detectors in gas chromatography

- minimum instantaneous concentration CmaX - maximum instantaneous concentration c' - instantaneous concentration in the effective volume of the detector C, - heat capacity at constan

DETECTORS IN GAS CHROMATOGRAPHY

JOURNAL OF CHROMATOGRAPHY LIBRARY

Volume 1 Chromatography of Antibiotics

by G H Wagman and M J Weinstein

Volume 2 Extraction Chromatography

edited by T Braun and G Ghersini

Volume 3 Liquid Column Chromatography A Survey of Modern Techniques

and Applications

edited by Z Deyl, K Macek and J Janak

Volume 4 Detectors in Gas Chromatography

by J SevCik

DETECTORS

JIkf SEVcfK

Department of Analytical Ciieunistrj., Ciiarles University, Prague

Distribution of this book is being handled by the following team of publishers

for the U.S.A and Canada AMERICAN ELSEVIER PUBLISHING COMPANY, INC

52 Vanderbilt Avenue New York New York 10017 for the East European Countries, China, Northern Korea, Cuba,

Vietnam and Mongolia SNTL, PUBLISHERS OF TECHNICAL LITERATURE

Prague for all remaining areas ELSEVIER SCIENTIFIC PUBLISHING COMPANY

335 Jan van Galenstraat

P 0 Box 211, Amsterdam, The Netherlands

Library of Congress Cataioging in Publication Data

Translation 0 KAREL STULfK 1976

ALL RIGHTS RESERVED NO PART OF THIS PUBLICATION

MAY BE REPRODUCED, STORED I N A RETRIEVAL SYSTEM,

OR TRANSMITTED I N ANY FORM OR BY ANY MEANS, ELECTRONIC, MECHANICAL, PHOTOCOPYING, RECORDING OR OTHERWISE,

WITHOUT PRIOR WRITTEN PERMISSION OF THE PUBLISHERS

Elsevier Scientific Publishing Company,

Jan van Galenstraat 335, Amsterdam

PRINTED IN CZECHOSLOVAKIA

List of symbols 9

Preface 12

1 Introduction 1 5 1.1 Concentration distribution of the eluted substance at the column outlet 1.2 Detector signal

1.2.1 Detector response

1.3 Effect of the measuring device on signal changes 1.4 Sample injection

1.5 Parameters characterizing detectors

1.5.1 Sensitivity

1.5.2 Detector linearity

1.5.3 Linear dynamic range

1.5.4 Lowest detectable amount

1.5.5 Detector selectivity

1.6 Literature

17

22

23

24

28

30

31

32

33

34

36

37

2 The Thermal Conductivity Detector (TCD) 39

2.1 Detection mechanism 39

2.2 TCD signal 42

2.2.1 TCD background current 43

2.2.2 TCD response 43

2.3 Effect of experimental parameters on the magnitude.and shape of the TCD signal 2.3.1 Carrier gas 46

2.3.2 Construction of the TCD 47

2.3.2.1 Sensor heating voltage

2.3.2.2 Sensor parameters 49

2.3.2.3 Cell geometric constant

2.3.2.4 Temperatures of the sensor and the cell walls

2.3.2.6 Measuring circuits 5 5 2.4 Applications of the TCD

2.5 Literature 57

46

49 5 1 52 52 56 2.3.2.5 Time constant of the TCD

3 Ionization Detectors 59

3.1 Physical principles of the detection 59

3.1.1 The collision 60

3.1.2 Effect of the electric field intensity 62

3.2 Ionization energy sources 65

3.3.1 The slow-down mechanism 68

3.3.2 Recombination 69

3.3.3 Background current of the ionization detector 71

3.4 Literature 71

3.3 Reactions in the ionization detector 68

4 The Electron Capture Detector ( E C D ) 72

4.1 Detection mechanism 72

4.2 ECD signal 74

4.2.1 ECD background current 74

4.2.2 ECD response 76

4.2.2.1 Linearity and linear dynamic range 77

4.2.2.2 Sensitivity and selectivity of the ECD

4.3 Experimental conditions affecting the E C D signal 79

4.3.1 Carrier gas 79

4.3.2 Construction of the ECD 80

4.4 Applications of the ECD 82

4.5 Literature 85

77 5 The Flame Ionization Detector (FID) 87

5.1 Detection mechanism 87

5.2 FID signal 91

5.2.1 FID background current 92

5.2.2 FID response 92

5.2.2.1 Linear dynamic range and linearity of the FID 94

5.2.2.2 Sensitivity and selectivity of the FID 95

95

5.3.1 Gas flow-rate 95

5.3.2 Geometry of the FID 97

5.4 FID applications 101

5.5 Literature 102

5.3 Experimental conditions affecting the magnitude and character of the FID signal 6 The Thermionic Detector Using an Alkali Metal Salt (TIDAj 105

6.1 Detection mechanism 106

6.2 TlDA signal 109

6.2.1 TlDA background current 110

6.2.2 TIDA response 112

6.2.2.1 Linearity and linear dynamic range of the TlDA 115

6.2.2.2 Sensitivity and selectivity of the TIDA 116

6.3 Effect of the experimental conditions on the magnitude and character of the TIDA signal 117 6.3.1 Gas flow-rate 117

6.3.2 Detector geometry 118

6.4 TIDA applications 120

6.5 Literature 121

7 The Photoionization Detector (PID) 123

7.1 Detection mechanism 123

7.2 PID signal i25

7.2.1 PID background current i27

7.2.2 PID response 127

128

7.3.1 Carrier gas 128

7.3.2 Geometric arrangement of the PID 129

7.3.2.1 Discharge compartment 129

7.3.2.2 Detection compartment 130

7.4 PID applications 131

7.5 Literature 133

7.3 Effect of the experimental conditions on the PID signal 8 The Helium Detector (HeD) 133

8.1 Detection mechanism

8.2 HeD signal

8.2.1 HeD background current

8.2.2 HeD response

8.2.2.1 Linearity and linear dynamic range of the HeD 8.2.2.2 Sensitivity and selectivity of the HeD

8.3 Effect of experimental conditions on the HeD signal

8.3.1 Carrier gas

8.4 HeD applications

8.5 Literature

8.3.2 Construction of the helium and argon detectors

133

135

135

136

139

140

140

140

141

143

144

9 The Flame Photometric Detector (FPD) 145

9.1 Detection mechanism 145

9.2 FPD signal 149

9.2.1 FPD background current 151

9.2.2 FPD response 152

9.2.2.1 Linearity and linear dynamic range of the FPD 153

9.2.2.2 Sensitivity and selectivity of the FPD 155

155

9.3.1 Composition of the gases and their flow-rates 155

9.3.2 Detector temperature 156

9.3.3 Construction 157

9.4 Use of the flame photometric detector 159

9.5 Literature 162

9.3 Effect of experimental conditions on the magnitude of the FPD signal 10 The Coulometric Detector ( C D ) 165

10.1 Detection mechanism 165

10.2 CD signal 169

10.2.1 CD background current 170

10.2.2 CD response 170

10.2.2.1 Linearity and linear dynamic range of the CD 171

10.3 Effect of experimental conditions on the magnitude of the CD signal 172

10.3.1 Gas flow-rate 172

10.2.2.2 Sensitivity and selectivity of the CD 171

8

10.3.2 Construction of the detector 172

10.3.2.1 Bias 175

10.3.3 Temperature 175

10.4 Applications of the CD 175

10.5 Literature 179

11 The Electrolytic Conductance Detector ( E l C D ) 181

11.1 Detection mechanism 181

11.2 ElCD signal 183

11.2.1 ElCD background current 185

11.2.2 ElCD response 185

11.3 Construction of the ElCD 186

11.4 Applications of the ElCD 187

11.5 Literature 188

Index 189

- minimum instantaneous concentration

CmaX - maximum instantaneous concentration

c' - instantaneous concentration in the effective volume of the detector

C, - heat capacity at constant pressure

Cy - heat capacity at constant volume

ic, - ionization current of alkali metal

i ~- ionization current of de-excited atomic states ~ ~ ~ ~

ic,,, - ionization current due to electron capture

- probability ratio for two processes

- analytical property of a

p i n

- excitation energy of the eluted substance

M" - metastable atomic state

M* - excited atomic state

n - number of electrons

n

n - number of measurements

N - number of moles

N' - number of moles in the effective volume of the detector

ORG - polyatomic molecule of organic compound

CI(0RG) - chlorine-containing ployatomic organic molecule

Q E - cross-section for electron capture ionization

Qc,s - cross-section of direct ionization of carrier gas, eluted substance

r - column radius

r - radius of the sensor

R - universal gas constant

R - amplifier resistance

- ionization current of impurities in the carrier gas

- ionization current of the carrier gas due to direct ionization

- intensity of emitted light

- intensity of fluorescence emitted light

- initial intensity of emitted light

- rate constant of dissociation, ionization

- statistical constant dependent on the number of measurements, n

- dissociation constant of acid, base

- length of heated filament

- number of theoretical plates

- partial pressures of the carrier gas, the eluted substance

- part per billion (American billion = lo9)

- heated filament resistance

- detector response to property a

- estimated standard deviation

- eluted substance

- signal of analytical probcrty a, of eluted substance S

- time

- the Student distribution

- time of the beginning and the end of elution

- time corresponding to the inflection points of the elution curve

- elution time

- elution curve width

- time of passage of gas through the effective volume of the detector

- temperature

- temperature of the detector walls

- temperature of the heated filament

- linear flow-rate of gas

- volume flow-rate of gas

- flow-rate expressed in moles

- mean molecular velocity

z,, z a - number of non-elastic collisions

- effective volume of detector

- the third particle in a three-body collision

PREFACE

Much attention has been paid to the theoretical aspects of gas chromatographic separation processes Considerably less theoretical work has been devoted to elu- cidating the function of the detection part of the instrument, the “black box” that yields the actual results for interpretation This leads to a situation in which, for example, experimenters are prepared t o calculate Kovfits retention index values to hundredths of a unit without considering the time constant of the measuring device and recorder employed

An ever increasing demand for high-precision measurements has evolved as

a result of rapidly developing research techniques Thermostats maintaining the temperature constant to within a few hundredths of a degree have been constructed and gas flow-rates are controlled with a precision of kO.1 cm3/min Simultaneously, detectors are regarded as ideal measuring devices; this assumption is obviously false and has been the cause of many erroneous conclusions

The principle aim of this book is to draw attention to a number of experimental conditions that exert a considerable effect on the magnitude and character of the detector response At present, there are only two bboks on gas chromatographic detectors The treatment in German by Jentzsch and Otte appeared in 1970; that by

D J David in English was published just as the manuscript of this book was sub- mitted to the publishers in the beginning of 1974 These two books survey experimen- tal work carried out up t o 1969 and 1970 respectively

This book brings the subject up to date (papers published up t o May 1974 are covered) but, in contrast to the above two books, is not intended t o give an exhaustive survey of the literature on gas chromatographic detectors Attention is centred rather

on the clarification of detection mechanisms and explanation of the dependences of the detector response on experimental conditions

General conclusions which are intended to serve as criteria in designing new detectors are drawn whenever possible

In the Introduction, response formation is considered in general terms and rela- tionships that characterize all detectors are derived; although some of these rela- tionships differ from those commonly used, it is felt that they represent a closer approximation to the actual situation that obtains because of the consistent general approach followed in their derivation The formation of such a general theory is based on the large amount of published data and the author wishes to acknowledge the important role played by these data The literature cited is by no means exhaustive;

only those references which have a direct bearing have been included There is no doubt that much of what has been written here will have to be considerably modified as

a result of further research; it is the author’s hope that this book will prove a stimu- lus for such work

The author is grateful to a1 those who assisted him during the preparation of the book, to all the authors and publishers who gave permission to reproduce original figures and tables, to Dr J Novik for critical discussions and to Dr M Stulikovi and Dr K Stulik for translating the manuscript into English and for comments on the text

This Page Intentionally Left Blank

Progress in gas chromatography very frequently depends on the development of measuring techniques that are capable of monitoring substances after their separation While the separation process has often been treated in detail, less attention has been devoted to the accuracy of results and the stability of various measuring devices, The latter are termed ‘‘detectors” and form an independent part of the gas chromato- graph Their use is based on the various physical and chemical properties of the eluted substances or products formed by their reaction inside the detector

The method first employed for the monitoring of substances separated gas chro- matographically was titration [35] Later, gas volumes were measured [36] or the

eluted substances were weighed after deposition on active charcoal placed on the pan

of a balance [5] None of these methods required calibration or electronic circuitry

and thus contributed to the development of gas chromatography in the 1950s How- ever, simple detection methods are still employed, such as visual observations [74]

The progress of gas chromatography has led to the application of numerous phy- sical detection principles Changes in the electrical properties of materials are utilized

in the thermal conductivity detector, in which changes in the resistance of a heated wire, thermistor or transistor are measured The piezoelectric detector [37, 39, 431

with a quartz crystal, the pyroelectric detector [68] containing lithium tantalate and semiconductor detectors employing the p-n junction of a silicon diode [18,28, 51, 711

or tunnel junction supra conductivity [ 131 have also been used in gas chromatography [17] A detector utilizing variations in the electric conductivity of a TiO, layer [27]

is based on an analogous principle In the literature can also be found descriptions of catalytic detectors [2] and detectors that measure differences in the dielectric constants

of substances [ 5 5 , SO]; the sorptiothermal detector [16] and thermal flux meters [64]

respond to heat changes in the system being studied, in a similar manner to the ther- mal conductivity detector

Detectors based on the measurement of variations in the gas density utilize the inverse proportionality between the flow-rates of gases and their molecular weights

[25, 531 These detectors are among the oldest gas chromatographic detectors;

they have been modified in various ways, e.g., the measurement of changes in the gas density has been performed on a diaphragm [59] or combined with other detectors [22]

Changes in electronic energy states and in rotafional and vibrational energies of niolecules have also been utilized for the construction of detectors [12] Atomic

emission spectra [76], atomic absorption [3, 321, chemiluminescence [75,77] and

detector [52]: a cross-section detector [14, 661 and a surface ionization detector [ll, 231 have been described The function of these detectors is based on collisions

of particles of the eluted substances with high-energy particles The latter are obtained from radioisotopes or from an electric discharge However, such a discharge causes not only ionization but also excitation of molecules and atoms, resulting in emission

of light Thus detectors employing an electric discharge as a source of energy are often called emission detectors [63] In addition t o measuring the intensity of the emitted light, it is possible to measure the voltage at which the discharge begins [61],

the discharge current [30, 501 or the ionization current [47-491 The ionization is

affected by the space charge [56] and by the intensity of the electric field in which the ions move [69] All of the above quantities have been employed in the design of GC detectors

For the elucidation of the ionization mechanism, plasma chromatography [40] has great importance for monitoring the fragments formed during ionization processes under conditions identical with those in the detector Mass spectrometry combined with gas chromatography has made possible the identification of components in complex mixtures

Emitted radiation is also measured in gas chromatography These detectors are termed radiogas detectors [79] and they measure the number of radioactive particles

or pulses [38] with proportional counters [17, 731 or scintillators [34] Detectors utilizing neutron activation analysis [9] can also be included in this group

In addition to the above-mentioned detectors, gas chromatography has employed the monitoring of differences in the magnetic properties of substances (paramagnetism

or diamagnetism) in the magnetic detector [21], changes in the velocity of sound in

the ultrasonic detector [24, 60, 821 and changes in the temperature of a flame [70],

in surface potential [26] and in a number of other physical properties

The chemical properties of the eluted substances are utilized less frequently [67, 72,

831 Coulometry and conductimetry are used, as are polarography [65] and poten- tiometry with ion-selective electrodes [44 - 461 I n analyses of the atmosphere, microbiological detectors [15] have been used In Martin’s opinion [54], the reactions that will find the greatest use in the future are those in which the eluted substances are converted into CO, or H,O The calibration will then be simple and, moreover, chemical amplification will be possible

It follows from this brief introduction that a very wide variety of physical principles are utilized for measuring purposes in gas chromatography The following chapters

are devoted to discussion of selected detectors which are the most frequently used and/or bear promise for future development The main attention is centred on clarification of the detection mechanism and of the dependence of the signal on the experimental conditions Familiarity with these principles is essential for' correct interpretation of the results obtained and for optimization of the design of the measuring device

A substance S passes through the column and, as a result of the establishment of multiple equilibria, its concentration in the gaseous phase varies If N , is the number

of moles of substance S, then its instantaneous concentration at the column outlet can

be expressed by the equation

where : 1' is the retention volume, V o is the volume of gas passed during the elution of

concentration cs and n is the number of theoretical plates

The volume of the eluted gas can be replaced by an expression containing the gas volume flow-rate, u , the elution time, t , and the retention time, r,:

The number of theoretical plates, n, can be expressed as the ratio of the column length, L, to the height equivalent to a theoretical plate, H:

On substitution of expressions (1.2) - (1.4) into equation (l.l), the relationship for

the instantaneous concentration of the eluted substance assumes the form

18

The integral time function of the concentration distribution of the eluted substance fulfils the condition of Gaussian random distribution

/')s.di = Y

and has a maximum, c;Ipx, at time t = t , and two inflection points at times t i n f

located symmetrically on both sides of t,:

L tinf = t + -

R - H

The time interval between the inflection points corresponds to the width of the elution curve, 2 x A t , and is given by

A t = J(L/H)

It follows from the Gaussian distribution law that the value of function Y equals

unity within the integrated time interval, (- co, + 0 0 ) When the integration limits are given by multiples of A t , the value of integral Y is always smaller than the

f

FIG 1.1 An ideal elution curve: t~ - elution time, A? - the elution curve width,

f 4 A? - elution time limits

value corresponding to the oveiall amount of eluted substance (Fig 1.1) When the multiple, 4 ( + A t ) , is used, the error is as little as + 3 x which is satisfactory when considering the experimental conditions

The actual separation of the eluted substance in the column is affected by the siinultaneous processes of dissolution and adsorption, by shifts in the equilibria in the presence of inert substances, by the amount of eluate [33] and especially by the amount of eluted substance, N , [31] The above relationships describing the theoretical separation must then be modified by introducing correcting terms [29, 841 in order

to satisfy better the experimental conditions With a change in the concentration of the eluted substance, a change in the concentration of the carrier gas occurs These

changes can be expressed in terms of changes in the partial pressures, as the overall pressure is constant in the open column separating system

If no other substance is present in the colun~n, then it contains only the carrier gas, C, the instantaneous concentration of which, c,, is constant along the whole column; its partial volume, V,, equals the overall column volume and its partial

pressure, p , at the column outlet On introduction of a substance S into the column, the partial pressure of the carrier gas changes: the overall pressure is unchanged, so that

where N,, N,, etc., are the numbers of moles of substances C, S, etc., respectively

For p s , the relationship

is obtained and the partial volume of the eluted substance is given by

N , R T

P

V, = ~

TABLE 1 1

THE AVERAGE VALUES OF THE CARRIER GAS A N D THE ELUTED

20

On introduction of substance S into the column, the volume of the gaseous phase changes The overall volume of the gaseous mixture equals the sum of the partial volumes of the individual components As the column volume cannot be increased in the gas chromatographic system, the volume changes occur as changes in the volume flow-rate of the gaseous mixture With a constant column cross-section, nr2, the

changes in the volume flow-rate appear as changes in the linear flow-rate, u The dependence of the height equivalent to a theoretical plate on the linear flow-rate is generally hyperbolic; it is evident that the maximum concentration of the eluted substance is also dependent on the flow-rate (equation (1.6))

of the flame ionization detector (FID) to 1-5 P I of carbon disulphide (Fig 1.2)

As CS, is only slightly ionized in the FID (see Chapter 5), the ionization current

decreases with an increase in the partial pressure The minimum ionization current corresponds to the maximum concentration of CS2; the greater the amount of sample injected, the more pronounced is the minimum in the ionization current

-t

FIG 1.3 The dependence of the instantaneous concentration of the eluted substance

on the overall amount, N,, the linear flow-rate, ri, and the elution time, t

Therefore, the instantaneous concentration of the eluted substance is a function of the overall amount of substance, N , , the linear flow-rate of the gaseous mixture and

the elution time, as depicted schematically in Fig 1.3 From the general equations,

(1.6) and (1.8), it follows that the maximum concentration of the eluted substance at the column outlet (the maximum on the elution curve):

- increases with the square root of the number of theoretical plates at a constant

- decreases with increasing flow-rate at constant elution time and number of theo-

- decreases with increasing elution time at a constant number of theoretical plates

flow-rate and elution time;

retical plates;

and flow-rate

The elution curve width:

- increases with increasing retention time at a constant number of theoretical

- decreases with an increasing number of theoretical plates at a constant retention plates;

time

The separated eluted substance is monitored by a detector at the column outlet

22

The basic requirement placed on the detector is that the recording should give a true picture of the concentration distribution in the column Therefore, the detector should not affect

the number of theoretical plates, i.e should not participate i n ' the separation process;

the retention time of the substance, i.e., should be placed at the last theoretical plate of the column:

the gas flow-rate, i.e., its internal diameter and hydrodynamic resistance should

be identical with those of the column

these theoretical requirements cannot be met completely in practice, the measured values deviate from the actual separation results

The detector measures variations in the magnitude of an analytical property, a', of the entering substance The detector records a measurable change after a molecule (radical, atom, ion) of the eluted substance collides with the detector sensor There- fore, it is evident that reactions which occur outside the range of the sensor do not contribute to the measured change The space in which the measurement takes place

is denoted as the detector effective space The measured change dependes only on those molecules present in this space which possess the required analytical property The measured change, denoted as the signal, is proportional to the magnitude

of the given analytical property (thermal conductivity, ionization cross-section, etc.) and to the number of molecules capable of reacting with the sensor The signal is thus proportional to the instantaneous concentration of the eluted substance and can be described by the equation

At a given time, other substances that possess the same analytical property as the eluted substance may be present in the effective volume of the detector The overall measured signal is then given by the sum of the signals for the eluted substance, the carrier gas and impurities:

Before the eluted substance enters the detector, only the molecules of the carrier gas and of impurities present in the carrier gas are present in the effective volume of the detector Stationary conditions are established in the detector and the measured signal is given by the equation

a: < a: > a:

If these values are comparable, the measured signal is very small [42] An ideal measuring arrangement would involve a carrier gas that does not possess the measured property; nitrogen is not ionized in the FID system

During elution of the studied substance, the partial pressures of the carrier gas and

of the impurities present decrease Therefore, the detector background current decreases during the elution, attains a minimum at the maximum concentration of the eluted substance, again increases on the further elution and finally reaches its original value

1.2.1 Detector response

The integral of the signal, s", over the interval ( t l , t z > , where f I and t , correspond

to the beginning and the end of the elution, respectively, is called the detector re-

sponse R":

(1.14)

The effective volume of the detector in which the measurement takes place can be expressed in terms of the gas flow-rate and the time, fder, required for the passage of the gaseous mixture through a volume I/,,,:

(1.15) The time tder can be calculated from the values following from the time distribution

of the concentration of the eluted substance; the substance concentration distribution

is characterized by values t , and A t (see Fig 1.1) If the limits, + 4 x A t , are applied

to function Y, the following relationship can be written for tder:

The newly established distribution of function St (see Fig 1.5) is given by the fol- lowing equation [62]:

When expressions for Pax are compared, it is observed that the maximum on the concentration distribution shifts by fdec/2 on passage of the gas through the detector

FIG 1.6 The shift in the elution time as

a function of the effective volume of the detector and the carrier gas flow-rate TABLE 1.2

THE TIME, tR[sec], OF PASSAGE OF GAS THROUGH THE EFFECTIVE VOLUME O F THE DETECTOR AT VARIOUS FLOW-RATES

The shift in the elution time, A t , , is inversely proportional to the gas flow-rate under

constant experimental conditions, i e., at a constant effective volume of the detector:

t R = - + A 'dec

26

where A is an experimental constant (see Fig 1.6) The shift in the elution time in-

creases with increasing detector volume; some values are given in Table 1.2

Because of the finite values of the effective volume of the detector and the gas flow-rate and because of the necessity of measuring the concentration using a sensor that yields an electric signal, the measured signal is distorted This distortion is called

in the tder/dt ratio [19]

the time constant of the device, T When it is assumed that each signal value corre- sponding t o instantaneous concentration, cs, is reached by a step, then the given signal is, in practice, reached only after a certain time (Fig 1.7) which is approximately

TABLE 1.3

DISTORTION OF THE DETECTOR SIGNAL AS A FUNCTION OF VARYING

tder/dt RATIOS FOR LAMINAR AND TURBULENT FLOW CONDITIONS

equal to three times the time constant Owing to this distortion the maximum of the output function, S""", is shifted with respect to the maximum of the input function,

c;lnx When the time constant increases, not only is the output function maximum shifted to longer times along the descending branch of the ideal Gaussian curve, but

it is generally distorted

The time integral of the output function, S", within limits ( t , t 2 ) has been defined as the response Therefore, distortion of the signal causes distortion of the detector response, which is thus not identical with the concentration distribution at the column outlet Figs 1.8 and 1.9 show changes in the detector response caused by

changes in the t,,,/dt ratio Mass transport also plays a role during the passage of the

eluted substance through the detector If there is a concentration gradient of the eluted substance in the direction of the flow, i.e if the mixture is transported by laminar flow alone and is not stirred, then the probability that all the eluted substances reach the sensor decreases with increasing detector volume and hence the response decreases (Fig 1.8 and Table 1.3) If there is no concentration gradient in the detector along the direction of flow, i.e., the mixture is stirred 'as a result of turbulent flow,

FIG 1.9 Distortion of the response of

a detector with turbulent flow as a func-

tion of changes in the f&t/dt ratio [19]

FIG 1.10 Changes in the response of

a detector with a large effective volume

as a function of the carrier gas flow-rate

then there is always a certain average concentration of the eluted substance in the effective volume of the detector; when the effective volume of the detector is increased

or the flow-rate decreased, the response is distorted, as is shown in Fig 1.9 Much attention has been paid to the dependence of signal distortion on the flow-rate of the mobile phase When the carrier gas flow-rate increases, tdet decreases

28

more rapidly than the elution peak width, A t , and therefore the distortion decreases

An increase in the flow-rate leads to a decrease in the influence of the effective volume

of the detector and the time constant of the device is determined by the time constants

of the other components of the apparatus, such as the recorder Response distortion for these cases is shown in Fig 1.9

The largest shift in elution times is exhibited by devices that employ diffusion processes and those with large effective volumes of the detector, operating at low gas flow-rates Large cells were employed in older types of thermal conductivity detectors; the average concentration is then measured An increase in the flow-rate results in an initial increase in the peak height, corresponding to the maximum number

of moles, N ; , present in the detector On increasing the flow-rate, tdet is increased and hence the elution curve area is decreased (Fig 1.10) It is thus desirable that the measuring device should have as low a time constant as possible and thus also the smallest possible effective volume of the detector

In the measuring device the signal S" is converted into an electrical impulse This impulse is usually a voltage formed on the amplifier resistor, R, through which the

current corresponding t o changes in the effective volume of the detector passes The current is sometimes measured directly using a current follower Therefore, the output signal can always be converted into a current and has the dimension of [A] As the response is defined as the time integral of the signal, it has the dimension of electric charge, [ C ]

The separation of the eluted substance at the column output is described by its concentration distribution Therefore, the gas chromatographic experiment requires evaluation of the relationship between the measured response (signal) and the con- centration of the eluted substance

For quantitative measurements, it is necessary to know the dependence of the response

on the concentration of the eluted substance, i.e., the calibration curve A number of

methods have been used for testing detectors [7, 811 The preparation of standard

mixtures using accurate control of the gas pressure and flow-rate [l] places very high demands on the control accuracy and constancy Calibration of the detector by means of absolute measurement [85] is also not suitable

Single injection of the sample into the column is the method most frequently employed for the calibration of measuring devices; however, it should be pointed out that this method is the least suitable compared with those described below The evaluation of a single injection is subject t o response distortion due to the time constant, T, to absorption processes and t o the fact that relatively large sample vol- umes are iiijected (Table 1.4), occupying a volume larger than that corresponding to one theoretical plate Calibration by a pulse method [ 5 8 ] is subject to similar drawbacks

TABLE 1.4

THE VOLUME OF THE BENZENE GASEOUS PHASE AT VARIOUS

TEMPERATURES AFTER INJECTION OF VARIOUS A M 0 UNTS

the former method [78], a constant amount of sample is injected into the detector for

FIG 1 1 1 Scheme of the diffusion FIG 1.12 Scheme of the logarithmic evaporator; a - permeable PTFE tube, diluter; 1 - vessel with an accurately

studied substance; 3 - container lid lever;

4 - carrier gas inlet; 5 -carrier gas outlet

a pre-selected time The diffusion evaporator utilizes a constant vapour pressure through permeable membranes or capillaries at constant pressure and temperature The diffusion evaporator may consist, for example, of a closed PTFE tube, filled with the studied substance and placed in the stream of carrier gas (Fig 1.11) At

constant temperature and pressure, a constant amount of the studied substance diffuses through the walls of the tube The loss in weight of the tube is determined and the concentration of the studied substance in the effective volume of the detector can then be calculated when the overall volume of the gas is known This method has the advantage that equilibrium conditions are established and the effects of the time constant, absorption, etc., are eliminated

The logarithmic diluter method [4] is based on dilution of the sample by the carrier gas at a constant flow-rate A container with the studied substance is placed

in the closed vessel of the diluter (Fig 1.12); a small amount of substance is introduced from the container into the vessel, the content of the vessel is homogenized and the carrier gas is introduced The concentration in the vessel decreases logarithmically and the actual concentration can be calculated from the equation

where c i is the initial concentration in the diluter vessel with volume Vand c, is the actual concentration at time t and carrier gas flow-rate v The measured signal values can be attributed to all the instantaneous concentration values and errors stemming from inaccurate reading of the response are avoided

The diffusion evaporator and logarithmic diluter methods are the only correct methods for testing detectors, as they exclude systematic and gross errors and, because of the slow variation in the concentration of the studied substance, invohe negligible distortion by the time constant of the device

The classification of detectors frequently encountered in the literature [20] divides detectors into concentration and mass detectors, destructive or non-destructive, integral or differential and universal or selective detectors One of the most frequently used classifications involves separation into two groups, concentration and mass detectors; however, we feel that this classification is not justified In support of this classification, dependences of the signal and response on the flow-rate and pressure

of the mobile phase have been cited However, a change in these parameters leads t o

a change in the concentration distribution of the eluted substance at the column outlet and in the time, fdet, during which the substance remains in the effective volume

of the detector The concentration distribution at the column outlet is described

by equation (1.5) Concentration cs is determined by the number of moles, N , =

= G / M , , where G is the mass of the eluted substance and M , is its molecular weight

Therefore, changes in the signal as a function of the experimental parameters must correspond to the concentration distribution at the end of the column The measured signal change corresponds to the number of molecules, i.e., to their concentration rather than their mass The dependence of this signal on the carrier gas flow-rate is

determined by the construction of the measuring device, chiefly by the V,,, and z values Classification into concentration and mass detectors should therefore be discouraged and the signal or the response should always be related to the concentra- tion of the eluted substance

Classification into destructive and non-destructive detectors is based on the evalu- ation of changes that take place i n the detector due to the detector mechanism Two cases may occur: properties either of the eluted substances or of products of reactions occurring in the detector are monitored This classification thus requires a thorough knowledge of the detection mechanism It should be remembered that irreversible changes may occur even in the thermal conductivity detector, as a result of contact of the substance with the heated filament If the substances entering and leaving the detector should be identical, for example in preparative chromatography, the exper- imental conditions must be carefully selected

Classification into integral and differential detectors I S groundless at 11ie present level of electronics development Differentiating and integrating circuits are common and the form of the recording can be chosen from the point of view of maximum accuracy; differential recording is employed for qualitative analysis and integral recording for quantitative analysis Detectors containing electronic circuits monitor the concentration distribution of the eluted substance at the column outlet, i.e., yield a chromatogram with the shape of a differential curve

The above classifications seem to us to be unsuitable as they d o not supply the experimenter with information for selecting the experimental conditions The classi- fication of detectors according to the analytical property (physical quantity) measured

is therefore more advantageous Detectors that measure changes in the conductivity, either thermal or electric, monitor variations in the sensor resistance Ionization detectors follow variations in the electric conductivity of a medium placed in an electric field Chemical detectors monitor the courses of chemical reactions, etc Although this classification of measuring devices yields more information, it is, of course, not indispensable for detector application

Gas chromatographic measuring devices have been compared according to a num- ber of criteria which have, however, been formulated differently by different workers Some relationships are defined in both linear and logarithmic coordinates, e.g., the linearity and linear dynamic range [6], the definitions of some concepts are not unified, so that units which are not interconvertible result, etc The following argu- ments and relationships are based on the generally valid equation for the chromatc- graphic detector signal (equation (1 lo)) as a universal criterion

1.5.1 Sensitivity

The concept of the sensitivity is derived from equation (1.1 l), which describes the overall detector signal as the sum of the contributions from the carrier gas, impurities and the eluted substance The sensitivity of the measuring device for the eluted sub-

32

stance is then defined as the change in the measured signal, AS", resulting from a change in the concentration, AN;', of the eluted substance:

(1.20) For the condition

in the measuring sensitivity For this reason, measurement with a flame ionization detector is generally more sensitive than with a thermal conductivity detector, as nitrogen as a carrier gas is not ionized in the FID The above phenomena should be borne in mind when the experimental conditions are selected and a carrier gas that contains no impurities which yield a signal in the particular detection mechanism should be used

1.5.2 Detector linearity

The linearity, I , is the proportionality constant in the relationship between the loga-

rithm of the signal and the logarithm of the concentration of the eluted substance

It is determined from equation (l.ll), which, in logarithmic form and under the validity of condition (1.21), can be written in the form

log So = log ( k x a;) + I x log N , , ( I 23)

By plottingequation (1.23), a straight line is obtained with a slope of I and an intercept

on the y-axis equal to log k x a:, i.e., to the logarithm of the sensitivity (Fig 1.13) Evaluation of the detector linearity in linear coordinates is frequently encountered

in the literature The corresponding devices are then called linear detectors, for which the proportionality constant between the signal and the concentration, i.e.,

the exponent I in equation (l.lO), is unity Detectors with slopes other than unity are

called non-linear This approach cannot be considered to be correct, by analogy

with a number of types of detectors that involve exponential dependences of the measured signal on the concentration, such as optical methods, and methods that employ radioactive decay Changes in the signal-to-concentration proportionality can also be caused by handling of the signal in electronic circuits Therefore, it is generally necessary to evaluate the detector linearity in logarithmic coordinates using equation (1.23)

~~ logNs

FIG 1 1 3 The log S, = f ( 1 log Ns) plot; log k a: - logarithm of the sensitivity;

I - detector linearity; (a2 - a,) - detector linear dynamic range

1.5.3 Linear dynamic range

The equation for the instantaneous concentration of the eluted substance and its dependence on the experimental parameters, especially on the carrier gas flow-rate, show that this function exhibits a maximum The detector linearity thus changes in

an analogous manner in the interval (+ 1, - 1) As quantitative evaluation can

be performed only at a constant linearity, it is necessary to find the concentration range in which this parameter is constant

The concentration interval, expressed as the logarithm of the concentration, within which the linearity does not change is called the detector linear dynamic range

If the amount of eluted substance varies in the concentration range - lo-"', the corresponding signals are given by the relationships

(1.26) The linear dynamic range is then expressed by the difference in the exponents of the concentrations (m2 -a1) (see Fig 1.13)

34

The linear dynamic range is important for the quantitative evaluation of chro- matograms However, it does not yield sufficient information, as it is not related to absolute concentrations In order the evaluate the applicability of a detector, it is necessary to know the absolute measurable amount of a substance This is expressed

in terms of the lowest detectable amount

1.5.4 Lowest detectable amount

The measured signal fluctuates owing to the inconstancy of the experimental pa- rameters (temperature, gas flow-rate, line voltage, thermal stability of the electronic circuitry, etc.) These variations can be considered to be random errors in discon- tinuous measurements performed at times t , to t , during the interval t, (see Fig 1.14) and are called the noise

~ ~ - ~ - ~

FIG 1.14 The change in the signal with time under equilibrium conditions; 1 - drift;

2 - noise; t , to t , - discontinuous measuring times in the interval t,

The noise oscillates around the average signal and the variation in individual values can be calculated from the assessed standard deviation :

where k, is a tabulated constant dependent on the number of measurements and the term in parentheses expresses the maximum and minimum deviation of the signal under equilibrium conditions The noise level varies randomly and it is necessary to consider the length of the time interval during which the noise is t o be monitored in order t o obtain a statistically significant value

The measurement time interval can be divided into portions identical with multiples

of & 4 A t (the elution time) which correspond to the number of discontinuous mea- surements, n When it is taken into account that the amount of information virtually does not increase for n > 5 and that the precision of the results increases very little,

two values can be selected as the number of measurements, n = 5 and n = 10; then the noise level, (Soma' - Somi"), is evaluated in time intervals of 40 A t and

80 A t , respectively

During evaluation of the chromatogram, it is important to know the change in the measured signal that can be considered to be statistically significant This value can

be found statistically by comparing the average values, 5” and (s” - AS”), assuming

the same variance Then the values of the Student distribution, t, are given by the

relationship

(1.28)

TABLE 1.5

THE CONFIDENCE LIMITS A N D COEFFICIENT K EXPRESSED I N TERMS

OF THE NOISE MAGNITUDE AT 95:4 A N D 99% PROBABILITY LEVELS

The statistically significant value of AS” can then be calculated from the tabulated

values o f t (see Table 1.5) As the variations in the above average values are identical

and their confidence limits must not overlap, a statistically significant AS“ value is

FIG I , 15 Evaluation of the chromatographic curve; Samax, Samin and 5“ - maximum, minimum and average signal values, respectively

greater than twice the noise level and depends on the interval within which the noise

is monitored (Table 1.5) This consideration is related t o the instantaneous signal This condition can be fulfilled by the maximum signal corresponding to the apex of

36

the elution peak, the other values being less than twice the noise level Then the elution

curve cannot be quantitatively evaluated, as the signal corresponding to time t , f dr

is not known Therefore, the latter value must also be statistically significant, i.e.,

must be valid, where K is a coefficient dependent on the required probability, given

in Table 1.5 Equation (1.29) defines the lowest detectable amount If 95% probability

is satisfactory and the noise is evaluated within a time interval corresponding to ten elution times (80dt), then any signal greater than twice the noise level can be treated

quantitatively (see Fig 1.15) When the probability is increased to 99%, then signals

greater than 2.65 times the noise level can be evaluated; when the noise is monitored over five elution times, then signals 6.72 times greater than the noise level are signifi- cant at the 99% level

1.5.5 Detector selectivity

The ability of the device t o react to only a limited number of substances is denoted as the selectivity, while a device that reacts with only a single substance is called specific The selectivity for substances J and K is expressed by comparing their signals based

on an identical analytical property, a The signals of the two substances are given by the relationships

The probability of various reactions taking place in the detector changes with changes in the experimental conditions In this way, the value of the measured prop- erty can be significantly increased or decreased and the device can become more or less selective A practical requirement for a selective detector is that I' 2 10

1.6 LITERATURE

1 Angely L., Levart E., Guiochon G., Peslerbe G.: Anal Chem 41, 1446 (1969)

2 Arutyunov Yu I., Zibrova L P., Karpov E F., Karbanov E M., Kravchenko V S., Korly-

3 Ballinger P R., Whittemore I M.: Gas Chrornatogr Abstr 1971, 1040

4 Beke R.: Ingenieurs 1970, 5

5 Bevan S C., Thorburn S.: J Chromatogr 11, 301 (1963)

6 Bowman M C., Beroza M.: J Chromatogr Sci 7, 484 (1969)

7 Brazhnikov V V., Sakodynskii K I.: Gazov Khromatogr 1969, 5

8 Burchfield H P., Wheeler R J., Bernos 3 B.: Anal Chem 43, 1976 (1971)

9 Cram S P., Brownlee J L., Jr.: J Gas Chromatogr 6, 313 (1968)

akov G A.: Zavod Lab 38, 660 (1972)

10 Crider W L., Slater R W., Jr.: Anal Chem 41, 531 (1969)

11 Cuderman J F.: Rev Sci Instrum 42, 583 (1971)

12 Dagnall R M., Johnson D J., West T S.: Spectrosc Lett 6, 87 (1973)

13 Dayem A H.: J Phys (Paris), Colloq 1972, 15

14 Del Campo Maldonaldo J L.: Nucl Sci Abstr 22, 28 207 (1968)

15 Druett H A., Packman L P.: Nature 218, 699 (1968)

16 Dudenbostel B F., Priestley W.: Ind Eng Chem 49, 99 A (1957)

17 Dupuis M C., Charrier G., Alba C., Massimino D.: Anal Abstr 20, 852 (1971)

18 Eden C., Margoninsky Y : J Gas Chromatogr 6, 349 (1968)

19 Esser R J E.: 2 Anal Chem 236, 59 (1968)

20 Ettre L S.: Theory Appl Gas Chromatogr Ind Med., Hahnemann Symp 1st 1966 (Pub 1968),

21 Farzane N G., Ilyasov L V.: Zh Fiz Khim 42, 3119 (1968)

22 Farzane N G., Ilyasov L V.: Zh Fiz Khim 45, 921 (1971)

23 Gillen K T., Eiernstein R B.: Chem Phys Lett 5, 275 (1970)

24 Grice H W., David D J.: J Chromatogr Sci 7, 239 (1969)

25 Griffiths J., James D., Phillips C S G.: Analyst 77, 901 (1952)

26 Griffiths J H., Phillips C S G.: J Chem Soc 1954, 3446

27 Guglya V G., Dergunov V V.: Zh Anal Khim 17, 2239 (1972)

28 Guglya V G.: Zaood Lab 39, 403 (1973)

29 Guiochon G., Jacob L., Valentin P.: J Chim Phys Physicochitn Biol 66, 1097 (1969)

30 Harley J., Pretorius V.: Nature 181, 177 (1958)

31 Harris W E.: J Chromatogr Sci 11, 184 (1973)

32 Hey H.: 2 Anal Chem 256, 361 (1971)

33 Huber J F K., Gerritse R G.: J Chrotnafogr 80, 25 (1973)

34 Itaya M.: Chem Abstr 68, 101 441d (1968)

35 James A T., Martin A J P.: Biochern J 50, 679 (1952)

36 Janik J.: CON Czech Chem Commun 19, 684, 700, 917 (1954)

37 Janghorbani M., Freund H.: Anal Chem 45, 325 (1973)

38 Jelen K., Lasa J., Ostrowski K.: Chem Anal (Warsaw) 53, 1239 (1968)

39 Karasek F W., Gibbins K R.: J Chromatogr Sci 9, 535 (1971)

40 Karasek F W., Tatone 0 S., Kane D M.: Anal Chem 45, 1210 (1973)

41 Karavaeva V G., Revel’skii I A,, Zhukhovitskii A A,: Zaood Lab 39, 275 (1973)

42 Karp S., Lowell S.: Anal Chem 43, 1910 (1971)

43 King W H.: Anal Chem 36, 1735 (1964)

44 Kneebone B M., Freiser H.: Anal Chem 45, 449 (1973)

45 Kojima T., Ichise M., Seo Y K.: Bunseki Kagaku 20, 20 (1971)

46 Kojima T., Ichise M., Seo Y.: Bunseki Kugakrr 22, 208 (1973)

23

38

47 Krico-Elektronic K G.: Ger Offen 1,958,751 (May 27, 1971 )

48 Lacase P.: Chromarographia 6, 32 (1973)

49 Lakshtanov V Z., Markevich A V., Dobychin S L.: GUZOC Khromatogr 1966, 46

50 Lakshtanov V Z., Markevich A V., Dobychin S L.: Zh Prikl Khim 40, 2492 (1967)

51 Lomashevich S A., Strokan N B., Fisnek N I.: Fiz Tekhn Poluprou 6, 2247 (1972)

52 Lovelock J E.: Nature 187, 49 (1960)

53 Martin A J P., James A T.: Biochern J 63, 138 (1956)

54 Martin A J P.: Pure Appl Chem 34, 83 (1973)

55 McCarthy W K., Lazarus M L.: Chem Instrum 1, 299 (1969)

56 MC Donald J R.: J Chem Phys 54, 2026 (1971)

57 McLean W R., Stanton D L., Penketh G E.: Analyst 98, 432 (1973)

58 Meyer A S., Knapp J Z.: ,Anal Biochern 33, 429 (1970)

59 Mulyarskii Ya V.: Zauod Lab 36, 1451 (1970)

60 Noble F W., Abel K., Cook P W.: Anal Chem 36, 1421 (1964)

61 Opregaarg M.: Norw Pat 120,095 (Apr 26, 1969)

62 Oster H., Ecker E.: Chromarographia 3, 220 (1971)

63 Overfield C V.: Chem Abstr 74, 106 792c (1971)

64 Patin C., Patin H.: C R Acad Sci., Sec C 276, 311 (1973)

65 Polesuk J., Howery D G.: J Chromatogr Sci 11, 226 (1973)

66 Pompeo D V., Otvos J W.: US Pat 2,641,710 (1973)

67 Poziomek E J., Crabtree E V.: J Assoc Off Anal Chem 56, 56 (1973)

68 Roundy C B., Byer R L.: J Appl Phys 44, 929 (1973)

69 Scolnick M E.: Chem Abstr 72, 28 251t (1970)

70 Scott R P W.: Nature 176, 793 (1955)

71 Seiyama T., Kagawa S.: Anal Chem 34, 1502 (1962)

72 Shcherban A N., Furman N I., Belogolovin N S., Primak A V., Skrynnik P M.: Inner

73 Simpson T H.: J Chromatogr 38 24 (1968)

74 Siuda J F., Debernardis J F., Cavestri R C.: J Chromatogr 75, 298 (1973)

75 Slawiiiska D., Kruh I.: Chem Anal (Warsaw) 18, 923 (1973)

76 Soerensen 0.: Haus Tech., Essen, Vortragsveroeff 283, 34 (1973)

77 Stedman D H., Daby E E.: J Air Contr Poll Assoc 22, 260 (1972)

78 Stevens R K., O’Keefe A E., Ortman G C.: Erioiron Sci Technol 3, 652 (1969)

79 Strong C., Dils R., Galliard T.: Column 1971, 2

80 Turner D W.: Nature 181, 1265 (1958)

81 Vermont J., Guillemin C L.: Anal Cheni 45, 775 (1973)

82 Volkov E F., Rabinovich S I., Breschenko V Y.: Ref Zh, Khim 1969, Abstr No 5 G 14

83 Vtorov B C., Kalmanovskii V I., Chulpava B V., Sheshenin V A., Yashin Y I.: Ref Zh

84 Wold S., Andersson K.: J Chromatogr 80, 43 (1973)

85 Zalkin V S : Zauod Lab 36, 129 (1970)

Tekh 1971, 7 8

Khim 1972, Abstr No 11 N 404

2.1 DETECTION MECHANISM

The thermal conductivity detector is among the most commonly used measuring devices in gas chromatography for monitoring substances separated in the column This detector measures changes in the thermal conductivity of the carrier gas, caused

by the presence of the eluted substances

The thermal conductivity, I f , of component S is determined by the gas density, e,

the mean molecular velocity, ij, the mean free path, 2, and the specific heat at constant

volume, C v :

It follows from equation (2.1) that the thermal conductivity is a function of the size

of the molecules, their mass and the temperature, as

where D is the diffusion coefficient The dependence of the thermal conductivity on the molecular masses of substances S and C is given by the relationship

Thermal conductivity values for some compounds are given in Table 2.1

of the components and their molar fractions, x [46]:

The thermal conductivity of binary mixtures is given by the thermal conductivities

where A and B are constants and .yS = 1 - sc

By solving equation (2.4) for a change in the conductivity of a binary mixture [22],

an expression that quantitatively describes the non-linear variation of the conductivity