Beruflich Dokumente
Kultur Dokumente
Chapter 5 analyses
Errors can sometimes be calamitous, as this picture of the famous train accident at Montparnasse
station in Paris illustrates. On October 22, 1895, a train from Granville, France, crashed through
the platform and the station wall because the brakes failed. The engine fell thirty feet into the
street below killing a woman. Fortunately, no one on the train was seriously hurt, although the
passengers were badly shaken. The story of the train derailment was featured in the children’s
story The Invention of Hugo Cabret, by Brian Selznick (2007) and part of Hugo’s nightmare in the
movie Hugo (2011), winner of 5 academy awards in 2012.
Errors in chemical analyses are seldom this dramatic, but they may have equally serious ef-
fects as described in this chapter. Among other applications, analytical results are often used in
the diagnosis of disease, in the assessment of hazardous wastes and pollution, in the solving of
major crimes, and in the quality control of industrial products. Errors in these results can have se-
rious personal and societal effects. This chapter considers the various types of errors encountered
in chemical analyses and the methods we can use to detect them.
M easurements invariably involve errors and uncertainties. Only a few of these are due
to mistakes on the part of the experimenter. More commonly, errors are caused by
faulty calibrations or standardizations or by random variations and uncertainties in results.
Frequent calibrations, standardizations, and analyses of known samples can sometimes be
ND/Roger Viollet/Getty Images used to lessen all but the random errors and uncertainties. However, measurement errors are
an inherent part of the quantized world in which we live. Because of this, it is impossible to
The term error has two slightly perform a chemical analysis that is totally free of errors or uncertainties. We can only hope
different meanings. First, error to minimize errors and estimate their size with acceptable accuracy.1 In this and the next two
refers to the difference between a
chapters, we explore the nature of experimental errors and their effects on chemical analyses.
measured value and the “true” or
“known” value. Second, error often The effect of errors in analytical data is illustrated in Figure 5-1, which shows results
denotes the estimated uncertainty in a for the quantitative determination of iron. Six equal portions of an aqueous solution with
measurement or experiment. a “known” concentration of 20.00 ppm of iron(III) were analyzed in exactly the same way. 2
1
Unfortunately these ideas are not widely understood. For example, when asked by defense attorney
Robert Shapiro in the O. J. Simpson case what the rate of error in a blood test was, Marcia Clark
the lead prosecutor replied that the state’s testing laboratories had no percentage of error because
“they have not committed any errors” San Francisco Chronicle, June 29, 1994, p. 4.
2
Although actual concentrations can never be “known” exactly, there are many situations in which
we are quite certain of the value, as, for example, when it is derived from a highly reliable reference
standard.
Copyright 2013 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
5A Some Important Terms 83
Note that the results range from a low of 19.4 ppm to a high of 20.3 ppm of iron. The average,
or mean value, x , of the data is 19.78 ppm, which rounds to 19.8 ppm (see Section 6D-1 for The symbol ppm stands for parts per
rounding numbers and the significant figures convention). million, that is, 20.00 parts of iron(III)
per million parts of solution. For
Every measurement is influenced by many uncertainties, which combine to produce a aqueous solutions, 20 ppm 5 20 mg/dL.
scatter of results like that in Figure 5-1. Because measurement uncertainties can never be
completely eliminated, measurement data can only give us an estimate of the “true” value.
However, the probable magnitude of the error in a measurement can often be evaluated. It is
❮ Measurement uncertainties
cause replicate results to vary.
then possible to define limits within which the true value of a measured quantity lies with a
given level of probability.
Estimating the reliability of experimental data is extremely important whenever we
collect laboratory results because data of unknown quality are worthless. On the other
hand, results that might not seem especially accurate may be of considerable value if
the limits of uncertainty are known.
Unfortunately, there is no simple and widely applicable method for determining the
reliability of data with absolute certainty. Often, estimating the quality of experimental
results requires as much effort as collecting the data. Reliability can be assessed in sev-
eral ways. Experiments designed to reveal the presence of errors can be performed. Stan-
dards of known composition can be analyzed, and the results compared with the known
composition. A few minutes consulting the chemical literature can reveal useful reliability
information. Calibrating equipment usually enhances the quality of data. Finally, statisti-
cal tests can be applied to the data. Because none of these options is perfect, we must
ultimately make judgments as to the probable accuracy of our results. These judgments
tend to become harsher and less optimistic with experience. The quality assurance of
analytical methods and the ways to validate and report results are discussed further in
Section 8E-3.
One of the first questions to answer before beginning an analysis is “What maximum error
can be tolerated in the result?” The answer to this question often determines the method
chosen and the time required to complete the analysis. For example, experiments to determine
whether the mercury concentration in a river water sample exceeds a certain value can
often be done more rapidly than those to accurately determine the specific concentration.
To increase the accuracy of a determination by a factor of ten may take hours, days, or even
weeks of added labor. No one can afford to waste time generating data that are more reliable
than is necessary for the job at hand.
x– = 19.8 xt = 20.0
Copyright 2013 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
84 Chapter 5 errors in chemical analyses
a xi
N
Example 5-1
Calculate the mean and median for the data shown in Figure 5-1.
Solution
5A-2 Precision
Precision is the closeness of results to Precision describes the reproducibility of measurements—in other words, the close-
others obtained in exactly the same way. ness of results that have been obtained in exactly the same way. Generally, the precision
Copyright 2013 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
5A Some Important Terms 85
5A-3 Accuracy
Accuracy indicates the closeness of the measurement to the true or accepted value Accuracy is the closeness of a mea-
and is expressed by the error. Figure 5-2 illustrates the difference between accuracy sured value to the true or accepted
and precision. Note that accuracy measures agreement between a result and the value.
accepted value. Precision, on the other hand, describes the agreement among several
results obtained in the same way. We can determine precision just by measuring rep-
licate samples. Accuracy is often more difficult to determine because the true value is
❮ The term “absolute” has a
different meaning in this text
than it does in mathematics. An
usually unknown. An accepted value must be used instead. Accuracy is expressed in absolute value in mathematics
terms of either absolute or relative error. means the magnitude of a
number ignoring its sign. As
Absolute Error used in this text, the absolute
The absolute error E in the measurement of a quantity x is given by the equation error is the difference between
an experimental result and an
E 5 xi 2 xt (5-3) accepted value including its sign.
where xt is the true or accepted value of the quantity. Returning to the data displayed
in Figure 5-1, the absolute error of the result immediately to the left of the true value The absolute error of a measurement
is the difference between the measured
value and the true value. The sign of
the absolute error tells you whether the
value in question is high or low. If the
measurement result is low, the sign is
negative; if the measurement result is
high, the sign is positive.
Copyright 2013 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
86 Chapter 5 errors in chemical analyses
of 20.0 ppm is 20.2 ppm Fe; the result at 20.1 ppm is in error by 10.1 ppm Fe.
Note that we keep the sign in stating the error. The negative sign in the first
case shows that the experimental result is smaller than the accepted value, and the
positive sign in the second case shows that the experimental result is larger than the
accepted value.
Relative Error
The relative error of a measurement is The relative error Er is often a more useful quantity than the absolute error. The
the absolute error divided by the true percent relative error is given by the expression
value. Relative error may be expressed
in percent, parts per thousand, or xi 2 xt
parts per million, depending on the Er 5 3 100% (5-4)
magnitude of the result. As used in xt
this chapter, relative error refers to the
relative absolute error. Relative random Relative error is also expressed in parts per thousand (ppt). For example, the relative
errors (relative uncertainties) are error for the mean of the data in Figure 5-1 is
discussed in Sections 6B and 8B.
19.8 2 20.0
Er 5 3 100% 5 21%, or 210 ppt
20.0
(x–2 – xt)
Analyst 2
Benzyl isothiourea
hydrochloride
(x–3 – xt)
Analyst 3
Figure 5-3 Absolute error in the Nicotinic acid
micro-Kjeldahl determination of
nitrogen. Each dot represents the error
(x–4 – xt)
associated with a single determination. Analyst 4
Each vertical line labeled (xi 2 xt) is Nicotinic acid
the absolute average deviation of the
set from the true value. (Data from
C. O. Willits and C. L. Ogg, J. Assoc. –2.0 –1.5 –1.0 – 0.5 0 0.5 1.0
Offic. Anal. Chem., 1949, 32, 561.) Absolute error (xi – xt), %N
Copyright 2013 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
5B Systematic Errors 87
numerical average for the data. Both the precision and the accuracy are poor for the
results of analyst 4.
Figures 5-1 and 5-3 suggest that chemical analyses are affected by at least two
types of errors. One type, called random (or indeterminate) error, causes data to be Random, or indeterminate, errors
scattered more or less symmetrically around a mean value. Refer again to Figure 5-3, affect measurement precision.
and notice that the scatter in the data, and thus the random error, for analysts 1 and 3
is significantly less than that for analysts 2 and 4. In general, then, the random error
in a measurement is reflected by its precision. Random errors are discussed in detail
in Chapter 6.
A second type of error, called systematic (or determinate) error, causes the mean Systematic, or determinate, errors
of a data set to differ from the accepted value. For example, the mean of the results in affect the accuracy of results.
Figure 5-1 has a systematic error of about 20.2 ppm Fe. The results of analysts 1 and 2
in Figure 5-3 have little systematic error, but the data of analysts 3 and 4 show sys-
tematic errors of about 20.7 and 21.2% nitrogen. In general, a systematic error in
a series of replicate measurements causes all the results to be too high or too low. An
example of a systematic error is the loss of a volatile analyte while heating a sample.
A third type of error is gross error. Gross errors differ from indeterminate and
determinate errors. They usually occur only occasionally, are often large, and may
cause a result to be either high or low. They are often the product of human errors.
For example, if part of a precipitate is lost before weighing, analytical results will be
low. Touching a weighing bottle with your fingers after its empty mass is determined
will cause a high mass reading for a solid weighed in the contaminated bottle. Gross
errors lead to outliers, results that appear to differ markedly from all other data in An outlier is an occasional result in
a set of replicate measurements. There is no evidence of a gross error in Figures 5-1 replicate measurements that differs
and 5-3. Had one of the results shown in Figure 5-1 occurred at, say, 21.2 ppm Fe, it significantly from the other results.
might have been an outlier. Various statistical tests can be performed to determine if
a result is an outlier (see Section 7D).
5B Systematic Errors
Systematic errors have a definite value, an assignable cause, and are of the same
magnitude for replicate measurements made in the same way. They lead to bias in Bias measures the systematic error
measurement results. Note that bias affects all of the data in a set in the same way associated with an analysis. It has a
and that it bears a sign. negative sign if it causes the results to
be low and a positive sign otherwise.
Instrumental Errors
All measuring devices are potential sources of systematic errors. For example, pipets,
burets, and volumetric flasks may hold or deliver volumes slightly different from
those indicated by their graduations. These differences arise from using glassware at a
Copyright 2013 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
88 Chapter 5 errors in chemical analyses
temperature that differs significantly from the calibration temperature, from distor-
tions in container walls due to heating while drying, from errors in the original cali-
bration, or from contaminants on the inner surfaces of the containers. Calibration
eliminates most systematic errors of this type.
Electronic instruments are also subject to systematic errors. These can arise from
several sources. For example, errors may emerge as the voltage of a battery-operated
power supply decreases with use. Errors can also occur if instruments are not cali-
brated frequently or if they are calibrated incorrectly. The experimenter may also
use an instrument under conditions where errors are large. For example, a pH meter
used in strongly acidic media is prone to an acid error as discussed in Chapter 21.
Temperature changes cause variation in many electronic components, which can
lead to drifts and errors. Some instruments are susceptible to noise induced from
the alternating current (ac) power lines, and this noise may influence precision and
accuracy. In many cases, errors of these types are detectable and correctable.
Method Errors
The nonideal chemical or physical behavior of the reagents and reactions on which an
analysis is based often introduce systematic method errors. Such sources of nonideality
include the slowness of some reactions, the incompleteness of others, the instability of
some species, the lack of specificity of most reagents, and the possible occurrence of side
reactions that interfere with the measurement process. As an example, a common method
error in volumetric analysis results from the small excess of reagent required to cause an
indicator to change color and signal the equivalence point. The accuracy of such an analy-
sis is thus limited by the very phenomenon that makes the titration possible.
Another example of method error is illustrated by the data in Figure 5-3 in which
the results by analysts 3 and 4 show a negative bias that can be traced to the chemical
nature of the sample, nicotinic acid. The analytical method used involves the decom-
position of the organic samples in hot concentrated sulfuric acid, which converts the
nitrogen in the samples to ammonium sulfate. Often a catalyst, such as mercuric ox-
ide or a selenium or copper salt, is added to speed the decomposition. The amount of
ammonia in the ammonium sulfate is then determined in the measurement step. Ex-
periments have shown that compounds containing a pyridine ring such as nicotinic
acid (see structural formula, page 86) are incompletely decomposed by the sulfuric
acid. With such compounds, potassium sulfate is used to raise the boiling tempera-
ture. Samples containing NiO or NiN linkages must be pretreated or subjected
to reducing conditions.3 Without these precautions, low results are obtained. It is
highly likely the negative errors, (x3 2 xt) and (x4 2 xt) in Figure 5-3 are systematic
errors resulting from incomplete decomposition of the samples.
Of the three types of systematic Errors inherent in a method are often difficult to detect and are thus the most
errors encountered in a chemical ❯ serious of the three types of systematic error.
analysis, method errors are
usually the most difficult to Personal Errors
identify and correct.
Many measurements require personal judgments. Examples include estimating the
position of a pointer between two scale divisions, the color of a solution at the end
Color blindness is a good
example of a limitation that ❯ point in a titration, or the level of a liquid with respect to a graduation in a pipet or
buret (see Figure 6-5 page 116). Judgments of this type are often subject to system-
could cause a personal error in
a volumetric analysis. A famous atic, unidirectional errors. For example, one person may read a pointer consistently
color-blind analytical chemist high, while another may be slightly slow in activating a timer. Yet, a third may be less
enlisted his wife to come to the sensitive to color changes, with an analyst who is insensitive to color changes tending
laboratory to help him detect
color changes at end points of
titrations. 3
J. A. Dean, Analytical Chemistry Handbook, New York: McGraw-Hill, 1995, section 17, p. 17.4.
Copyright 2013 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
5B Systematic Errors 89
Constant Errors
The effect of a constant error becomes more serious as the size of the quantity mea-
sured decreases. The effect of solubility losses on the results of a gravimetric analysis,
shown in Example 5-2, illustrates this behavior.
Example 5-2
The excess of reagent needed to bring about a color change during a titration
is another example of constant error. This volume, usually small, remains the same
regardless of the total volume of reagent required for the titration. Again, the relative
error from this source becomes more serious as the total volume decreases. One way
of reducing the effect of constant error is to increase the sample size until the error
is acceptable.
Proportional Errors
A common cause of proportional errors is the presence of interfering contaminants
in the sample. For example, a widely used method for the determination of copper
is based on the reaction of copper(II) ion with potassium iodide to give iodine (see
Sections 20B-2, 38H-3, and 38H-4.). The quantity of iodine is then measured and
is proportional to the amount of copper. Iron(III), if present, also liberates iodine
Copyright 2013 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
90 Chapter 5 errors in chemical analyses
from potassium iodide. Unless steps are taken to prevent this interference, high re-
sults are observed for the percentage of copper because the iodine produced will be
a measure of the copper(II) and iron(III) in the sample. The size of this error is fixed
by the fraction of iron contamination, which is independent of the size of sample
taken. If the sample size is doubled, for example, the amount of iodine liberated by
both the copper and the iron contaminant is also doubled. Thus, the magnitude of
the reported percentage of copper is independent of sample size.
4
S ee U.S. Department of Commerce, NIST Standard Reference Materials Catalog, 2011 ed., NIST
Special Publication 260, Washington, D.C.: U.S. Government Printing Office, 2011; see also http://
www.nist.gov.
Copyright 2013 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
5B Systematic Errors 91
the components in these materials has been determined in one of three ways: (1) by
analysis with a previously validated reference method, (2) by analysis by two or more
independent, reliable measurement methods, or (3) by analysis by a network of co-
operating laboratories that are technically competent and thoroughly knowledgeable
with the material being tested. Several commercial supply houses also offer analyzed
materials for method testing.5
Often, analysis of standard reference materials gives results that differ from the
accepted value. The question then becomes one of establishing whether such a differ-
ence is due to bias or to random error. In Section 7B-1, we demonstrate a statistical
test that can help answer this question.
© Kresimir Juraga
Independent Analysis
If standard samples are not available, a second independent and reliable analytical
method can be used in parallel with the method being evaluated. The independent Standard reference materials from
method should differ as much as possible from the one under study. This practice NIST.
minimizes the possibility that some common factor in the sample has the same effect
on both methods. Again, a statistical test must be used to determine whether any dif-
ference is a result of random errors in the two methods or due to bias in the method
❮ Indifficult
using SRMs, it is often
to separate bias from
ordinary random error.
under study (see Section 7B-2).
Blank Determinations
A blank contains the reagents and solvents used in a determination, but no analyte. A blank solution contains the solvent
Often, many of the sample constituents are added to simulate the analyte environ- and all of the reagents in an analysis.
ment, which is called the sample matrix. In a blank determination, all steps of the Whenever feasible, blanks may also
contain added constituents to simulate
analysis are performed on the blank material. The results are then applied as a cor- the sample matrix.
rection to the sample measurements. Blank determinations reveal errors due to inter-
fering contaminants from the reagents and vessels employed in the analysis. Blanks The term matrix refers to the
are also used to correct titration data for the volume of reagent needed to cause an collection of all the constituents
indicator to change color. in the sample.
WEB Statistical methods are extremely important, not only in chemistry but in all walks of life.
WORKS Newspapers, magazines, television, and the internet bombard us with confusing and often
misleading statistics. Go to www.cengage.com/chemistry/skoog/fac9, choose Chapter 5,
and go to the Web Works. There you will find a link to a website that contains an interest-
ing presentation of statistics for writers. Use the links there to look up the definitions of
mean and median. You will find some nice examples of salary data that clarify the distinc-
tion between the two measures of central tendency, show the utility of comparing the two,
and point out the importance of using the appropriate measure for a particular data set.
For the nine salaries given, which is larger, the mean or the median? Why are the two so
different in this case?
5
or example, in the clinical and biological sciences area, see Sigma-Aldrich Chemical Co., 3050 Spruce
F
St., St. Louis, MO 63103, or Bio-Rad Laboratories, 1000 Alfred Nobel Dr., Hercules, CA 94547.
Copyright 2013 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
92 Chapter 5 errors in chemical analyses
Copyright 2013 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.