Liquid
`Chromatography
`for the Analyst
`
`Raymond E W. Scott
`Ge rgetown Umve
`y
`
`Marcel Dekker, Inc.
`
`New York-Basel-Hong Kong
`
`FRESENIUS KABI 1024-0001
`
`

`
`Library of Congress Cata1ogirsg—in~Publication Data
`
`Scott, Raymond P. W. (Raymond Peter William)
`Liquid chromatography for the analyst / Raymond P. W. Scott.
`p.
`cm. —— (Chromatographic science series; V. 67)
`Includes bibliograpiical references and index.
`ISBN (L824?-9184-3 (alk. paper)
`1. Liquid chromatograpliy.
`1. Series: Chromatograpliie science; v. 67.
`QD79.C454S37
`199—-
`/,
`5
`
`543 ’ .(}894~—dc20
`‘-‘iii 1?,
`
`93-43899
`CIP
`
`ISBN: U—8247~9184—3
`
`The publisher offers discounts on this book when ordered in bulk quantities.
`For more information, write to Speciai Sales/Professional Marketing at the
`address below.
`
`This book is printed on acid—free paper.
`
`Copyright © 1994 by MARCEL I)EKKER, WC. All R'ghts Reserved.
`
`Neither this book nor any part may be reproduced or transmitted in any form
`or by any means, electronic or mechanical,
`including photocopying, micro»
`filming, and recording, or by any information storage and retrieval system,
`without permission in writing from the publisher.
`
`MARCEL DEKKER, INC.
`2?0 Madison Avenue, New York, New York 10016
`
`Current printing (last digit):
`10 9 8 7 6 5 4 3 2 1
`
`PRINTED IN THE UNITED STATES OF AMERICA
`
`.....
`FRESENIUS KABI 1024-OOO2
`
`
`
`

`
`Contents
`
`Preface
`Acknowledgments
`
`1. An Introduction to Chromatography
`
`A Short History of LC
`The Separation Process
`The Different Forms of Chromatography
`Chromatography Nomenclature
`References
`
`2. Resolution, Retention and Selectivity
`
`The Plate Theory
`The Retention Volume of a Solute
`Factors that Control the Distribution Coefficient of a Solute
`Molecular Interactions
`The Thermodynamic Explanation of Retention
`Factors that Control the Availability of the
`Stationary Phase
`
`iii
`v
`
`1
`
`2
`4
`7
`9
`13
`
`15
`
`17
`21
`23
`23
`29
`
`33
`
`vii
`
`FRESENIUS KABI 1024-0003
`
`

`
`viii
`
`Chromatographic Methods of identification
`The Capacity Ratio of a Solute
`The Separation Ratio
`Column Efficiency
`References
`
`Liquid Chromatography Phase Systems
`
`The Production of Silica Gel
`Solute/Solvent Interactions on the Surface of Silica Gel
`
`Mono—Layer Adsorption of Solvents on the Surface
`of Silica Gel
`Solute Interactions with the Silica Gel Surface
`
`(Mobile Phase n—Heptane/Chloroform)
`Bi~Layer Adsorption of Soivents on the Surface of
`Silica Gel
`Solute Interactions. with the Silica Gel Surface
`
`(Mobile Phase n~HeptanefEthyl Acetate)
`Silica Gel as an Exclusion Medium
`
`Application Areas for Silica Gel as a Stationary Phase
`Silica Gel as a Stationary Phase in Elution
`Chromatography
`Silica Gel as a Stationary Phase in Exclusion
`Chromatography
`B011 ed Phases
`The Different Classes of Bonded Phases
`"Reverse Phases
`
`The Interaction of Reverse Phases with Solvents
`and Solutes
`
`‘3‘aired Ion Reagents
`The Use of Bonded Phases in Elution Chromatography
`Aqueous Solvent Mixtures
`The Use of Macro~Poroos holymers as Stationary Phases
`Polymer Stationary Phases in Elution Chroma-
`tography
`General Advice for the Unknown Sample
`References
`
`40
`4 l
`42
`44
`48
`
`51
`
`55
`S8
`
`58
`
`60
`
`63
`
`65
`67
`69
`
`69
`
`70
`71
`73
`76
`
`7'?
`79
`81
`82
`84
`
`90
`91
`91
`
`FRESENIUS KABI 1024-OOO4
`
`

`
`The Liquid Chromatography Column
`
`The Rate Theory
`The Summation of Variances
`The Maximum Sample Volume
`The Van Deemter Equation
`The Significance of the HETP Equation
`The Reduced Chromatogram
`Resolution
`Calculating the Efficiency Required to Achieve a
`Specific Resolution
`Peak Asymmetry
`Column Selection
`Column Diameter
`Preparative Columns
`Volume Overload
`References
`
`The Liquid Chromatograph
`
`The Basic Liquid Chromatograph
`The Mobile Phase Supply System
`The Gradient Programmer
`The LC Pump
`The Sample Valve
`The Column and Column Oven
`Detectors
`Data Acquisition and Processing
`References
`
`Liquid Chromatography Detectors
`
`Detector Specifications
`Detector Linearity and Response Index (oz)
`Linear Dynamic Range
`Detector Noise Level
`
`ix
`
`93
`
`94
`94
`95
`97
`105
`106
`108
`
`109
`1 11
`114
`116
`117
`118
`121
`
`123
`
`124
`124
`125
`128
`138
`144
`149
`152
`155
`
`157
`
`158
`158
`161
`162
`
`FRESENIUS KABI 1024-0005
`
`

`
`
`
`Measurement of Detector Noise
`Detector Sensitivity or the Miniinurn Detectable
`Concentration
`
`Pressure Sensitivity
`Flow Sensitivity
`Temperature Sensitivity
`The UV Detector
`
`The Fixed Wavelength Detector
`The Multi—WaVeiength Detector
`The Electrical Conductivity Detector
`The Fluorescence Detector
`The Refractive Index Detector
`
`The Tridet Multifunctional Detector
`References
`
`Sample Preparation
`
`Sample Preparation Techniques
`Filtration Techniques
`Centrifugation Techniques
`Concentration and Extraction Techniques
`An Automatic Sample Extraction and Concen-
`tration Procedure
`
`A Sample Separation Protocol
`LC Sample Preparation for Solid Materials
`Materials of interest Present in High Concentration
`Materials of Interest Present in Low Concentration
`The Preparation of Liquid Samples for LC Anaiysis
`Materials of Interest Present in High Concentration
`Materials of Interest Present in Low Concentration
`The Preparation of Liquid/Solid Samples for LC Analysis
`Materials of Interest Present in High Concentration
`Materials Present in Samples at Low Concentration
`General Comments on Choice of Mobile Phase
`Derivatization Techniques
`Pre—Column Derivatization
`Post~Coiuinn Derivatization
`References
`
`63
`
`64
`l65
`65
`65
`167'
`
`76
`180
`184
`189
`193
`
`I95
`
`195
`
`196
`197
`198
`
`205
`210
`212
`213
`217
`221
`221
`225
`228
`229
`232
`235
`237
`238
`
`245
`247
`
`FRESENIUS KABI 1024-OOOH6
`
`

`
`8. Qualitative and Quantitative Analysis
`
`Qualitative Analysis
`The Effect of Temperature on Retention
`Volume Measurement
`
`The Effect of Solvent Composition on Retention
`Volume Measurement
`
`Quantitative Analysis
`Peak Height Measurements
`Peak Area Measurements
`
`Procedures for Quantitative Analysis
`Quantitative Analysis Using Reference Standards
`The Relative Precision of Peak Height and Peak
`Area Measurements
`Peak De—Convolution
`
`Reporting Analytical Results
`References
`
`9. LC Applications
`
`Separations Based on Exclusion Chromatography
`Exclusion Chromatography Employing Silica Gel
`Exclusion Chromatography Employing Micro-
`Reticulated Cross—Linked Polystyrene Gels
`Chiral Separations
`Interactive LC Systems
`Dispersive Interaction Chromatography
`Polar Interaction Chromatography
`Ionic Interaction Chromatography
`Mixed Interaction Chromatography
`Summary
`
`Index
`
`X1
`
`251
`
`252
`
`260
`
`262
`
`265
`265
`266
`
`267
`267
`
`272
`273
`
`277
`279
`
`281
`
`282
`283
`
`286
`290
`296
`297
`304
`309
`314
`319
`
`32 J
`
`FRESENIUS KABI 1024-0007
`
`

`
`
`
`8 Q
`
`ualitative and Quantitative Analysis
`
`The object of an LC analysis is to establish the probable identity, and
`determine the precise amount, of each ofthe pertinent components
`present in the sample. The pertinent components may include all the
`substances present in the mixture or only those of specific interest.
`The identity of the peak is determined from its position on the
`chromatogram, that is, the time required for it to be eluted, whereas,
`the quantity ofa component present is determined from the peak
`height or peak area. It must be emphasized that a single LC analysis on
`a hitherto unknown sample can not unambiguously confirm the
`presence of a particular compound on the basis of retention data alone.
`Retention data, whether it is corrected retention volume, capacity
`factor or the separation ratio of the solute to that of a standard, can
`only indicate the probability of substance identity. Retention data from
`a second analysis, using a different phase system, increases the
`confidence level but absolute verification requires confirmation by
`another analytical technique. This might include infrared spectrometry,
`mass spectrometry or nuclear magnetic resonance spectroscopy. Such
`evidence would be essential for litigation purposes.
`
`There are in—line LC/spectroscopic systems available, but in most
`cases it is easier to carry out a semi—preparative separation, collect the
`material and carry out the spectroscopic examination off—line.
`However, for routine quality control analyses, where the sample
`
`251
`
`FRESENIUS KABI 1024-0008
`
`

`
`252
`
`clzaracteristics are already well established, retention data can be
`
`quite sufficient for identification purpose.
`
`Qualitative Analysis
`
`In LC both quantitative and qualitative accuracy depends heavily on
`the components ofthe sample being adequately resolved from one
`another. The subject of resolution has already been discussed, but it is
`necessary to consider those areas where uncertainty can still arise.
`Unfortunately, unless the analyst is aware of the pitfalls and how to
`deal with them, false assumptions of resolution can be made very
`easily.
`
`Consider the liquid chromatography peak shown in figure l.
`
`Figure 1
`
`A Single LC Peak
`
`1.25
`
`1
`
`E
`g
`I:
`
`8 0.75
`fl
`
`OD g
`
`GW
`
`0.5 _
`
`0.25
`
`0
`
`'
`
`
`
`3500 Plates
`
`2500 Plates
`
`4.8
`
`5
`
`5.2
`
`5.4
`
`.
`
`Volume Flow of Mobile Phase (ml)
`
`It is seen that the peak shown in figure 1 is asymmetric, which is
`typical of many peaks eluted from LC columns. This particular peak is
`not grossly asymmetric, the front half of the peak having an efficiency
`equivalent to 3500 theoretical plates and the latter half an efficiency Of
`
`FRESENIUS KABI 1024-0009
`
`

`
`253
`
`2500 theoretical plates. This situation frequently occurs in LC and can
`be caused by a number of different effects. The two major causes of
`peak asymmetry have already been touched upon and can arise from a
`difference in peak dispersion occurring in the front half ofthe peak
`relative to the rear half or to the solute distribution coefficient being
`different for the two halves of the peak.
`
`Re-iterating the HETP equation (10) given on page 104,
`
`H=2?tdp+
`
`u
`
`27D
`fi(k)d
`f(k')d2
`um + ——E—3u+ —2——5S——fu
`In
`
`2
`
`(1)
`
`Where the symbols have the meanings previously defined.
`
`Now, it was also shown on page 145 that for a given column, solute,
`mobile phase and flow rate, equation (1) can be reduced to an alternative
`abbreviated form which is given as follows,
`
`(2)
`
`On page 6, it was shown that in the front half of the peak, there will
`be a net transfer of solute from the mobile phase to the stationary
`phase and thus the resistance to mass transfer in the mobile phase will
`dominate. At the rear half of the peak there is a net transfer of solute
`from the stationary phase to the mobile phase and in this case the
`resistance to mass transfer in the stationary phase will dominate. Then
`if the resistance to mass transfer in the stationary phase is greater than
`that for the mobile phase, the rear part of the peak will be broader
`than the front half. In which case,
`
`2
`cldp < czdg
`Dm
`DS
`
`(3)
`
`FRESENIUS KABI 1024-0010
`
`

`
`254
`
`Consequently, the peak will exhibit the asymmetry shown in figure 1.
`It is seen that the relative values of the resistance to mass transfer
`2
`2
`
`terms is controlled by the functions, Fig and
`
`Dm
`
`DS
`
`It would seem that, in practice, the inequality defined in (3) can
`frequently occur but the converse does not appear to be true. Thus,
`peak asymmetry (in part or whole) resulting from inequality in mass
`transfer between the two phases manifests itself in the form shown in
`
`figure 1.
`
`Alternatively, peak asymmetry could arise from thermal effects.
`During the passage of a solute along the column the heats of
`adsorption and desorption that are evolved and adsorbed as the solute
`distributes itself between the phases. At the front of the peak, where
`the solute is being continually adsorbed, the heat of adsorption will be
`evolved and thus the front of the peak will be at a temperature above
`
`its surroundings. Conversely, at the rear of the peak, where there will
`be a net desorption of solute, heat will be adsorbed and the
`temperature or the rear of the peak will fall below its surroundings.
`
`Figure 2
`
`Temperature Profile of a Peak Passing Through a Heat of
`Adsorption Detector
`
`Temperature above that of
`the surroundings
`
`A
`
`Temperature
`
`Temperature below that of
`the surroundings
`
`..?___._?__.>
`Time
`
`FRESENIUS KABI 1024-0011
`
`

`
`255
`
`In fact, this phenomenon has been used as the basis of a very sensitive
`detecting system. An example of the temperature profile of an
`adsorbent as a peak passes over it is shown in figure 2. Unfortunately,
`it was found almost impossible to produce a true simulation of the
`concentration profile of the peak from the temperature profile and
`interest in the detector declined.
`
`Nevertheless, it can be seen from figure 2 that the evolution and
`adsorption of heat will cause the front of the peak to travel through
`the column at an elevated temperature relative to the rear of the peak.
`Furthermore as the heat evolved will be proportional to the mass
`adsorbed, the largest temperature changes will occur at the highest
`concentrations in the peak. As a result, the distribution coefficient of
`the solute with respect to the stationary phase will be slightly smaller
`at the front of the peak and at the high concentrations, i.e. the peak
`maximum. Now, it has already been established that the velocity of a
`solute band through a column is inversely proportional to its
`distribution coefficient. Hence, the high concentrations of the peak-
`front will move slightly faster through the column than the lower
`concentrations of the peak. This will cause the peak maximum to move
`towards the front of the peak and thus produce peak asymmetry. The
`effect of the heat of adsorption on peak shape has been treated
`quantitatively for both LC (1) and GC (2).
`
`In practice, it is probable that both of the effects discussed contribute
`to the overall peak asymmetry. Unfortunately, peak asymmetry varies
`in extent from the very obvious to the barely noticeable and because
`of this, peak asymmetry is often dismissed as the normal shape ofa
`single solute peak. Such an assumption can cause serious errors in both
`qualitative and quantitative analysis.
`
`Under circumstances where two solutes are incompletely resolved and
`one of the pair is present at a much lower concentration than the
`other, the profile of the pair often resembles a normal peak with slight
`asymmetry. Consider the combined elution profile of the two peaks
`shown in figure 3.
`
`FRESENIUS KABI 1024-0012
`
`

`
`256
`
`Figure 3
`
`The Combined Elution Profile of Two Unresolved Peaks
`
`1.25
`
`0.75
`
`3 (II Solute
`Concentration
`
`I
`
`.° I») U‘!
`
`0
`
`»
`
`i
`
`4.7 4.8 4.9
`
`5
`
`5.1 5.2 5.3 5.4 5.5
`
`Volume of Mobile Phase (ml)
`
`It is seen that the peak profile shown in figure 1, resulting from an
`asymmetric single peak, is indistinguishable from the asymmetric peak
`shown in figure 3 that results from a 10% impurity eluting relatively
`close to the parent peak. Care must be taken not to interpret peak
`asymmetry as the normal elution profile when, in fact,
`it
`is an
`indication of incomplete resolution. If there is any uncertainty, the
`splvent should be changed and the peak in question carefully
`monitored. It does not matter if, in changing the solvent mixture, the
`resolution of other components is lost, as only the integrity of the peak
`in question is being examined. The composition of the mobile phase
`should be changed at least twice to be sure that only a single solute is
`present. If no second component can be detected from the peak
`profiles, then the original mobile phase can be used for the analysis
`and the integrity of the peak assumed.
`
`FRESENIUS KABI 1024-0013
`
`

`
`257
`
`This an excellent example of the value of the diode array detector. If
`the chromatogram shown in figure 3 was monitored at two different
`wavelengths, then a peak ratio curve would immediately disclose the
`presence of the second peak (see page 175) and it would no longer be
`necessary to resort to changes in mobile phase composition to establish
`the presence of the impurity.
`
`A second component, even at low concentration, will not only give an
`erroneous value for the amount of solute present but, on examining
`figure 3 it is seen that the retention time ofthe major peak is also
`significantly changed. This effect can be used for quantitative
`estimation of the unresolved mixture providing the chromatographic
`sytem can provide the necessary high precision.
`
`Scott and Reese (3) measured the accuracy and precision that could be
`obtained from a well designed liquid chromatograph and their work
`will be referred to again later. In the course of their work they found
`that, providing the retention time of a mixed peak could be measured
`with sufficient accuracy and the retention times of the two pure
`components were known, then the proportion of the two components
`could be determined, even if only a single peak was apparent. The
`theory behind this work is a little complex (4), and it is not
`appropriate to discuss the details here. The basis of the method was to
`calculate the theoretical retention time from the combined elution
`curves for the two solutes over a range of mixtures. The curve
`relating retention time to composition was then used as a calibration
`curve to determine the composition of an unknown mixture from the
`retention time of the combined peak. The calibration curve that was
`calculated and the experimental points they obtained are shown in
`figure 4.
`
`It is seen that the solutes used were nitrobenzene and deutero-
`nitrobenzene. These two solutes are so similar in physical and
`chemical characteristics that neither can be exclusively detected in a
`solvent without interference by the other. It is also seen that the
`retention times of the two solutes are 8.927 minutes and 9.061 minutes
`giving a difference of only 8.04 sec.
`
`FRESENIUS KABI 1024-0014
`
`

`
`_
`
`258
`
`Figure 4
`
`Graph of Retention Time Difference against Sample
`Composition for Mixtures of Nitrobenzene and Deutero-
`Nitrobenzene
`
`Deutero-
`Nitro-
`Benzene
`
`9.061m.
`
`
`
`.\' U1
`
`
`
`RetentionTimeDiff(secs) NV‘‘U10
`
`8.927
`
`0
`
`20
`
`
`60
`40
`0
`80
`100
`% w/w Deutero—Nitrobenzene
`
`It follows that measurements must be made with a precision of about
`0.2 second if quantitative results are to be of any value. It is seen from
`figure 4 that the experimental points lie very close to the line and a
`fairly accurate measurement of the distribution of the two isotopes can
`' be obtained from retention time measurements. This method has very
`limited areas of application and is given here, more to demonstrate the
`effect of unresolved impurities on retention time, than to suggest it as
`an alternative to adequate chromatographic resolution. In some cases,
`however, particularly in the analysis of isotopes, it may be the only
`practical way to obtain a quantitative evaluation of the mixture by a
`liquid chromatographic method.
`
`FRESENIUS KABI 1024-0015
`
`

`
`Finally, it must be emphasized that the uncertainty arising from the
`slight asymmetry shown by the presence of a 10% impurity in figure
`3 does not depend on the magnitude of the impurity. In figure 5 the
`elution curve resulting from a 50% mixture oftwo closely eluting
`components is shown.
`
`259
`
`Figure 5
`
`The Elution Curve ofa Binary Mixture Containing Equal
`Quantities of Each Solute
`
`1.5
`
`Concentration P (II
`Solute
`
`0
`
`4.7 4.8 4.9
`
`5
`
`t
`5.1 5.2 5.3 5.4 5.5
`
`Volume Flow of Mobile Phase (ml)
`
`It is seen that the profile of the combined peaks is perfectly
`symmetrical and displays no hint that there are two solutes present.
`Obviously an absorption ratio curve from a diode array detector
`would quickly disclose the presence of the two components, as would
`an appropriate changes in mobile phase composition. However, there
`would be a further clue for the analyst to follow that would give
`warning of the "duplicity" ofthe peak. The double peak would be
`very broad and be inconsistent with the change in peak width of the
`other solute peaks with retention time. The peak width of a solute
`increases regularly with retention time but, unfortunately, the
`relationship is not smooth. There are good reasons for this, but they
`
`FRESENIUS KABI 1024-0016
`
`

`
`260
`
`are not pertinent to this discussion. However, a peak such as that
`shown in figure 5 would be excessively broad and obviously out of
`order with the peak widths of neighboring solutes. This should alert
`the analyst to the possibility of an unresolved pair of solutes. The
`analyst should always be alert to the width of each peak in relationship
`to its position on the chromatogram, as this is the first indication of
`the presence of a composite peak.
`
`The Effect of Temperature on Retention Volume Measurement
`
`There are two major factors that influence retention volume
`
`measurement and they are temperature and solvent composition. In
`order to measure retention volume with adequate precision it is
`necessary to know the relationship between retention time and
`temperature so that the control limits of the column temperature can
`be specified.
`
`The effect of temperature on retention time was investigated by Scott
`and Reese (3), who measured the retention volume of the solutes o-
`dinitro-benzene, 2—ethoxy naphthalene and p-chlorophenatole over a
`range of temperatures. The chromatographic conditions used are as
`follows,
`
`Chromatographic Conditions
`
`Column
`
`Silerex 1
`
`Column Length
`Column Diameter
`
`25 cm
`4.6 mm
`
`Column Packing
`l\/Iobile Phase
`Flow Rate
`
`Silica Gel (particle size 10pm)
`44% butyl chloride and 56% n—heptane
`1 ml/min.
`
`Detector
`Sample Volume
`
`UV adsorption at 254 nm
`1 ul
`
`Scott and Reese chose to monitor retention volume as opposed to
`retention time, as retention volume is always the primary dependent
`variable in LC. Retention time is not a primary measurement because
`it must also include the reproducibility of the flow—rate delivered by
`
`FRESENIUS KABI 1024-0017
`
`

`
`261
`
`the pump. The retention volume was measured absolutely by means of
`a burette connected to the end of the column. The column was situated
`in a water bath that was controlled to +/- 0.050C. Precautions were taken
`to ensure the mobile phase attained the column temperature before
`entering the column. The results obtained are shown in figure 6A.
`
`Figure 6A
`
`Graphs of Retention Volume against Temperature for Three
`Solutes
`
`0-Dinitrobenzene Mean Retention Volume 17.19 ml
`Slope -0.262 ml/C 1.53%/C
`
`.$__\\\\\
`
`2-Ethoxy Naphthalene Mean Retention Volume 4.93 ml
`Slope -0.070 ml/C 1.44%/C
`
`
`
`p-chlorophenatol Mean Retention Volume 3.077 ml
`Slope -0.044 ml/C 1.42%/C
`
`_____fi“‘____*—“;7_‘-__
`
`t\) G
`
`».—- (J1
`
`anS
`
`
`
`
`
`RetentionVolume(ml)
`
`20
`
`26
`24
`22
`Temperature (C)
`
`28
`
`It is seen that in order to measure retention volumes with a precision
`of 0.1%, the temperature control must be +/- 0.0400 This level of
`temperature control on a thermostat bath is not difficult to achieve but
`it is extremely difficult, if not impossible, to return to a specific
`temperature to within +/- 0.040C after prior change. To achieve a
`preoision of retention volume measurement of 1%, the temperature
`control must be +/- 0.400 This is far more practical as most column
`oven temperature can be set to a given temperature to within +/-
`0.250C. Although the data was obtained for three specific solutes, the
`results can be taken as reasonably representative for all solutes and
`phase systems. In most practical analyses, the precision limits of
`retention volume measurement will be about 1% but this will not
`include the reproducibility of the flow rate given by the pump. As
`
`FRESENIUS KABI 1024-0018
`
`

`
`262
`
`retention times are usually measured in analytical LC, the precision of
`
`measurement may be significantly greater than +/—1%. It follows that
`
`ifthe identity ofthe peak must be confirmed, and retention data is
`
`being used for the purpose, then it is essential that the column is
`
`carefully thermostatted.
`
`The Effect of Solvent Com_p_osition on Retention Volume Measurement
`
`Scott and Reese (3) also measured the change in retention volume with
`
`solvent composition using the same LC apparatus as that used for
`
`investigating the effect of temperature. The column was thermostatted
`
`at 247°C and the results that were obtained are shown in figure 6B.
`
`Figure 6B
`
`Graphs of Retention Volume against Solvent Composition
`for Three Solutes
`
`o—Dinitrobenzene Mean 17.57 ml
`Slope 5.27% per unit change in concentration
`
`9 —&
`E\i
`
`
`
`I5
`
`to
`
`u
`
`E Ec > =
`
`1
`.2
`‘E
`3
`é
`
`0
`
`2—Ethoxy Naphthalene Mean 5.01 ml
`Slope 5.02% per unit change in concentration
`5
`
`—0-—Z0-:-—0:2——o——:-O-
`
`p-chlorophenatole Mean 2.857 ml
`Slope 3.71% per unit change in concentration
`
`43
`
`if
`
`44
`
`45
`
`Solvent Composition % w/v Butyl Chloride
`
`FRESENIUS KABI 1024-0019
`
`

`
`263
`
`It is seen that small changes in solute composition can have a profound
`effect on the retention volume, but it must be borne in mind that the
`magnitude of the effect will vary somewhat, between different solutes
`and different phase systems. The results shown in figure 6B indicate
`that if a retention volume is to be measured with a precision of 0.1%,
`the solvent composition must be maintained constant to within 0.02%
`w/v. Normally, the solvent composition can be kept constant within
`those limits providing a closed solvent system is used to prevent
`evaporation. However, it would be extremely difficult, if not impossible
`to make up another solvent mixture of the same composition within
`those limits of precision. This would be particularly difficult if either
`of the solvents were volatile. If the precision required for retention
`volume measurement was 1%, the solvent composition would have to
`be maintained constant with a precision of +/- 0.2 % w/v which should
`be quite practical. Furthermore, it would be fairly straightforward to
`make up replacement solvents to the same concentration within the
`same precision limits. For a routine analysis, however, it might well
`be advantageous to prepare the mobile phase in large volumes and
`store it in an appropriate manner. Scott and Reese (3) examined the
`repeatability of retention time measurements by two procedures, one
`employing computer processed data and the other by manual
`measurement of the distance in centimeters from the injection point to
`the peak maximum on a potentiometric recorder chart. They carried
`out twelve replicate analyses and the resulting statistical analysis is
`shown in table 1.
`
`Table 1
`
`The Precision of Retention Data Measurement Made
`Manually and by Computer
`
`Parameter
`Capacity factor
`
`Peak 1
`0.22
`
`Mean (cm)
`S.D.(% Mean)
`Mean (min.)
`S.D.(% Mean)
`
`10.17
`0.85
`3.97
`0.31
`
`Peak 2
`0.94
`
`16.0
`0.245
`6.27
`0.20
`
`Peak 3
`1.50
`
`20.59
`0.19
`8.11
`0.17
`
`Peak 4
`5.21
`
`51.14
`0.15
`20.14
`0.33
`
`FRESENIUS KABI 1024-0020
`
`

`
`Figure 7
`
`Tips of Peaks Having Extreme Retention Values
`Reconstructed by the Computer
`
`Peak Crests (between 99.9 and 100% of the peak heights)
`
`Retention Time Difference
`Recorded
`
`,<———4.4 sec. -1-»;1
`
`I l1l i
`
`<—2.1 sec.
`
`Retention Time Difference
`Actual
`
`By taking the mean positions of the peaks, it is seen that they are only
`2.1 sec apart. However, as a result ofa noise spike on the front of the
`
`FRESENIUS KABI 1024-0021
`
`

`
`265
`
`first peak and a similar noise spike on the back of the second peak, the
`computer interpreted each noise spike as the peak maxima, which
`resulted in a retention difference of 4.4 see. It is clear that a
`significant part of the scatter of retention data can be due to detector
`noise, either arising from the detector itself or from changes in
`operating conditions of the chromatograph. This once more emphasizes
`the importance of using a chromatograph that can maintain stable
`operating conditions in conjunction with a detector having as little
`long term noise as possible. The results also indicate that unless the
`maximum sensitivity is essential for the analysis, the sample size
`should be adjusted so that the detector can be operated well below its
`maximum sensitivity to reduce the consequence of the noise.
`
`Quantitative Analysis
`
`Quantitative estimates of the mass of a particular solute present in a
`sample are obtained from either peak height or peak area measurements.
`The values obtained are then compared with the peak height or area of
`a reference solute present in the sample at a known concentration or
`mass. In this chapter quantitative analysis by LC will be discussed but
`the procedures described should not be considered as entirely
`appropriate for other types of chromatographic analysis. Those
`interested in general quantitative chromatographic analysis including
`GC and TLC are referred to the book by Katz (4).
`
`Peak Height Measurements
`
`Most detectors are concentration sensitive devices and thus the peak
`height will be proportional to the maximum concentration in the peak,
`which, in turn, will be proportional to the total area of the peak. The
`total area of the peak is proportional to the total mass of solute
`contained in the peak providing it is not excessively tailing. As the
`peak height is inversely related to the peak width, then, if peak heights
`are to be used for analytical purposes, all parameters that can affect the
`peak width must be held constant. This means that the capacity factor
`of the solute (k') must remain constant and, consequently, the solvent
`
`FRESENIUS KABI 1024-0022
`
`

`
`266
`
`composition is held stable. The temperature must also be held constant
`and an highly repeatable method of sample injection must be used. If
`computer data acquisition and processing are employed, then a direct
`printout of the peak heights is obtained and, with most systems, the
`calculated analysis is also presented. Ifthe peak heights are to be
`measured manually, which even today is carried out in the majority of
`LC analyses, then the base line is produced beneath the peak and the
`height between the extended base line and the peak maximum
`measured. In general, the measurements should be made estimating to
`the nearest 0.1 mm.
`
`Peak Area Measurements
`
`The area of a peak is the integration of the peak height (concentration)
`with respect to time (volume flow of mobile phase) and thus is
`proportional to the total mass of solute eluted. Measurement of peak
`area accommodates peak asymmetry and even peak tailing without
`compromising the simple relationship between peak area and mass.
`Consequently, peak area measurements give more accurate results
`under conditions where the chromatography is not perfect and the
`peak profiles not truly Gaussian or Poisson.
`
`Unfortunately, neither the computer nor the potentiometric recorder
`measures the primary variable, volume of mobile phase, but does
`measure the secondary variable, time. This places stringent demands
`on the LC pump as the necessary accurate and proportional
`relationship between time and volume flow depends on a constant flow
`rate. Thus, peak area measurements should never be made unless a
`,good quality pump is used to control the mobile phase flow rate.
`Furthermore, the pump must be a constant flow pump and not a
`constant pressure pump.
`
`Peak areas can be measured manually in a number of ways, the
`simplest being the product ofthe peak height and the peak width at
`0.6065 of the peak height. This does not give the true peak area but
`providing the peak is Poisson, Gaussian or close to Gaussian it will
`
`FRESENIUS KABI 1024-0023
`
`

`
`267
`
`always give accurately the same proportion of the peak area. Other
`methods involve the use of a planimeter (an instrument that provides a
`numerical value for the area contained within a perimeter traced out
`by a stylus), and the measurement of the product of the peak height
`and the peak width at the base. The first method is extremely tedious
`and the second somewhat inaccurate. The most accurate manual
`method of measuring peak area is to cut the peak out and weigh it. A
`copy of the chromatogram should be taken and the peaks cut out of the
`copy. This procedure is also a little tedious (not as tedious as the use
`of a planimeter) but does provide very accurate values for the peak
`area. It is particularly effective for skewed or malformed peaks where
`other methods of manual peak area measurement (with the exception
`of the planimeter) fail dismally and give very inaccurate results. The
`recommended method is to use the product of the peak height and the
`peak width at 0.6065 of the peak height but this does require adequate
`resolution of the components of the mixture.
`
`Procedures forQuantitative Analysis
`
`There are two basic methods used in quantitative analysis; one uses a
`reference standard with which the peak areas (peak heights) of the
`other solutes in the sample are compared; the other is a normalization
`procedure where the area (height) of any one peak is expressed as a
`percentage ofthe total area (heights) of all the peaks. There are
`certain circumstances where each method is advantageous, and
`providing they are used carefully and appropriately all give
`approximately the same accuracy and precision.
`
`Qugtitative Analysis Using_Reference Standards
`
`Reference standards can be used in two ways; a weighed amount of the
`standard can be added directly to the sample and the area of the peaks
`of interest compared with that of the standard; alternatively, a
`weighed amount of the standard can be made up in a know