequence, the detector must have a low time constant
Summary 169
an electron capture detector for high-speed separations are evaluated. If this detector is
used in combination with narrow-bore columns, the make-up flow has to be
sufficiently high to avoid tailing caused by the large detection cell. For electron
capture detectors, the make-up flow actively participates in the detection mechanism.
For this reason it is important to evaluate the sensitivity at higher make-up flow rates.
It was found that even at the high flow rates required to minimise peak tailing, a very
good sensitivity could be obtained.
In chapter 5, different types of mass analysers, including the ion trap, sector
instrument, reflectron time-of-flight and the orthogonal acceleration time-of-flight
mass spectrometer, have been evaluated as mass spectrometric detectors in
combination with narrow-bore capillaries. The potentials and limitations of all these
mass spectrometers are discussed in detail. Special emphasis is paid to the maximum
scan speed, detection limits, mass spectrometric resolution and the quality of the mass
spectra obtained. With rnass-scanning instruments including the quadrupole, the ion
trap and the sector machine, the acquisition rates are limited to a maximum of some 20
scans per second. This maximum scan rate is sufficient for chromatographic
separations in the minute range. With non rnass-scanning techniques, such as the time
of-flight mass spectrometer and a microchannel plate mass spectrometer, a complete
spectrum can be recorded in 100 J.I.Sec or less. The actual scan speed unfortunately is
still limited because the computer time required for data transfer and data handling is
far too long. In the future faster computers will enable scan speeds over a few hundred
spectra per second. With most of the mass spectrometers described here, the detection
limits are in the ]ow pg range. Apart from having a high sensitivity, it is also important
to have a high mass resolution. With each of the instruments considered, a compromise
has to bemadebetween sensitivity, mass resolution (selectivity) and scanning rates. In
our opinion, the orthogonal acceleration time-of-flight mass analyser currently offers
the best possibilities as a mass spectrometer for hyphenation to high-speed narrow
bore capillary columns. Several suggestions are formulated to further improve the
performance of this instrument.
In chapter 6, the applicability of high-speed GC in comprehensive two-dimensional
gas chromatography is demonstrated for the analysis of very complex mixtures such as
petroleum products. Separations of such mixtures in practice are limited by the
170
required column length. Increasing the column length to imprave the chromatographic
resolving power results in a proportional time penalty. It is demonstrated that
comprehensive multi-dimensional chromatography provides a high peak capacity in a
reasonable time.
Samenvatting
Chromatografische scheiding is het gevolg van het verschil van de migratiesnelheid
van de componenten in de scheidingskolom. De chromatografische scheiding wordt
tegengewerkt door verschillende migratiesnelheden van de moleculen van een
component hetgeen bandverbreding veroorzaakt. Bij een gegeven chromatografische
resolutie en een constante capaciteits· en scheidingsfactor, is de analysetijd
evenredig met de verhouding van de schotelhoogte en de gemiddelde lineaire
snelheid van de mobiele fase. Om deze verhouding kleiner te maken, en dus de
analysetijd te reduceren, moet het radiale massatransport in de kolom verhoogd
worden. De tijd nodig voor het instellen van radiaal evenwicht is een complexe functie
van verschillende parameters zoals de diffusiecoëfficiënt van een component in de
mobiele en de stationaire fase, de kolomdimensies, het stromingsprofiel van de
mobiele fase in de kolom, etc.
In hoofdstuk 2 worden de verschillende mogelijkheden voor de reductie van de
analysetijd in gaschromatografie (GC) beschreven. Open capillaire kolommen kunnen
door hun hoge permeabiliteit hoge chromatografische scheidingsefficiëntie opleveren.
Voor de meeste toepassingen is het gebruik van dunne film kolommen de beste
oplossing. Het gebruik van capillairen met een kleine inwendige diameter met
waterstof als draaggas is de meest geschikte methode om de analysetijd te reduceren.
Bij dit type kolommen is de winst in analysetijd door de kolomuitgang aan te sluiten
op een vacuumsysteem verwaarloosbaar. Vacuum GC levert alleen een interessante
tijdswinst op, indien slechts een beperkt aantal schotels nodig is. Het gebruik van
geometrisch vervormde kolommen of turbulente stroming in open capillairen of met
zeer kleine deeltjes gepakte kolommen, biedt slechts beperkte mogelijkheden voor de
reductie van de analysetijd. Bovendien vormen de hoge inlaatdrukken voor deze
systemen extra problemen voor de toepasbaarheid. De kolomdimensies hebben niet
172
alleen invloed op de analysetijd, maar ook worden de minimaal detecteerbare
hoeveelheid, de kolombelaadbaarheid en de minimaal detecteerbare concentratie
hierdoor beïnvloed. Bovendien worden steeds hogere eisen gesteld aan de benodigde
instrumentatie naarmate de kolomdiameter kleiner wordt. De bandverbreding door het
injectie- en detectiesysteem moet voldoende klein zijn om de chromatografische
resolutie niet of nauwelijks te beïnvloeden. Tevens moet de gevoeligheid en de
selectiviteit van de detector voldoende zijn. Het gebrek aan compatibele instrumentatie
heeft het algemeen gebruik van capillairen met een kleine inwendige diameter
verhinderd. In dit proefschrift worden verschillende injectie- en detectiesystemen
gekoppeld met deze capillairen en de mogelijkheden en beperkingen van deze
systemen worden uitvoerig bestudeerd.
In de literatuur zijn al verschillende injectiesystemen beschreven, waarvan de
injectiebandbreedtes voldoende klein en compatibel zijn met de kleine
chromatografische bandverbreding van nauwe capillairen. Met deze injectiesystemen
echter kunnen slechts zeer kleine hoeveelheden monster op de kolom geïnjecteerd
worden. Ondanks de interessantere minimaal detecteerbare hoeveelheid die behaald
kan worden met deze kleine capillairen, is de voor de praktijk meer relevante minimaal
detecteerbare concentratie te hoog voor vele applicaties. In hoofdstuk 3 worden de
mogelijkheden en beperkingen van injectietechnieken voor grotere monster
hoeveelheden geëvalueerd. Er wordt aangetoond dat ondanks de lage kolomstroming
voor nauwe capillairen, de "splitless"-tijd die nodig is voor kwantitatieve opbrengst
voor splitless injecties verrassend klein kan zijn, indien "liners" gebruikt worden met
een kleine inwendige diameter. Voor "hot splitless" injecties is het focusserende effect
van de chromatografische pieken beperkt en treedt verder een hoge graad van
discriminatie en/ of thermische degradatie op. In dit opzicht bieden "cold splitless" en
"on-column" injecties betere resultaten. Tevens worden de mogelijkheden van deze
injectietechnieken gebruik makend van een "retention gap" gekoppeld met kolommen
met een kleine inwendige diameter bestudeerd. Voor "cold splitless" injecties kunnen
succesvol grotere volumina (tot enkele microliters) geïnjecteerd worden. Door
eliminatie van het oplosmiddel kan het injectievolume nog verder opgevoerd worden.
Voor "on-column" injecties laat het gebruik van een "retention gap" directe
monsterintroductie toe op de kolom. Goede focusserende effecten worden verkregen
Samenvatting 173
zolang de volledige "solvent"film past in de "retention gap". Een probleem bij het
gebruik van een "retention gap11 is dat de chromatografische pieken altijd iets "tailing"
vertonen, veroorzaakt door het koppelsysteem. De enige manier om dit probleem te
voorkomen is de plaatsing van een koude val aan het begin van de analytische kolom.
Er worden tevens hoge eisen aan het detectiesysteem gesteld. Het gebruikte
detectiesysteem moet voldoende gevoelig zijn om een interessant werkgebied over te
houden. Bovendien moet de tijdsconstante van de detector voldoende klein zijn om de
chromatografische resolutie te behouden. In hoofdstuk 4 worden de mogelijkheden
van "electron capture" detectie voor snelle chromatografische scheidingen
geëvalueerd. Indien deze detector gekoppeld wordt met capillairen met een kleine
inwendige diameter, moet de make-up flow voldoende hoog zijn om "tailing" te
minimaliseren. Bij deze detector speelt het make-up gas een essentiële rol in het
detectiemechanisme. Hierdoor is het van belang om de invloed van deze hogere
debieten op de gevoeligheid te evalueren. Het is aangetoond dat ook bij een groot
debiet een hoge gevoeligheid kan worden verkregen.
In hoofdstuk 5 worden verschillende massaspectrometers (MS) zoals de "ion trap", een
sectorinstrument, een "reflectron time-of-flight" MS en een "orthogonal acceleration
time-of-flight" MS ( oa-TOF MS) geëvalueerd als detectoren voor snelle
chromatografische scheidingen. Hierbij is vooral aandacht besteed aan de
scansnelheid, detectielimieten, massaspectrometrische resolutie en de kwaliteit van de
spectra. Met massa-scannende instrumenten zoals de "quadrupole11, de "ion trap" en
een sectorinstrument, is de scansnelheid beperkt tot ongeveer 20 spectra per seconde.
Deze scansnelheid is voldoende voor chromatografische scheidingen die enkele
minuten in beslag nemen. Met niet-scannende massaspectrometers, zoals de "time-of
flight" MS en de "microchannel plate" MS, kan een volledig spectrum geregistreerd
worden in minder dan 100 microseconden. Dit betekent echter niet dat er meer dan
1000 spectra kunnen worden geregistreerd per tijdseenheid. De scansnelheid die tot nu
toe behaald is, wordt beperkt door de traagheid van de dataoverdracht naar de
computer. Met de meeste massaspectrometers die hier beschreven zijn, worden
detectielimieten behaald op het lage picogramniveau. Naast een hoge gevoeligheid, is
ook een goede massaresolutie van belang. Hierdoor kan extra selectiviteit
geïntroduceerd worden. Bij elke MS moet een compromis gezocht worden tussen
174
gevoeligheid, selectiviteit ( massaresolutie) en scansnelheid. Bij de huidige stand van
de techniek biedt de oa-TOF MS de meeste mogelijkheden als massaspectrametrische
detector voor snelle chromatografische scheidingen. In hoofdstuk 5 worden tevens een
aantal suggesties gegeven om de mogelijkheden van deze MS verder te verbeteren.
In hoofdstuk 6 wordt de toepasbaarheid van snelle GC geïllustreerd voor de analyse
van complexe mengsels zoals petroleumprodukten door middel van 11comprehensive"
twee-dimensionale gas chromatografie. Complexe scheidingen worden in de praktijk
beperkt door de vereiste kolomlengte. Verlenging van de kolom heeft echter een veel
langere analysetijd tot gevolg. Het is aangetoond dat "comprehensive" twee
dimensionale gaschromatografie een hoge piekcapaciteit biedt die de analyse mogelijk
maakt van zeer complexe mengsels binnen een aanvaardbare tijd.
List of symbols
A area response
B magnetic field strength
BM,o 2DM,o
Co minimum detectable concentration
c~ minimum detectable concentration for a concentration sensitive
detector
egt minimum detectable concentration fora mass flow sensitive detector
CM,o resistance to mass transfer in the mobile phase at column outlet pressure
Cs resistance to mass transfer in the stationary phase
de column inside diameter
dcoil coil radius
dp partiele diameter of the packing material
D term descrihing the extra band broadening effects
D A,B binary diffusion coefficient of the analyte in the mobile phase
DM,o diffusion coefficient in the mobile phase under outlet conditions
DR,o radial dispersion coefficient
Ds diffusion coefficient in the stationary phase
DsF secondary-flow dispersion coefficient
De Dean number
f1 pressure correction factor after Giddings
f2 pressure correction factor after James-Martin
F atm flow under ambient conditions
F det detector flow at detection pressure and temperature
F o column flow under outlet conditions
176
h
H
Hextra
k
K
L
mx
m/z
N
p p
Patm
Pdet
Pi
Po p
Q
Qo
Qg Qjf
Qs
r
reduced plate height
plate height
increased plate height due to instromental band broadening
retention factor
distribution constant
column length
molecular weight of compound X
mass-to-charge ratio
plate number
number of plates required for a given chromatographic resolution
local pressure
average pressure ofthe column
ambient pressure
pressure in the detector cell
column inlet pressure
column outlet pressure
relative pressure
sample amount injected onto the column
minimum detectable amount
minimum detectable amount for a concentration sensitive detector
minimum detectable amount for a mass flow sensitive detector
sample capacity of the column
radius of the magnet
mass spectrometric resolution
noise of the detector
peak resolution
Reynolds number
sensitivity for a concentration sensitive detector
sensitivity for a mass flow sensitive detector
Schmidt number
List of Symbols
tM
tR
Tatm
Tdet
Tinj
Toven
Uo,opt
u
w a
1(
K'
À!
Po
cr
crchrom
mobile-phase hold-up time
total retention time
ambient temperature
temperature of the detector
temperature of the injector
temperature of the GC oven
average linear carrier gas velocity
carrier gas velocity under column outlet conditions
carrier gas velocity under column outlet conditions for maximum
separation efficiency
average linear carrier gas velocity for maximum separation efficiency
ion extraction voltage
detector cell volume
sample volume introduced onto the column
working range
separation factor
phase ratio
mobile phase viscosity
difference in total relention time between two adjacent peaks
correction factor for the turtuosity of gas channels in packed columns
velocity profile factor for the separation column
velocity profile factor for the detector
aspect ratio
geometry factor
Time constant for radial diffusion
gas density of the mobile phase under outlet conditions
standard deviation of a chromatographic peak
average standard deviation of two adjacent peaks
standard deviation due to chromatographic dispersion
standard deviation due to connecting tubes, frits, etc.
standard deviation due to the detector
177
178
cre1 standard deviation due to the fini te response time of the detector
electtonics and the data acquisition system
Oextra total standard deviation due to instrumentation
Oïnj standard deviation due to the injector
Otot total standard deviation of a chromatographic peak
ux diffusion volume
AGC
API
ARC
Cl
Da
ECD EI
FEC FID
FTMS
FWHM
GC
HCB HSGC I.D.
MCP
MDA
MDC
MS
oa-TOFMS
PCB PID
PTV
RF
SEM
List of abbreviations
automatic gain control
atmospheric pressure ionisation
automatic reaction control
chemical ionisation
Dalton
electron capture detector
electron impact
filament emission current
flame ionisation detector
F ourier Transform mass spectrometer
full width half maximum
gas chromatography
hexachlorobenzene
high-speed gas chromatography
inside diameter
microchannel plate
minimum detectable amounts
minimum detectable concentration
mass spectrometer
orthogonal acceleration time-of-flight mass spectrometer
polychlroronatedbiphenyl
photoionization detector
temperature programmabie vaporizer
radio frequency
secondary electron multiplier
180
SFE supercritical fluid extraction
SIM selected ion monitoring
SSL split/splitless
TCD thermal conductivity detector
TDC time-to-digital converter
TMS tetramethyl-silyl
TOF time-of-flight
TV target value
VIT variabie ionisation time
voc volatile organic compounds
2G-GC two dimensional gas chromatography
Dankwoord
Het einde van dit boekje is voor mij de ideale gelegenheid om vele mensen te danken
voor hun bijdrage bij het tot standkomen van dit werk.
Op de eerste plaats wil ik mijn promotor Prof.dr.ir. Carel Cramers danken voor het
vertrouwen die hij in mij stelde en de mogelijkheid die hij mij bood om te
promoveren. Speciale dank ben ik verschuldigd aan Hans-Gerd lanssen die mij de
kunst van het chromatograferen bijbracht en waarover we veelvuldig discussies
hebben gevoerd. Aan jou kon ik mijn wetenschappelijke bevindingen kwijt en kon ik
in gedachten "stoeien" over allerlei nieuwe experimenten die niet ontkwamen aan
jouw "superkritische" opmerkingsgeest. Dankzij de vele contacten van Piet Leclercq
heb ik enkele keren de mogelijkheid gekregen om op verschillende buitenlandse
laboratoria te werken en kennis en ervaring uit te wisselen op het gebied van
GC/MS. Tevens wil ik Henri Snijders danken voor de vele discussies over Willem II
en snelle GC, hoewel we het in de meeste gevallen niet eens waren. Mark van
Lieshout ken ik zeer erkentelijk voor de pratische assistentie. Anton Bombeeck, die
voor elk probleem wel een Multilabje kon verzinnen, jij kon met volle teugen
genieten (en tevens menig andere aanwezigen) als ik er in slaagde jouw schakelingen
in rode rook te laten opgaan. Bij Harrie Maathuis kon ik voor alles en nog wat
terecht: voor kabels en computers, e-mail en belastingsinfo. Huub van Leuken, jij
toverde met uiterste precisie enkele kunstwerkjes op je draaibank. Denise Tjallema
dank ik voor de secretariële steuntjes in de rug en het vernederlandsen van mijn
Belgische uitdrukkingen. Op Hans van Rijsewijk kon ik rekenen voor de aanschaf
182
van lab-benodigdheden en de administratieve en financiële afhandelingen van de
reizen die ik mocht maken tijdens mijn promotiestudie. Han Martens was de
alwetende dokter voor overspannen computersystemen.
Een bijzonder woord van dank gaat uit naar Alex Scholten met wie ik het "AiO
bestaan" kon delen: samen naar de Mensa) de frustaties van het onderzoek, maar ook
af en toe eens ontspannen met een potje snooker. Tevens wil ik mijn overige
collega's danken) in het bijzonder Hans Mol, Manuel Mertens, Xianwen Lou, Hai
Pham Tuan en Gerard Rutten, voor de collegiale) aangename sfeer en de vele
gedachtenwisselingen die we voerden bij het nuttigen van "hoegjes" of "wittekes".
Additionally, I like to thank some other people who are not a memher of our
Laboratory in Eindhoven.
First of all, I want to thank Dr. F. Munari and Dr. S. Tretianu from CE Instruments
(Milan, Italy) and René Leenders from Interscience (Breda, the Netherlands) for the
discussions and the instrumentation they supplied for my Ph.D. research.
A special word ofthanks to Professor H. WoBnik (University ofGiessen, Germany)
and Professor M. Guilhaus (University ofNew South Wales) for inviting me to work
in their laboratodes which gave me the opportunity to exchange a lot of knowledge
with their students and to visit their countries.
I want to thank Huub van Cruchten and Jeff Brown for their fruitful discussions and
their advise about several GC/MS experiments.
Tenslotte) ben ik Anja grote dank verschuldigd voor de vele tekeningen, posters en
dia's die je voor me maakte. Met veel knip-, plak-, kunst- en vliegwerk en vooral
eindeloos geduld heb jij de lay-out van dit proefschrift verzorgd. Jij bent voor mij
een echte steun voor alles, tijdens alles, en vooral na alles.
Mijn ouders wil ik graag danken omdat zij het voor mij mogelijk hebben gemaakt
een academische studie te volgen. Jullie interesse hebben mij gesteund en
gestimuleerd tijdens mijn studie.
Curriculum Vitae
Peter Van Y sacker is geboren op 28 oktober 1967 te Roeselare in het Belgische
Vlaanderen. In 1985 rondde hij zijn middelbare studie (Moderne Humaniora, richting
wetenschappelijk B), af aan het Klein Seminarie te Roeselare. Daarna vertrok hij naar
de Katholieke Universiteit Campus Kortrijk en behaalde in 1987 het
kandidaatsdiploma in de wetenscahppen, groep scheikunde. Hij vervolgde zijn studie
aan de Katholieke Universiteit Leuven en behaalde in 1989 het licentiaatsdiploma in
de scheikundige wetenschappen in de richting organische chemie met een
afstudeeropdracht in de polymeerchemie onder leiding van Prof.dr. M. Van Beylen. In
1990 slaagde hij aan dezelfde universiteit voor het baccalaureaatsexamen in de
bedrijfseconomie.
Na enige bedrijfservaring bij Dow Benelux in Terneuzen werkzaam was als analist op
het laboratorium voor de kwaliteitscontrole van de produktie van olefinederivaten,
startte hij in mei 1992 hij met zijn promotieondrzoek op het gebied van snelle
capillaire gas chromatografie op het laboratorium voor instrumentele analyse van de
Technische Universiteit in Eindhoven, onder leiding van Prof.dr.ir. C.A. Cramers.
184
Bibliografie
High-speed GC/MS using narrow-bore columns and ion trap detection, P. G. Van
Ysacker, J G.M Janssen, HMJ Snijders, P.A. Leclercq, C.A. Cramers and HJM
van Cruchten, J. MicrocoL Sep., 5 {1993) 413-419 and Proc. 15th International
Symposium on Capillary Chromatography, Riva del Garda, Italy, Hüthig Verlag,
Heidelberg, 1993, p. 988-997.
A high-speed gas chromatograph coupled to a time-of-flight mass analyzer, H
Wollnik, R. Becker, H Gotz, A. Kraft, H Jung, C.-C. Chen, P.G. Van Ysacker,
JG.M Janssen, HMJ Snijders, P.A. Leclercq and C.A. Cramers, Int. J. Mass
Spectrom. Ion Processes, 130 (1994) L7-Lll.
Comparison of different mass spectrometers in combination with high-speed narrow
bore capillary gas chromatography, P.G. Van Ysacker, JG.M Janssen, HMJ
Snijders P.A. Leclercq, H WoUnik and C.A. Cramers, Proc. 16th International
Symposium on Capillary Chromatography, Riva del Garda, Italy, Hüthig Verlag,
Heidelberg, 1994, p. 785-796.
Electron capture detection in high-speed narrow-bore capillary gas chromatography:
Fast and sensitive analysis of PCBs and pesticides, P.G. Van Ysacker, JG.M
Janssen, HMJ Snijders and C.A. Cramers, J. High Resol. Chromatogr., 18 (1995)
397-402 and Proc. 16th International Symposium on Capillary Chromatography,
Rivadel Garda, Jtaly, Hüthig Verlag, Heidelberg, 1994, p. 1359-1370.
Trace analysis in high-speed narrow-bore capillary gas chromatography. Preliminary
results of hot and cold splitless injections, HMJ Snijders, JG.M Janssen, P.G. Van
186
Ysacker and C.A. Cramers, Proc. 16th International Symposium on Capillary
Chromatography, Rivadel Garda, Italy, Hüthig Verlag, Heidelberg, 1994, p. 1137-
1147.
High-speed narrow-bore capillary gas chromatography in combination with a fast
double focusing mass spectrometer, P.G. Van Ysacker, J. Brown, J.G.M Janssen,
P.A. Leclercq, A. Phillips and C.A. Cramers, J. High Resolut. Chromatogr., 18
(1995) 517-524.
High-speed narrow·bore capillary gas chromatography in combination with
orthogonal acceleration time-of.flight mass spectrometric detection, P.G. Van
Ysacker, M Guilhaus, L. Roach, V. Mlinsky, J. G.M Janssen, P.A. Leclercq and C.A.
Cramers, Proc. 18th International Symposium on Capillary Chromatography, Riva
del Garda, Italy, Hüthig Verlag, Heidelberg, 1996, p. 1496.
Trace analysis in narrow-bore capillary gas chromatography: Hot and cold splitless,
on-column and large volume sampling, P. G. Van Ysacker, H.MJ. Snijders, J. G.M
Janssen and CA. Cramers, Proc. 18th International Symposium on Capillary
Chromatography, Rivadel Garda, Italy, Hüthig Verlag, Heidelberg, 1996, p. 638.
Non-splitting injection techniques for narrow-bore capillary gas chromatography,
P.G. Van Ysacker, H.MJ. Snijders, J.G.M Janssen and C.A. Cramers,submitted for
publication.
STELLINGEN
1) Bij het bespreken van de verschillende mogelijkheden voor het versnellen van een
chromatografische analyse is het onjuist de invloed van de kolomdiameter buiten
beschouwing te laten.
A. Peters, M Klemp, L. Puig, C. Rankin, R. Sacks, Analyst, 166 (1991) 1313.
2) Grob's bewering dat het onmogelijk is om "splitless" injecties uit te voeren op
kolommen met een inwendige diameter kleiner dan 200 IJID, is niet correct.
K. Grob, in "Classica/ split and splitless injection in capillary gas
chromatography", Hüthig Verlag, Heide/berg, 1986, 132.
3) Om succesvol "splitless" of "on-column" injecties uit te voeren van
monsterhoeveelheden van 1 Jll of meer op capillairen met een kleine inwendige
diameter (LD. :s;IOO !Jm), is het gebruik van een "retention gap" noodzakelijk.
Dit proefschrift, hoofdstuk 3.
4) De "electron capture" detector gedraagt zich onder bepaalde experimentele
condities als een massastroom-gevoelige detector.
Dit proefschrift, hoofdstuk 4.
5) De belangrijkste belemmering voor snelle capillaire gaschromatografie gekoppeld
met massaspectrometrie, is het ontbreken van voldoend snelle computers.
Dit proefschrift, hoofdstuk 5.
6) Bij de huidige stand van zaken is de "orthogonal-acceleration time-of-flight"
massaspectrometer de meest geschikte massaspectrometer voor koppeling met
snelle capillaire gaschromatografie.
Dit proefschrift, hoofdstuk 5.
7) Indien ionen tijdelijk worden opgeslagen in een ionenbron van een
massaspectrometer, kunnen verstoorde spectra verkregen worden.
S.A. McLuckey, G.L. Glish, K.G. Asano, G.J. Van Berkel, Anal. Chem., 60 (1988)
2314.
8) De introductie van de multicapillaire kolom zal het gebruik van snelle capillaire
gaschromatografie in een stroomversnelling brengen.
9) De nomenclatuur voor de chromatografie van de International Uni on of Pure and
Applied Chemistry (IUP AC) is niet eenduidig.
L.S. Ettre, Pure & Appl. Chem., 65 (1993) 819.
I 0) In het gebied van etherische oliën, in het bijzonder sesquiterpenen, is het gevaarlijk
alleen te vertrouwen op GC retentieindices en massaspectra.
Lê Thanh, Nguyên Xuán Düng, Ange Bighelli, Joseph Casanova, Piet A. Leclercq,
submittedfor publication in Spectroscopy (Amsterdam).
11) Het gebruik van de term "replaceable" bij lineaire polyacrylamide matrices in
capillaire zone electroforese is misleidend.
MC. Ruiz-Martinez, J. Berka, A. Belenkii, F. Foret, A.W. Miller, B.L. Karger.,
Anal. Chem., 65 (1993) 2851.
12) Indien het vroegere Joegoslavië over de olierijkdommen beschikte van Koeweit,
was het conflict allang opgelost.
13) Vrouwe Justitia mag dan wel geblinddoekt zijn, ze beschikt over zeer verfijnde
luisterapparatuur.
Peter Van Ysacker
Eindhoven, 31 oktober 1996