ll
`
`HIGH PRESSURE
`
`LIQUID
`
`CHROMATOGRAPHY
`
`Biochemical and Biomedical
`
`Applications
`
`ll
`
`PHYLLIS R. BRO WN
`
`Department of Biochemical Pharmacology
`Division of Biological and Medical Sciences
`Brown University
`Providence, Rlzode Ixlcmd
`
`ACADEMIC PRESS
`
`I9 73
`
`New York and London
`
`FRESENIUS KABI 1031-0001
`
`

`
`COPYRIGHT © 1973, BY ACADEMIC PRESS, INC.
`ALL RIGHTS RESERVED.
`No PART OF THIS PUBLICATION MAY BE REPRODUCED OR
`TRANSMITTED IN ANY FORM OR BY ANY MEANS, ELECTRONIC
`OR MECHANICAL, INCLUDING PHOTOCOPY, RECORDING, OR ANY
`INFORMATION STORAGE AND RETRIEVAL SYSTEM, WITHOUT
`PERMISSION IN wRITING FROM THE PUBLISHER.
`
`ACADEMIC PRESS, INC.
`111 Fifth Avenue, New York, New York 10003
`
`United Kingdom Edition published by
`ACADEMIC PRESS, INC. (LONDON) LTD.
`24/28 Oval Road, London NW1
`
`LIBRARY OF CONGRESS CATALOG CARD NUMBER: 72-77361
`
`PRINTED IN THE UNITED STATES OF AMERICA
`
`—
`FRESENIUS KABI 1031-0002
`
`

`
`CONTENTS
`
`1’ref(IL'e
`Abbreviations
`
`Chapter 1.
`
`INTRODUCTION
`
`I. Description
`11. History
`III. Comparison of Chromatographic Methods
`IV. Advantages of High Pressure Liquid
`Chromatography
`
`Chapter 2.
`
`INSTRUMENTATION
`
`1. Hardware
`
`II. Pumps
`III. Detectors
`IV. Columns
`
`V. Recorders, Fraction Collectors,
`and Integrators
`
`Chapter 3.
`
`EXPERIMENTAL METHODS
`
`I. Standardization of Conditions
`
`II. Operating Parameters
`III. Preparation of Sample
`
`vii
`
`IX
`Xi
`
`I
`11
`13
`
`14
`
`25
`
`27
`28
`44
`
`49
`
`52
`
`55
`78
`
`FRESENIUS KABI 1031-0003
`
`

`
`viii
`
`CON FENTS
`
`IV. Use of Fraction Collector
`V. Automation
`VI. Column Maintenance
`VII.
`Instrument Maintenance
`
`Chapter 4.
`
`IDENTIFICATION OF PEAKS
`
`I. Use of Retention Data
`II. Use of Internal Standards
`III.
`Isotopic Labeling
`IV. Collection and Characterization
`of Eluent Peaks
`
`V. Enzymic Peak—Shift
`VI. Derivatization
`
`Chapter 5.
`
`QUANTITATION
`
`I. Chromatography
`ll.
`Integration
`III. Calculation
`IV. Statistical Treatment of Data
`
`Chapter 6.
`
`APPLICATIONS
`
`I. General Applications
`II.
`Specific Applications
`
`BIBLIOGRAPHY
`
`Author Index
`
`Subject Index
`
`80
`81
`82
`84
`
`87
`96
`99
`
`100
`
`101
`1 1 1
`
`1 13
`115
`122
`126
`
`128
`130
`
`185
`191
`
`19 5
`
`FRESENIUS KABI 1o31—ooo4[
`
`

`
`CHAPTER 5
`
`QUANTITATION
`
`In order to perform quantitative analyses or kinetic studies by high
`pressure liquid chromatography, it is necessary to be able to determine
`accurately the concentration of one or more components present. With
`the measurement and calculation methods available, high pressure liquid
`chromatography is a highly accurate analytical technique, comparable
`in accuracy to gas chromatography. The quantitative results obtained
`depend on the complete process of analysis from preparation of the sample
`to interpretation of the results. As in a chain, the analysis is no better than
`the weakest link. Therefore, each step must be considered in developing
`techniques and interpreting the final results.
`There are four basic steps in quantitative liquid chromatography:
`(1) chromatography, (2) integration, (3) determination of sample composi-
`tion, and (4) statistical analysis of the data. In the chromatography step,
`an analog signal is generated by the detector and recorded in the form
`of chromatographic peaks. The area under the peak is converted to digital
`data in the integration process. Peak areas can be integrated either by
`manual methods or by using integrating devices. The calculation step con-
`sists of relating these data to the composition of the sample. In the final step,
`the data are expressed through statistics.
`
`112
`
`FRESENIUS KABI 1031-0005A
`
`

`
`1. Chromatography
`
`113
`
`I. Chromatography
`
`A. SAMPLE PREPARATION
`
`Using high pressure liquid chromatography in biochemical and bio-
`medical research, one of the most important parts of the analytical pro-
`
`the enzymes present in tissues. Pipetting precision is essential because
`large errors in accuracy can be caused by relatively small pipetting errors.
`In preparing cell extracts using the diethyl ether extraction of TCA from a
`sample, volume changes can cause as much as 10% difference in repro-
`ducibility. The presence of a large concentration of salts also causes dif-
`ferences in chromatograms which result in quantitative errors. In making
`up solutions of samples or standards, the careful calculation of molarity or
`concentration is necessary. Too concentrated a solution may cause over-
`loading of the column; thus poor peak shapes are obtained, or there will be
`problems in the subsequent chromatograms. Also errors in the final data can
`result from faulty weighings or dilutions of standards, reagents, or samples.
`Not only is sample preparation important, but care must be taken in sample
`storage because problems such as changes in concentration due to evapo-
`ration, contamination, and salting-out effects can prevent accurate analyses.
`In addition, some nucleotides may catalyze their own decomposition if not
`stored under appropriate pH conditions.
`It is ofthe utmost importance that not only the procedure for preparation
`and handling of the sample be standardized, but also the standard or re-
`ferences should be handled and stored in the very same way as the samples.
`The probability of accurate results will be increased if extreme care is taken
`in the standardization procedure.
`
`B. SAMPLE INTRODUCTION
`
`Since the sample volume used in high pressure liquid chromatography
`‘is so small, it is critically important that the sample be carefully and quanti-
`tatively introduced into the column. The syringe must be absolutely clean
`and not blocked, no air bubbles must be present, and the sample volume
`must be measured precisely. If possible, the syringe should be rinsed several
`times with the sample solution so that there is no dilution effect. The needle
`must be wiped dry to prevent extra drops from being injected. For this
`purpose, Kimwipes should be used to prevent contamination by a used
`
`FRESENIUS KABI 1031-0006
`
`

`
`1 14
`
`5. QUANTITA1 ION
`
`cloth. In most instruments, the needle fits right into the column so that no
`sample is lost during the injection process. If the stopped-flow technique
`of injection is used, the pressure is stopped for the injection process. The
`injection technique is not as critical using this technique as it is when the
`sample is injected into the flowing stream through a septum, i.e., pressure is
`not stopped. It is also important to make sure the syringe is emptied com-
`pletely. The sample must be injected slowly so that all the liquid is introduced
`evenly in the column, thus preventing the formation of a pool of sample at
`the injection port.
`
`C. SEPARATION OF PEAKS
`
`In the chromatography process, in which the compounds migrate dif-
`ferentially, there may be decomposition of components during the chro-
`matographic process. Care must be taken to avoid thermal or chemical
`degradation of any of the compounds present. Decomposition may lead to
`the presence of new peaks, the absence of expected peaks, or the masking of
`peaks normally in the chromatogram. If there is peak overlap or tailing,
`different eluents, column packings, or change in temperature should be
`tried. If, however, separations which were previously good deteriorate, it
`is possible that the column packing has changed. If the problem is due to an
`accumulation of impurities on the column, the column should be washed
`well. If, however, washing does not help, it is possible that the packing
`material has decomposed irreversibly and a new column should be tried. For
`an example of the poor separation caused by column deterioration see
`Fig. 3-23. In the upper chromatogram there is excellent separation of
`ATP and GTP in a cell extract of rat liver. However, these peaks in the
`same sample are poorly resolved in the lower chromatogram which had been
`run a few days before on a column that had been used for 2% years.
`One problem that always faces chromatographers is whether there is
`any retention of solutes on the column that will bleed off in subsequent
`runs. The process is referred to as “ghosting.” In order to determine whether
`or not there is “ghosting,” it is possible to run an experiment using isotopes.
`An experiment such as this was carried out by Brown (1970). A standard
`solution .of adenine nucleotides was run followed by a cell extract contain-
`ing “C adenine and guanine nucleotides. Fractions were collected for both
`the samples and the low concentrate wash after each sample. Another stan-
`dard solution of adenine nucleotides was run and fractions of this sample
`and the subsequent wash were also collected. The radioactivity of all the
`fractions from each solution was counted and it was found that peaks of
`the radioactive sample as plotted corresponded to the chromatogram from
`the analyzer (see Fig. 4—4). The wash solutions and the standard solution
`
`FRESENIUS KABI 1031-00074
`
`

`
`II. Integration
`
`115
`
`run both before and after the “C samples were found to have only back-
`ground levels of radioactivity. Therefore, it was concluded that the nucleo-
`tides were eluted completely.
`
`D. DETECTION AND AMPLIFICATION
`
`Important detector specifications are sensitivity, linearity, and speci-
`ficity. Errors in detection and amplification can be due to lack of familiarity
`with electronics. Knowledge of detector specificity is important in quanti-
`tative analysis. For example, all compounds do not absorb in the UV.
`Therefore, some compounds will not have peaks in a chromatogram ob-
`tained using a UV detector. Also, as was discussed in the section on detectors,
`column flow rate, detector electronics, and temperature all affect the quanti-
`tative results. It is important to know which parameters are critical with
`the detector being used so that proper precautions can be taken. It is neces-
`sary to know if the concentration ofthe compound to be quantitated is linear
`with the area of the peak. Modified Beer’s laws plots should be determined
`for each component in a sample. An example of the plot of concentration
`of AMP (in nanomoles) vs the area (in square centimeters) of the AMP is
`shown in Fig. 5-1.
`
`II. Integration
`
`The integration step consists of converting the detector signal into
`numerical data. Since the detector signal is usually in the form of peaks in
`a chromatogram from a strip chart recorder, it is necessary to convert the
`
`-NU‘-5GI05~l
`
`
`
`AREA(cmzl
`
`I_I_.;Ifl.L
`0.: 0.2 0.3 0.4 6.5 0.6 0.7 0.5 0.9 l.O LI‘
`CONCENTRATION lnmoles)
`
`Fig. 5-1. Plot of concentration (in nanomoles) of AMP vs area (cm’) of AMP peak.
`[Reproduced from Brown 0970).]
`
`FRESENIUS KABI 1031-0008
`
`

`
`116
`
`5. QUANTITATION
`
`peak area or heights to numbers. This can be done manually or by electronic
`techniques. If electronic techniques are used, neither a recorder nor an
`operator may be necessary for the conversion step. However, it is useful to
`have the chromatogram as it is sometimes possible to spot differences in peak
`areas or shapes that might not be detectable as easily with numbers alone.
`These differences could be due to actual reactions in the cell or cell extracts
`
`or could be symptoms of trouble in the chromatographic system. When a
`strip chart recorder is used, characteristics that can affect the accuracy of
`the results are the dead band, linear range, pen speed, shifting ofthe balance
`point, chart paper degree of filtering, or irregular pen movements. The first
`two characteristics are fully described in Varian’s “Basic Liquid Chroma-
`tography.”
`
`‘
`
`A. MANUAL METHODS
`
`The manual methods for converting detector signals into numerical
`data are: (1) planimetry; (2) measurement of peak height; (3) measurement
`of height times width at half-height; (4) triangulation; and (5) cut and weigh.
`A planimeter is a device which measures area by tracing the perimeter
`of the peak. (Fig. 5-2). The area is presented digitally on a dial. Since the
`precision and accuracy of this method depends not only on the device itself
`but largely on the steadiness and skill of the operator, the peaks should be
`traced several times and the results averaged. This method is not as accurate
`as the other methods. It is tedious and time-consuming as well. The second
`manual method is the measurement of peak height (Fig. 5—3). Although this
`procedure is simpler than the measurement of peak areas, it is not as accurate
`since peak height does not always change linearly with sample size. This is
`especially true if operating conditions vary or when the column is over-
`loaded. In the third method, the area is approximated by multiplying the
`height of the peak by the width at half—height (Fig. 5—4). The accuracy of this
`method is affected by the measurement of the width since a narrow peak can
`adversely affect the precision. Rather than using the base, the width at half-
`height is used to reduce the errors due to tailing of the peaks and any baseline
`irregularities. The fourth method is that of ‘triangulation in which a triangle
`is constructed by drawing tangents to the slope of the peak (Fig. 5-5). The
`area is calculated by the triangle formulaA = %BH, where the base is taken as
`the distance between the intersection of the two tangents with baseline and
`the height is measured by the shortest distance between the baseline and the
`intersection of the two tangents. With this method, it is assumed that the
`peak is symmetrical. The accuracy and precision of this method is dependent
`on the skill of constructing the triangle and the complexity of drawing the
`tangent lines. A slight error in the placement of the tangents can have a large
`
`FRESENIUS KABI 1031-0009“
`
`

`
`II. Integration
`
`117
`
`Fig. 5-2. Use of planimetcr.
`
`E
`
`Fig. 5-3. Measurement of peak height,
`
`
`
`Fig. 5-4. Measurement of area using peak height and width at half-height.
`
`FRESENIUS KABI 1031-0010
`
`

`
`118
`
`5 QUAN’I'l'IA'1“ION
`
`I
`
`\
`I
`/-4-—— B ——->\
`
`Fig. 5-5. Triangulation.
`
`effect on the results. The fifth manual method used is the cut—and-weigh
`method; the peak is cut out of the chromatogram and weighed on an analyt-
`ical balance. The constancy, weight and moisture content of the chart
`paper, and the skill in cutting are factors that affect the accuracy. More-
`over, the chromatogram is destroyed; this can be a serious disadvantage. It
`has been found that cutting and weighing of a xerox copy of the chromato-
`gram will minimize these disadvantages.
`
`B.
`
`INTEGRATING DEVICES
`
`A number of integrating devices have been designed for direct attach-
`ment to recorders so that the peak areas are recorded and integrated simul-
`taneously. Examples of these are electronic ball and disc, and analog and
`voltage integrating devices. One of the most widely used integrators in
`gas chromatography is the ball and disc type, manufactured by Disc Instru-
`ments Inc., because it provides automation relatively inexpensively. The
`accuracy and precision obtained is dependent on careful adjustment of the
`operation of the recorder and is limited by the mechanical performance
`of the recorder. However, although the results are excellent when the base-
`line of the chromatogram is steady and the separation of the peaks is good,
`the disc recorder is difficult to use under other conditions. The electronic
`digital integrator is one in which the chromatographic input signal is fed
`into a voltage to frequency converter and an output pulse rate, proportion to
`the peak area, is generated. The pulses from the voltage to frequency con—
`verter are accumulated and printed out as a measure of peak area when the
`
`FRESENIUS KABI 1031-0011*
`
`

`
`II. Integration
`
`119
`
`slope detector senses a peak. The major advantages ofthis integrator are the
`wide linear range, high count rate, and sensitive power detection. Although
`they are expensive, the cost may be worthwhile because of their high pre-
`cision, sensitivity, and the rapidity with which they work.
`Computers are being used to integrate peak areas. Although digital
`integrators are accurate and sensitive devices for measuring the peak areas,
`they do not handle the calculations or interpret the data. Computer systems
`that have been used include off-line, hybrid, dedicated computers, or multi-
`channel dedicated, or time- shared computers. In the off-line, batch-process-
`ing approach, the data on printed or punched paper tape from the digital
`integrator is manually transferred to a computer for processing. The hybrid
`has an integrator on-line to a computer. In the dedicated computer, a single
`computer is attached to a liquid chromatograph. This type ofsystem is used
`mainly for research studies. In the multichannel dedicated computer many
`chromatographs are on-line to one computer dedicated to this work. In the
`time-shared computer, chromatographs, as well as other analytical instru-
`ments, are on-line to one large computer. Time-shared services have made
`off-line computation available to small laboratories and can provide good
`computorial powers and storage of such data as compound name, retention
`time, and response factor.
`Burtis and Gere (1970) investigated six integration techniques and
`tabulated the precision of the results obtained by the various integration
`methods. The time required to perform the integration varied over a
`wide range. Ten chromatograms were picked at random for evaluation of
`these techniques and operators skilled in the various techniques performed
`the integration. The results (Table 5-1) show that the manual methods are
`expensive in terms of time expended on the integration. The precision ofthe
`digital integration was far better than the majority of the manual methods.
`The disc ® integrator, although not as fast or as precise as the digital inte-
`grator, was more rapid and more precise than manual methods. The digital
`integrator can correct the peak area for baseline drift and offset, sense and
`separate shoulders and overlapped peaks, and identify peaks by relative
`retention time. The on-line-computer also has many advantages. This system
`not only detects peaks by relative retention times, but it also calculates area
`percentage of each peak, calculates composition results in desired units,
`and types an output of the quantitative result. Although the time ofintegra-
`tion and precision are about as good as the digital integrators, its main
`advantage is that it has greater manipulative power.
`When quantitating nucleotide peaks, at least 10 samples of a standard
`solution of each nucleotide should be run. Because absorbance is directly
`related to molar extinction coefficients at the wavelength used, separate
`calibrations are required for each nucleotide. The concentration of a
`
`FRESENIUS KABI 1031-0012
`
`

`
`120
`
`5. QUANTITATION
`
`
`00.003
`
` 00.000.0_w.m:Bo.m><
`
`R0mi:_.00.000.:
`
`00.0$0
`
`00.02.0m
`
`$08.2
`
`0.3.0V0707$0NM:
`
`_S_m5
`
`35
`
`
`
`ammo?newEU
`
`$23:xm
`
`0o_§:w§§
`
`.HOuD~C1.—w_n.~
`
`Tm59$
`
`
`
`
`
`umoozhzzzc§<m0mez~mozomE<&2oU
`
` 1
`
`.83:050was05:5Eoumu
`
`
`
`
`._€00muufioxme
`
`ex
`
`30_2o
`
`.m><
`
`.
`
`$0.2o.92
`
`00.2o.m><
`
`2
`
`$0
`
`00._
`
`3.0
`
`03
`
`00.0E._
`
`3.0
`
`00.0
`
`Re
`
`8.:
`
`03:
`
`8.0
`
`3.0m
`
`8.2
`
`02:
`
`0300.0380.0
`
`00.03.3
`
`mm;03:
`
`N_.N5.0
`
`0300.8
`
`El3.:
`
`9._00.:
`
`,
`
`03N3200.0
`
`NS00.2
`
`8.000.0.
`
`Nam03
`
`00.08.00
`
`:03.2
`
`mm::2
`
`R003E00.0
`
`03052
`
`03N2:
`
`EM000
`
`mamN000
`
`0200.2
`
`:203:
`
`3WO
`
`00
`
`00.0
`
`$0
`
`03
`
`$6
`
`9&9
`
`8.0
`
`8.0
`
`.m><
`
`3.0
`
`3?
`
`3.3
`
`02:
`
`5.0
`
`0_.0~
`
`2.2
`
`0::
`
`008000.8.0008.00viiv0.3.0mV300N003;_080
`
`:
`
`00:2
`
`.03
`
`o_9E_007000
`
`0000
`
`swam
`
`00:3
`
`$00.0
`
`00.2
`
`$000
`
`AEECcom:8End.
`
`«Q8ovaoiooum
`
`FRESENIUS KABI 1031-0013
`
`
`
`
`
`

`
`121
`
`II. Integration
`
`TABLE 5-2
`
`TABULATION OF AREAS UNDER
`THE PEAKS OF A STANDARD
`SOLUTION OF AMP”"’
`______________
`
`Area/nmole
`Spectrum no.
`_________,_j_
`
`163
`171
`173
`174
`175
`176
`180
`181
`182
`183
`
`6.42
`6.81
`6,74
`6.87
`6.92
`6.49
`6.76
`6.81
`6.52
`6.69
`Mean 6.71
`
`0.18
`Standard deviation
`Coefficient of variation 2.8
`__._____________
`
`"From Brown (1970).
`[The AMP solution was 6.08 X
`10" M. The volume of sample
`solution injected ranged from 6
`to 8 ul.
`Instrument: Varian LCS
`1000. Starting volume: 50 ml. Flow
`rates:
`12 and 6 ml/hr. Eluents:
`0.015 M KHZPOA and 0.25 M
`KHZPO4 in 2.2 MKCI. UV output:
`0.08.
`
`nucleotide in a cell extract is then determined by relating the peak area to
`the peak area of 1 nmole of the standard nucleotide. In all cases, the same
`method of calculation should be used for standards as for samples. An
`example of the use of this method of calculating the peak area by multiply-
`ing the height times the width at half-height,
`is the AMP run by Brown
`(1970) (Table 5-2). The standard deviation of the 10 samples was 0.18 and
`the coefficient of variation 2.8%. The concentration of all adenine nucleo-
`tides and cell extracts can be calculated by relating peak areas to that ofthe
`standard AMP since all adenine nucleotides have the same molar extinction
`coefficient. A comparison of values of total adenine nucleotide content as
`determined by high pressure liquid chromatography and by an enzymic assay
`in four different blood samples showed close agreement. These values are
`shown in Table 5-3 (Brown, 1970).
`
`FRESENIUS KABI 1031-0014
`
`

`
`122
`
`5. QLANIITATION
`
`TABLE 5-3
`
`To'rAL ADENINh NUCLEOTIDES (AMP + ADP + ATP) (nrnoles/;il)"»b
`
`Sample
`
`P
`
`0.344
`0.332
`
`H
`
`0.387
`0.393
`
`A
`
`0.354
`0.348
`
`S
`
`0.229
`0.230
`
`Varian LCS 1000
`Enzymic analysis
`
`"From Brown (1970).
`1 mm X 3 m; packed with pellicular anion ex-
`blnstrumentz Varian LCS 1000. Colzmm:
`change resin. Eluents: 0.015 M Kl-IZPO4; 0.25 M KH2PO., in 2.2 M KCl. Flow rates: 12 ml/hr,
`6 ml/ hr. Starting volume of/ow concentration eluent: 50 ml. UV output: 0.08.
`
`III. Calculation
`
`In liquid chromatography, as in other chromatographic methods, the
`composition of the solute must be calculated. There are three principal
`methods used for relating the digital data to the composition of the sample:
`normalization, internal standardization, and calibration techniques.
`
`A. NORMALIZATION
`
`In the normalization method it is assumed that the entire sample is eluted
`
`and detected. The percentage of peak x, in Fig. 5-6, is given by the formula:
`A
`o
`_
`x
`/"x“Ax+Ay+AzX10O
`
`Ay, Ax, and AZ represent the individual peak areas. This method can be
`useful when analyzing complex mixtures. Since the relative peak areas
`obtained in liquid chromatography are not always related to composition,
`because the response of a‘ given detector may be different for each molecular
`type or class'of compound, it is necessary to use response factors. This will
`
`
`
`Fig. 5-6. Normalization.
`
`g
`
`l
`
`l
`
`FRESENIUS KABI 1031 -001 5‘.
`
`

`
`[1]. Calculation
`
`I23
`
`spectively, and F, is the response factor assigned to the reference compound.
`Corrected areas are obtained by multiplying the area by the relative response
`factor. Since different detectors operate on different principles, different
`factors must be calculated for different detectors.
`
`B.
`
`INTERNAL STANDARDS
`
`In order to eliminate apparatus and procedure errors, internal standards
`can be used. The requirements for an internal standard are as follows:
`1. It must be completely resolved from all the unknowns
`2. It must elute near the peak of interest
`3. It must be similar in concentration to the peak of interest
`4. It must be chemically similar but not present
`in the original
`
`The areas are represented by Ax and by A,-,, the correction action by
`Fx and ES of the unknown peak and internal standard, respectively,
`
`C. CALIBRATION
`
`In calibration, one external standard is used so that there can be direct
`comparison of the peak area of the sample to that of a standard which has
`
`“
`
`FRESENIUS KABI 1031-0016
`
`

`
`124
`
`5. QUANTITATION
`
`
`
`AREA(COMPONENT)
`
`STANDARD
`
`PERCENTAGE
`
`Fig. 5-7. Use of internal standard.
`
`been injected directly. A calibration factor K is determined by injecting the
`standard solutions and the results are expressed by the formula:
`
`%x = AXK
`
`This technique depends upon the precise injection of the standards. It is
`particularly useful in analyzing simple mixtures or for trace analysis. If the
`peak areas are plotted versus the known weight of the compound, this is
`known as absolute calibration (Fig. 5-8). If, however, the areas of the sample
`divided by the standard are plotted against the weight of the sample divided
`by the standard, this is known as a relative calibration curve (Fig. 5-9). In
`practice, an accurately known amount of an internal standard is added to
`an unknown sample and the mixture is chromatographed. The area ratios
`are measured and from the calibration graph, the weight ratios of sample to
`standard are obtained. Since a known amount of standard was added, the
`amount of unknown sample can be calculated. The advantages of this tech-
`nique are that the quantities injected need not be accurately measured and
`the detector response does not have to be known nor remain constant. The
`major disadvantage, especially in dealing with nucleotide pools of cell
`extracts, is the difficulty in finding a standard that does not interfere with
`the components in the sample.
`
`FRESENIUS KABI 1031-0017
`
`

`
`[1]. Calculation
`
`125
`
`A
`
`5
`
`WEIGHT OF COMPONENT
`
`l
`
`Fig. 5-8. Absolute calibration curve.
`
`
`
`STANDARD <:7>MP0FEfiT ‘
`
`COMPONENT
`WEIGHT (STANDARD )
`
`Fig. 5-9. Relative calibration curve.
`
`FRESENIUS KABI 1031-0018
`
`<L
`
`IJ
`01
`<[
`[LIJ
`0.
`
`K<
`
`TI
`
`-Z
`
`O
`LIJ
`[Z
`z <O O
`n. z
`2 <O l-0 (0
`\_/
`<tL|J
`
`I<
`
`

`
`126
`
`5. QUANTITATION
`
`IV. Statistical Treatment of Data
`
`A. DEFINITION OF TERMS
`
`In quantitative analysis, the terms accuracy, precision, repeatability,
`and reproducibility are frequently used. Accuracy is defined as the measure-
`ment of difference between a trial value and the true value. Precision is an
`expression of exactness or a measure of how well replicate values agree.
`Accuracy is difficult without precision but precision does not insure ac-
`curacy. However, precision with proper calibration gives accuracy. When a
`single operator repeats an analysis on the same apparatus, repeatability
`is the term used to define the precision. If the same analysis is run by dif-
`ferent operators on different instruments, possibly in different laboratories,
`the term reproducibility is used to define the precision.
`
`"B. SOURCES OF ERRoR
`
`In all measurements, there are errors. Errors can be divided into two
`categories: indeterminate and determinate. Indeterminate errors are ran-
`dom errors which cannot be eliminated and are inherent in the analytical
`technique. If minimized, they give high precision. Determinate errors are
`those errors whose cause and magnitude can be determined. If the determin-
`ate errors can be minimized, high accuracy is achieved. However, they can
`never be completely eliminated. In high pressure liquid chromatography,
`some determinate errors are: (1) poor sampling techniques; (2) decom-
`position on the column; (3) change in detector response; (4) recorder perfor—
`
`mance;(5)calculationerrors;and(6)operatorprejudiceandcarelessness.
`
`Errors in poor sampling techniques can result from poor preparation of
`the sample or improper storage which may cause decomposition or change
`in concentration. It is possible that all the sample is not injected completely
`into the column or that there is inaccurate measurement in the syringe.
`The sample may decompose on the column, causing inaccurate results, or
`there may be contamination of the sample before injection. It is even possible
`that the wrong'sample may be taken for analysis or the wrong solution may
`be prepared. The incorrect calculation of the concentration may cause large
`errors. Overloading of the column cannot only cause poor peak shape and
`masking of peaks in that particular chromatogram, but in many chromato-
`grams thereafter. As in all analytical techniques, complete and accurate
`labeling of samples is absolutely necessary. Changes in detector responses
`may be caused by changes in temperature or flow rates. Also it is important
`to realize that different detectors give different responses to the same
`
`‘
`:
`
`l
`
`FRESENIUS KABI 1031-0019
`
`

`
`IV. Statistical Treatment of Data
`
`127
`
`compound and there are different responses by one detector for different
`compounds. For example, when a UV detector is used, the absorbance is
`proportional to the molar extinction coefficient. Therefore, for each com-
`pound separate calibrations are required. It is also possible that the recorder
`performance is defective. Calculation and calibration errors are possible
`sources of inaccuracies. Calculation errors are especially common when
`there is peak overlap or a drifting baseline. There may be errors in the
`integrating operation. Another possible source of determinate errors is
`operator prejudice; the results may be predetermined in the mind of the
`operator. As was pointed out in Chapter 3, one of the most important
`factors in obtaining the best performance from a liquid chromatograph is
`an alert, careful, well-trained technician or operator who has the respon-
`sibility for the overall operation.
`
`"”
`
`FRESENIUS KABI 1031-0020