Chapter 15

Quality Assurance
Chapter Overview
15A The Analytical Perspective—Revisited
15B Quality Control
15C Quality Assessment
15D Evaluating Quality Assurance Data
15E Key Terms
15F Chapter Summary
15G Problems
15H Solutions to Practice Exercises

In Chapter 14 we discussed the process of developing a standard method, including optimizing

the experimental procedure, verifying that the method produces acceptable precision and
accuracy in the hands of a signal analyst, and validating the method for general use by the
broader analytical community. Knowing that a method meets suitable standards is important if
we are to have confidence in our results. Even so, using a standard method does not guarantee
that the result of an analysis is acceptable. In this chapter we introduce the quality assurance
procedures used in industry and government labs for the real-time monitoring of routine
chemical analyses.

953
954 Analytical Chemistry 2.1

Step 1. Identify and Define Problem Step 5. Propose Solution to Problem

What is the problem’s context? Is the answer sufficient?
What type of information is needed? Does answer suggest a new problem?

Step 2. Design Experimental Procedure Step 4. Analyze Experimental Data

Establish design criteria. Reduce and transform data.
Identify potential interferents. Complete statistical analysis.
Establish validation criteria. Verify results.
Select analytical method. Interpret results.
Establish sampling strategy.

Feedback
Loop

Step 3. Conduct Experiment & Gather Data

Calibrate instruments and equipment.
Standardize reagents.
Gather data.
Figure 15.1 Flow diagram showing one view of the analytical approach to solving problems. This dia-
gram is modified after Atkinson, G. F. J. Chem. Educ. 1982, 59, 201–202.
15A  The Analytical Perspective—Revisited
Figure 15.1 is the same as Figure 1.3. You
As we noted in Chapter 1, each area of chemistry brings a unique perspec-
may wish to review our earlier discussion tive to the broader discipline of chemistry. For analytical chemistry this
of this figure and of the analytical ap- perspective is as an approach to solving problem, one representation of
proach to solving problem.
which is shown in Figure 15.1.
If you examine an analytical method it often seems that its development
was a straightforward process of moving from a problem to its solution.
Unfortunately—or, perhaps, fortunately for those who consider themselves
analytical chemists!—developing an analytical method seldom is routine.
Even a well-established standard analytical method, carefully followed, can
yield poor data.
An important feature of the analytical approach in Figure 15.1 is the
feedback loop that includes steps 2, 3, and 4, in which the outcome of one
step may lead us to reevaluate the other steps. For example, after standard-
izing a spectrophotometric method for the analysis of iron (step 3), we may
find that its sensitivity does not meet our original design criteria (step 2). In
response, we might choose a different method, change the original design
criteria, or work to improve the sensitivity.
The feedback loop in Figure 15.1 is maintained by a quality assurance
program, whose objective is to control systematic and random sources of
 Chapter 15 Quality Assurance 955

error.1 The underlying assumption of a quality assurance program is that

An analysis is in a state of statistical con-
results obtained when an analysis is under statistical control are free of trol when it is reproducible and free from
bias and are characterized by well-defined confidence intervals. When used bias.
properly, a quality assurance program identifies the practices necessary to
bring a system into statistical control, allows us to determine if the system
remains in statistical control, and suggests a course of corrective action if
the system falls out of statistical control.
The focus of this chapter is on the two principal components of a qual-
ity assurance program: quality control and quality assessment. In ad-
dition, we will give considerable attention to the use of control charts for
monitoring the quality of analytical data.

15B  Quality Control

Quality control encompasses all activities that bring an analysis into sta-
For one example of quality control, see
tistical control. The most important facet of quality control is a set of writ- Keith, L. H.; Crummett, W.; Deegan, J.,
ten directives that describe relevant laboratory-specific, technique-specific, Jr.; Libby, R. A.; Taylor, J. K.; Wentler, G.
sample-specific, method-specific, and protocol-specific operations. Good “Principles of Environmental Analysis,”
Anal. Chem. 1983, 55, 2210–2218. This
laboratory practices (GLPs) describe the general laboratory operations article describes guidelines developed by
that we must follow in any analysis. These practices include properly re- the Subcommittee on Environmental An-
cording data and maintaining records, using chain-of-custody forms for alytical Chemistry, a subcommittee of the
American Chemical Society’s Committee
samples, specifying and purifying chemical reagents, preparing commonly on Environmental Improvement.
used reagents, cleaning and calibrating glassware, training laboratory per-
sonnel, and maintaining the laboratory facilities and general laboratory
equipment.
Good measurement practices (GMPs) describe those operations spe-
cific to a technique. In general, GMPs provide instructions for maintain-
ing, calibrating, and using equipment and instrumentation. For example, a
GMP for a titration describes how to calibrate the buret (if required), how
to fill the buret with titrant, the correct way to read the volume of titrant
in the buret, and the correct way to dispense the titrant.
The directions for analyzing a specific analyte in a specific matrix are
described by a standard operations procedure (SOP). The SOP indi-
cates how we process the sample in the laboratory, how we separate the
analyte from potential interferents, how we standardize the method, how
we measure the analytical signal, how we transform the data into the desired
result, and how we use quality assessment tools to maintain quality control.
If the laboratory is responsible for sampling, then the SOP also states how
we must collect, process, and preserve the sample in the field. An SOP
may be developed and used by a single laboratory, or it may be a standard
procedure approved by an organization such as the American Society for

1 (a) Taylor, J. K. Anal. Chem. 1981, 53, 1588A–1596A; (b) Taylor, J. K. Anal. Chem. 1983,
55, 600A–608A; (c) Taylor, J. K. Am. Lab October 1985, 53, 67–75; (d) Nadkarni, R. A.
Anal. Chem. 1991, 63, 675A–682A; (e) Valcárcel, M.; Ríos, A. Trends Anal. Chem. 1994, 13,
17–23.
956 Analytical Chemistry 2.1

Testing Materials or the Federal Food and Drug Administration. A typical

SOP is provided in the following example.

Example 15.1
Provide an SOP for the determination of cadmium in lake sediments using
atomic absorption spectroscopy and a normal calibration curve.
Solution
Figure 7.7 in Chapter 7 shows an example Collect sediment samples using a bottom grab sampler and store them
of a bottom grab sampler.
at 4 oC in acid-washed polyethylene bottles during transportation to the
laboratory. Dry the samples to constant weight at 105 oC and grind them
to a uniform particle size. Extract the cadmium in a 1-g sample of sediment
by adding the sediment and 25 mL of 0.5 M HCl to an acid-washed 100-
mL polyethylene bottle and shaking for 24 h. After filtering, analyze the
sample by atomic absorption spectroscopy using an air–acetylene flame,
a wavelength of 228.8 nm, and a slit width of 0.5 nm. Prepare a normal
calibration curve using five standards with nominal concentrations of 0.20,
0.50, 1.00, 2.00, and 3.00 ppm. Periodically check the accuracy of the
calibration curve by analyzing the 1.00-ppm standard. An accuracy of
±10% is considered acceptable.

Although an SOP provides a written procedure, it is not necessary to

follow the procedure exactly as long as we are careful to identify any modi-
fications. On the other hand, we must follow all instructions in a protocol
for a specific purpose (PSP)—the most detailed of the written quality
control directives—before an agency or a client will accept our results. In
many cases the required elements of a PSP are established by the agency that
sponsors the analysis. For example, a lab working under contract with the
Environmental Protection Agency must develop a PSP that addresses such
items as sampling and sample custody, frequency of calibration, schedules
for the preventive maintenance of equipment and instrumentation, and
management of the quality assurance program.
Two additional aspects of a quality control program deserve mention.
The first is that the individuals responsible for collecting and analyzing
the samples can critically examine and reject individual samples, measure-
ments, and results. For example, when analyzing sediments for cadmium
(see the SOP in Example 15.1) we might choose to screen sediment samples,
discarding a sample that contains foreign objects—such as rocks, twigs,
or trash—replacing it with an additional sample. If we observe a sudden
change in the performance of the atomic absorption spectrometer, we may
choose to reanalyze the affected samples. We may also decide to reanalyze
a sample if the result of its analysis clearly is unreasonable. By identifying
those samples, measurements, and results subject to gross systematic errors,
inspection helps control the quality of an analysis.
 Chapter 15 Quality Assurance 957

The second additional consideration is the certification of an analyst’s

competence to perform the analysis for which he or she is responsible. Be-
fore an analyst is allowed to perform a new analytical method, he or she
may be required to analyze successfully an independent check sample with
acceptable accuracy and precision. The check sample is similar in composi-
tion to samples that the analyst will analyze later, with a concentration that
is 5 to 50 times that of the method’s detection limit.

15C  Quality Assessment

The written directives of a quality control program are a necessary, but not
a sufficient condition for obtaining and maintaining a state of statistical
control. Although quality control directives explain how to conduct an
analysis, they do not indicate whether the system is under statistical control.
This is the role of quality assessment, the second component of a quality
assurance program.
The goals of quality assessment are to determine when an analysis has
reached a state of statistical control, to detect when an analysis falls out of
statistical control, and to suggest possible reasons for this loss of statistical
control. For convenience, we divide quality assessment into two categories:
internal methods coordinated within the laboratory, and external methods
organized and maintained by an outside agency.

15C.1  Internal Methods of Quality Assessment

The most useful methods for quality assessment are those coordinated by
the laboratory, which provide immediate feedback about the analytical
method’s state of statistical control. Internal methods of quality assessment
include the analysis of duplicate samples, the analysis of blanks, the analysis
of standard samples, and spike recoveries.

Analysis of Duplicate Samples

An effective method for determining the precision of an analysis is to an-
alyze duplicate samples. Duplicate samples are obtained by dividing a A split sample is another name for dupli-
single gross sample into two parts, although in some cases the duplicate cate samples created from a single gross
sample.
samples are independently collected gross samples. We report the results
for the duplicate samples, X1 and X2, by determining the difference, d, or
the relative difference, (d)r, between the two samples
d = X1 - X2
(d) r = d # 100
( X 1 + X 2) / 2
and comparing to an accepted value, such as those in Table 15.1 for the
analysis of waters and wastewaters. Alternatively, we can estimate the stan-
dard deviation using the results for a set of n duplicates
958 Analytical Chemistry 2.1

Table 15.1 Quality Assessment Limits for the Analysis of Waters and Wastewaters
(d)r when (d)r when
analyte [analyte] < 20�MDL (±%) [analyte] > 20�MDL (±%) spike recovery limit (%)
acids 40 20 60–140
anions 25 10 80–120
bases or neutrals 40 20 70–130
carbamate pesticides 40 20 50–150
herbicides 40 20 40–160
metals 25 10 80–120
other inorganics 25 10 80–120
volatile organics 40 20 70–130
Abbreviation: MDL = method’s detection limit
Source: Table 1020.I in Standard Methods for the Analysis of Water and Wastewater, American Public Health Association: Washington, D. C.,
18th Ed., 1992.
n
/d 2
i

s=
2n
i=1

where di is the difference between the ith pair of duplicates. The degrees of
freedom for the standard deviation is the same as the number of duplicate
samples. If we combine duplicate samples from several sources, then the
precision of the measurement process must be approximately the same for
each.

Example 15.2
To evaluate the precision for the determination of potassium in blood
serum, duplicate analyses were performed on six samples, yielding the fol-
lowing results in mg K/L.
duplicate X1 X2
1 160 147
2 196 202
3 207 196
4 185 193
5 172 188
6 133 119

Estimate the standard deviation for the analysis.

Solution
To estimate the standard deviation we first calculate the difference, d, and
the squared difference, d 2, for each duplicate. The results of these calcula-
tions are summarized in the following table.
 Chapter 15 Quality Assurance 959

duplicate d = X1 – X2 d 2

1 13 169
2 –6 36
3 11 121
4 –8 64
5 –16 256
6 14 196

Finally, we calculate the standard deviation.

s= 169 + 36 + 121 + 64 + 256 + 196 = 8.4

2#6

Practice Exercise 15.1

To evaluate the precision of a glucometer—a device a patient uses at
home to monitor his or her blood glucose level—duplicate analyses are
performed on samples drawn from five individuals, yielding the following
results in mg glucose/100 mL.
duplicate X1 X2
1 148.5 149.1
2 96.5 98.8
3 174.9 174.5
4 118.1 118.9
5 72.7 70.4
Estimate the standard deviation for the analysis.
Click here to review your answer to this exercise.
The Analysis of Blanks
We introduced the use of a blank in Chapter 3 as a way to correct the signal
for contributions from sources other than the analyte. The most common
A method blank also is called a reagent
blank is a method blank in which we take an analyte free sample through blank
the analysis using the same reagents, glassware, and instrumentation. A
method blank allows us to identify and to correct systematic errors due to
impurities in the reagents, contaminated glassware, and poorly calibrated
instrumentation. At a minimum, a new method blank is analyzed whenever The contamination of reagents over time
is a significant concern. The regular use
we prepare a new reagent, or after we analyze a sample with a high concen-
of a method blank compensates for this
tration of analyte as residual carryover of analyte may produce a positive contamination.
determinate error.
When we collect samples in the field, additional blanks are needed to
correct for potential sampling errors.2 A field blank is an analyte-free
2 Keith, L. H. Environmental Sampling and Analysis: A Practical Guide, Lewis Publishers: Chel-
sea, MI, 1991.
960 Analytical Chemistry 2.1

sample carried from the laboratory to the sampling site. At the sampling
site the blank is transferred to a clean sample container, which exposes it to
the local environment. The field blank is then preserved and transported
back to the laboratory for analysis. A field blank helps identify systematic
errors due to sampling, transport, and analysis. A trip blank is an analyte-
free sample carried from the laboratory to the sampling site and back to the
laboratory without being opened. A trip blank helps to identify systematic
errors due to cross-contamination of volatile organic compounds during
transport, handling, storage, and analysis.

Analysis of Standards
Another tool for monitoring an analytical method’s state of statistical con-
trol is to analyze a standard that contains a known concentration of analyte.
Table 4.7 in Chapter 4 provides a sum- A standard reference material (SRM) is the ideal choice, provided that the
mary of SRM 2346, a standard sample of SRM’s matrix is similar to that of our samples. A variety of SRMs are avail-
Gingko biloba leaves with certified values able from the National Institute of Standards and Technology (NIST). If a
for the concentrations of flavonoids, ter-
pene ketones, and toxic elements, such as suitable SRM is not available, then we can use an independently prepared
mercury and lead. synthetic sample if it is prepared from reagents of known purity. In all cases,
the analyte’s experimentally determined concentration in the standard must
fall within predetermined limits before the analysis is considered under
statistical control.

Spike Recoveries
One of the most important quality assessment tools is the recovery of a
known addition, or spike, of analyte to a method blank, a field blank, or a
sample. To determine a spike recovery, the blank or sample is split into
two portions and a known amount of a standard solution of analyte is
added to one portion. The analyte’s concentration is determined for both
the spiked, F, and unspiked portions, I, and the percent recovery, %R, is
calculated as
%R = F - I
A # 100
where A is the concentration of analyte added to the spiked portion.

Example 15.3
A spike recovery for the analysis of chloride in well water was performed
by adding 5.00 mL of a 250.0 ppm solution of Cl– to a 50-mL volumet-
ric flask and diluting to volume with the sample. An unspiked sample
was prepared by adding 5.00 mL of distilled water to a separate 50-mL
volumetric flask and diluting to volume with the sample. Analysis of the
sample and the spiked sample return chloride concentrations of 18.3 ppm
and 40.9 ppm, respectively. Determine the spike recovery.
 Chapter 15 Quality Assurance 961

Solution
To calculate the concentration of the analyte added in the spike, we take
into account the effect of dilution.
A = 250.0 ppm # 550.00 mL
.0 mL = 25.0 ppm
Thus, the spike recovery is
%R = 40.925 - 18.3 # 100 = 90.4%
.0

Practice Exercise 15.2

To test a glucometer, a spike recovery is carried out by measuring the
amount of glucose in a sample of a patient’s blood before and after
spiking it with a standard solution of glucose. Before spiking the sam-
ple the glucose level is 86.7 mg/100 mL and after spiking the sample
it is 110.3 mg/100 mL. The spike is prepared by adding 10.0 µL of a
25 000 mg/100mL standard to a 10.0-mL portion of the blood. What is
the spike recovery for this sample.
Click here to review your answer to this exercise.
We can use a spike recovery on a method blank and a field blank to
evaluate the general performance of an analytical procedure. A known con- Figure 15.2, which we will discuss in
centration of analyte is added to each blank at a concentration that is 5 to Section 15D, illustrates the use of spike
50 times the method’s detection limit. A systematic error during sampling recoveries as part of a quality assessment
program.
and transport will result in an unacceptable recovery for the field blank, but
not for the method blank. A systematic error in the laboratory, however,
affects the recoveries for both the field blank and the method blank.
Spike recoveries on a sample are used to detect systematic errors due to
the sample’s matrix, or to evaluate the stability of a sample after its collec-
tion. Ideally, samples are spiked in the field at a concentration that is 1 to
10 times the analyte’s expected concentration or 5 to 50 times the method’s
detection limit, whichever is larger. If the recovery for a field spike is unac-
ceptable, then a duplicate sample is spiked in the laboratory and analyzed
immediately. If the laboratory spike’s recovery is acceptable, then the poor
recovery for the field spike likely is the result of the sample’s deterioration
during storage. If the recovery for the laboratory spike also is unacceptable,
the most probable cause is a matrix-dependent relationship between the
analytical signal and the analyte’s concentration. In this case the sample is
analyzed by the method of standard additions. Typical limits for spike re-
coveries for the analysis of waters and wastewaters are shown in Table 15.1.

15C.2  External Methods of Quality Assessment

Internal methods of quality assessment always carry some level of suspi-
cion because there is a potential for bias in their execution and interpre-
962 Analytical Chemistry 2.1

tation. For this reason, external methods of quality assessment also play
an important role in a quality assurance program. One external method
of quality assessment is the certification of a laboratory by a sponsoring
agency. Certification of a lab is based on its successful analysis of a set of
proficiency standards prepared by the sponsoring agency. For example,
laboratories involved in environmental analyses may be required to analyze
standard samples prepared by the Environmental Protection Agency. A sec-
ond example of an external method of quality assessment is a laboratory’s
See Chapter 14 for a more detailed de-
scription of collaborative testing.
voluntary participation in a collaborative test sponsored by a professional
organization, such as the Association of Official Analytical Chemists. Fi-
nally, an individual contracting with a laboratory can perform his or her
own external quality assessment by submitting blind duplicate samples and
blind standards to the laboratory for analysis. If the results for the quality
assessment samples are unacceptable, then there is good reason to question
the laboratory’s results for other samples.

15D  Evaluating Quality Assurance Data

In the previous section we described several internal methods of quality
assessment that provide quantitative estimates of the systematic errors and
the random errors in an analytical method. Now we turn our attention
to how we incorporate this quality assessment data into a complete qual-
ity assurance program. There are two general approaches to developing a
quality assurance program: a prescriptive approach, in which we prescribe
an exact method of quality assessment, and a performance-based approach
in which we can use any form of quality assessment, provided that we can
demonstrate an acceptable level of statistical control.3

15D.1  Prescriptive Approach

With a prescriptive approach to quality assessment, duplicate samples,
blanks, standards, and spike recoveries are measured using a specific proto-
col. We compare the result of each analysis to a single predetermined limit,
taking an appropriate corrective action if the limit is exceeded. Prescriptive
approaches to quality assurance are common for programs and laboratories
subject to federal regulation. For example, the Food and Drug Administra-
tion (FDA) specifies quality assurance practices that must be followed by
laboratories that analyze products regulated by the FDA.
Figure 15.2 provides a typical example of a prescriptive approach to
quality assessment. Two samples, A and B, are collected at the sample site.
Sample A is split into two equal-volume samples, A1 and A2. Sample B is
also split into two equal-volume samples, one of which, BSF, is spiked in
the field with a known amount of analyte. A field blank, DF, also is spiked
with the same amount of analyte. All five samples (A1, A2, B, BSF, and DF)
are preserved if necessary and transported to the laboratory for analysis.
3 Poppiti, J. Environ. Sci. Technol. 1994, 28, 151A–152A.
 Chapter 15 Quality Assurance 963

no DF recovery yes
within limits

B> MDL, or
no DL recovery yes no B>0.1×[spike], and yes
within limits
B<10×[spike]

systematic error systematic error

in laboratory in field

yes A1 – A2 yes BSF recovery no

within limits within limits

no

valid yes A1 – B poor

within limits replication
data

no immediate no BSL recovery

systematic error within limits
poor
duplicate samples
yes

time-dependent
systematic error

Figure 15.2 Example of a prescriptive approach to quality assurance for laboratories monitoring
waters and wastewaters. Adapted from Environmental Monitoring and Support Laboratory, U.
S. Environmental Protection Agency, “Handbook for Analytical Quality Control in Water and
Wastewater Laboratories,” March 1979.
After returning to the lab, the first sample that is analyzed is the field
blank. If its spike recovery is unacceptable—an indication of a systematic
error in the field or in the lab—then a laboratory method blank, DL, is
prepared and analyzed. If the spike recovery for the method blank is unsat-
isfactory, then the systematic error originated in the laboratory; this is error
the analyst can find and correct before proceeding with the analysis. An
acceptable spike recovery for the method blank, however, indicates that the
systematic error occurred in the field or during transport to the laboratory,
casting uncertainty on the quality of the samples. The only recourse is to
discard the samples and return to the field to collect new samples.
If the field blank is satisfactory, then sample B is analyzed. If the result
for sample B is above the method’s detection limit, or if it is within the
range of 0.1 to 10 times the amount of analyte spiked into BSF, then a
964 Analytical Chemistry 2.1

spike recovery for BSF is determined. An unacceptable spike recovery for

BSF indicates the presence of a systematic error that involves the sample.
To determine the source of the systematic error, a laboratory spike, BSL, is
prepared using sample B and analyzed. If the spike recovery for BSL is ac-
ceptable, then the systematic error requires a long time to have a noticeable
effect on the spike recovery. One possible explanation is that the analyte
has not been preserved properly or it has been held beyond the acceptable
holding time. An unacceptable spike recovery for BSL suggests an immedi-
ate systematic error, such as that due to the influence of the sample’s matrix.
In either case the systematic errors are fatal and must be corrected before
the sample is reanalyzed.
If the spike recovery for BSF is acceptable, or if the result for sample
B is below the method’s detection limit, or outside the range of 0.1 to 10
times the amount of analyte spiked in BSF, then the duplicate samples A1
and A2 are analyzed. The results for A1 and A2 are discarded if the difference
between their values is excessive. If the difference between the results for A1
and A2 is within the accepted limits, then the results for samples A1 and B
are compared. Because samples collected from the same sampling site at the
same time should be identical in composition, the results are discarded if
the difference between their values is unsatisfactory and the results accepted
if the difference is satisfactory.
The protocol in Figure 15.2 requires four to five evaluations of quality
This is one reason that environmental test- assessment data before the result for a single sample is accepted, a process
ing is so expensive. that we must repeat for each analyte and for each sample. Other prescrip-
tive protocols are equally demanding. For example, Figure 3.7 in Chapter
3 shows a portion of a quality assurance protocol for the graphite furnace
atomic absorption analysis of trace metals in aqueous solutions. This pro-
tocol involves the analysis of an initial calibration verification standard and
an initial calibration blank, followed by the analysis of samples in groups of
ten. Each group of samples is preceded and followed by continuing calibra-
tion verification (CCV) and continuing calibration blank (CCB) quality
assessment samples. Results for each group of ten samples are accepted only
if both sets of CCV and CCB quality assessment samples are acceptable.
The advantage of a prescriptive approach to quality assurance is that all
laboratories use a single consistent set of guideline. A significant disadvan-
tage is that it does not take into account a laboratory’s ability to produce
quality results when determining the frequency of collecting and analyz-
ing quality assessment data. A laboratory with a record of producing high
quality results is forced to spend more time and money on quality assess-
ment than perhaps is necessary. At the same time, the frequency of quality
assessment may be insufficient for a laboratory with a history of producing
results of poor quality.
 Chapter 15 Quality Assurance 965

15D.2  Performance-Based Approach

In a performance-based approach to quality assurance, a laboratory is free
to use its experience to determine the best way to gather and monitor
quality assessment data. The tools of quality assessment remain the same—
duplicate samples, blanks, standards, and spike recoveries—because they
provide the necessary information about precision and bias. What a labora-
tory can control is the frequency with which it analyzes quality assessment
samples and the conditions it chooses to signal when an analysis no longer
is in a state of statistical control.
The principal tool for performance-based quality assessment is a con-
trol chart, which provides a continuous record of quality assessment data.
The fundamental assumption is that if an analysis is under statistical control,
individual quality assessment results are distributed randomly around a
known mean with a known standard deviation. When an analysis moves
out of statistical control, the quality assessment data is influenced by ad-
ditional sources of error, which increases the standard deviation or changes
the mean value.
Control charts were developed in the 1920s as a quality assurance tool
for the control of manufactured products.4 Although there are many types
of control charts, two are common in quality assessment programs: a prop-
erty control chart, in which we record single measurements or the means
for several replicate measurements, and a precision control chart, in which
we record ranges or standard deviations. In either case, the control chart
consists of a line that represents the experimental result and two or more
boundary lines whose positions are determined by the precision of the
measurement process. The position of the data points about the boundary
lines determines whether the analysis is in statistical control.

Constructing a Property Control Chart

The simplest property control chart is a sequence of points, each of which The more samples in the original control
represents a single determination of the property we are monitoring. To chart, the easier it is to detect when an
analysis is beginning to drift out of statisti-
construct the control chart, we analyze a minimum of 7–15 samples while cal control. Building a control chart with
the system is under statistical control. The center line (CL) of the control an initial run of 30 or more samples is not
chart is the average of these n samples. an unusual choice.
n
/X i

CL = X =
n
i=1

Boundary lines around the center line are determined by the standard de-
viation, S, of the n points
n
/ (X - X )
i
2

S= i=1
n-1
4 Shewhart, W. A. Economic Control of the Quality of Manufactured Products, Macmillan: London,
1931.
966 Analytical Chemistry 2.1

The upper and lower warning limits (UWL and LWL) and the upper and
lower control limits (UCL and LCL) are given by the following equations.
Why these limits? Examine Table 4.12 in UWL = CL + 2S
Chapter 4 and consider your answer to
this question. We will return to this point
LWL = CL - 2S
later in this chapter when we consider how UCL = CL + 3S
to use a control chart.
  LCL = CL - 3S
Example 15.4
Construct a property control chart using the following spike recovery data
(all values are for percentage of spike recovered).
sample: 1 2 3 4 5
result: 97.3 98.1 100.3 99.5 100.9
sample: 6 7 8 9 10
result: 98.6 96.9 99.6 101.1 100.4
sample: 11 12 13 14 15
result: 100.0 95.9 98.3 99.2 102.1
sample: 16 17 18 19 20
result: 98.5 101.7 100.4 99.1 100.3

Solution
The mean and the standard deviation for the 20 data points are 99.4% and
1.6%, respectively. Using these values, we find that the UCL is 104.2%, the
UWL is 102.6%, the LWL is 96.2%, and the LCL is 94.6%. To construct
the control chart, we plot the data points sequentially and draw horizontal
lines for the center line and the four boundary lines. The resulting property
control chart is shown in Figure 15.3.

Practice Exercise 15.3

A control chart is a useful method for monitoring a glucometer’s performance over time. One ap-
proach is to use the glucometer to measure the glucose level of a standard solution. An initial analysis
of the standard yields a mean value of 249.4 mg/100 mL and a standard deviation of 2.5 mg/100 mL.
An analysis of the standard over 20 consecutive days gives the following results.
day: 1 2 3 4 5 6 7 8 9 10
result: 248.1 246.0 247.9 249.4 250.9 249.7 250.2 250.3 247.3 245.6
day: 11 12 13 14 15 16 17 18 19 20
result: 246.2 250.8 249.0 254.3 246.1 250.8 248.1 246.7 253.5 251.0
Construct a control chart of the glucometer’s performance.
Click here to review your answer to this exercise.
 Chapter 15 Quality Assurance 967

110

105
UCL
percent recovery

UWL

100
CL

1
LWL
95 LCL

90
0 5 10 15 20
sample number
Figure 15.3 Property control chart for Example 15.4. The warning limits are
shown in orange and the control limits in red.
We also can construct a control chart using the mean for a set of repli-
cate determinations on each sample. The mean for the ith sample is
n rep

/X ij
j=1
Xi = n
rep When using means to construct a prop-
erty control chart, all samples must have
where Xij is the jth replicate and nrep is the number of replicate determina- the same number of replicates.
tions for each sample. The control chart’s center line is
n
/X i

CL = n
i=1

where n is the number of samples used to construct the control chart. To

determine the standard deviation for the warning limits and the control
limits, we first calculate the variance for each sample.
n rep

/ (X ij - X i) 2
j=1
s 2i = n rep - 1
The overall standard deviation, S, is the square root of the average variance
for the samples used to construct the control plot.
n
/s 2
i

S= n
i=1

The resulting warning and control limits are given by the following four
equations.
968 Analytical Chemistry 2.1

UWL = CL + 2S
n rep
LWL = CL - 2S
n rep
UCL = CL + 3S
n rep
LCL = CL - 3S
  n rep

Constructing a Precision Control Chart

A precision control chart shows how the precision of an analysis changes
over time. The most common measure of precision is the range, R, between
the largest and the smallest results for nrep analyses on a sample.
R = X largest - X smallest
The more samples in the original control
chart, the easier it is to detect when an To construct the control chart, we analyze a minimum of 15–20 samples
analysis is beginning to drift our of statis-
tical control. Building a control chart with while the system is under statistical control. The center line (CL) of the
an initial run of 30 or more samples is not control chart is the average range of these n samples.
an unusual choice. n
/R i

R= n
i=1

The upper warning line and the upper control line are given by the follow-
ing equations
UWL = fUWL # R
UCL = fUCL # R
where fUWL and fUCL are statistical factors determined by the number of
replicates used to determine the range. Table 15.2 provides representative
values for fUWL and fUCL. Because the range is greater than or equal to zero,
there is no lower control limit and no lower warning limit.

Table 15.2 Statistical Factors for the Upper Warning Limit and
the Upper Control Limit of a Precision Control Chart
replicates fUWL fUCL
2 2.512 3.267
3 2.050 2.575
4 1.855 2.282
5 1.743 2.115
6 1.669 2.004
 Chapter 15 Quality Assurance 969

0.7

0.6 UCL
0.5
UWL
0.4
range

0.3

0.2
CL
0.1

0.0 Figure 15.4 Precision control chart for Example

15.5. The warning limits are shown in orange
0 5 10 15 20
sample number and the control limits in red.

Example 15.5
Construct a precision control chart using the following ranges, each deter-
mined from a duplicate analysis of a 10.0-ppm calibration standard.
sample: 1 2 3 4 5
result: 0.36 0.09 0.11 0.06 0.25
sample: 6 7 8 9 10
result: 0.15 0.28 0.27 0.03 0.28
sample: 11 12 13 14 15
result: 0.21 0.19 0.06 0.13 0.37
sample: 16 17 18 19 20
result: 0.01 0.19 0.39 0.05 0.05

Solution
The average range for the duplicate samples is 0.176. Because two repli-
cates were used for each point the UWL and UCL are
UWL = 2.512 # 0.176 = 0.44
UCL = 3.267 # 0.176 = 0.57
The resulting property control chart is shown in Figure 15.4.

The precision control chart in Figure 15.4 is strictly valid only for
the replicate analysis of identical samples, such as a calibration standard
or a standard reference material. Its use for the analysis of nonidentical
samples—as often is the case in clinical analyses and environmental analy-
ses—is complicated by the fact that the range usually is not independent
of the magnitude of the measurements. For example, Table 15.3 shows the
970 Analytical Chemistry 2.1

Table 15.3 Average Range for the Concentration of Chromium

in Duplicate Water Samples
number of
[Cr] (ppb) duplicate samples R
5 to < 10 32 0.32
10 to < 25 15 0.57
25 to < 50 16 1.12
50 to < 150 15 3.80
150 to < 500 8 5.25
> 500 5 76.0
Source: Environmental Monitoring and Support Laboratory, U. S. Environmental Protection Agency,
“Handbook for Analytical Quality Control in Water and Wastewater Laboratories,” March 1979.
relationship between the average range and the concentration of chromium
in 91 water samples. The significant difference in the average range for
different concentrations of chromium makes impossible a single precision
control chart. As shown in Figure 15.5, one solution is to prepare separate
precision control charts, each of which covers a range of concentrations for
which R is approximately constant.

Interpreting Control Charts

The purpose of a control chart is to determine if an analysis is in a state of
statistical control. We make this determination by examining the location

(a) (b) (c) UCL

UWL

UCL
range

UWL
CL
CL
UCL
UWL
CL

sample number sample number sample number

Figure 15.5 Example showing the use of a precision control chart for samples that span a range of
analyte concentrations. The precision control charts are for (a) low concentrations of analyte; (b)
intermediate concentrations of analyte; and (c) high concentrations of analyte.
 Chapter 15 Quality Assurance 971

of individual results relative to the warning limits and the control limits, (a)
and by examining the distribution of results around the central line. If we
assume that the individual results are normally distributed, then the proba- UCL
bility of finding a point at any distance from the control limit is determined UWL

by the properties of a normal distribution.5 We set the upper and the lower

result
CL
control limits for a property control chart to CL ± 3S because 99.74% of LWL
a normally distributed population falls within three standard deviations of LCL

the population’s mean. This means that there is only a 0.26% probability
of obtaining a result larger than the UCL or smaller than the LCL. When sample number
a result exceeds a control limit, the most likely explanation is a systematic
error in the analysis or a loss of precision. In either case, we assume that the (b)
analysis no longer is in a state of statistical control.
Rule 1. An analysis is no longer under statistical control if any single UCL
UWL
point exceeds either the UCL or the LCL.

result
CL
By setting the upper and lower warning limits to CL ± 2S, we expect that
no more than 5% of the results will exceed one of these limits; thus LWL
LCL

Rule 2. An analysis is no longer under statistical control if two out of

three consecutive points are between the UWL and the UCL or
sample number
between the LWL and the LCL.
(c)
If an analysis is under statistical control, then we expect a random dis-
tribution of results around the center line. The presence of an unlikely pat-
UCL
tern in the data is another indication that the analysis is no longer under UWL
statistical control.
result
CL
Rule 3. An analysis is no longer under statistical control if seven consecu- LWL
tive results are completely above or completely below the center LCL

line.
Rule 4. An analysis is no longer under statistical control if six consecutive
sample number
results increase (or decrease) in value. Figure 15.6 Examples of property control
Rule 5. An analysis is no longer under statistical control if 14 consecutive charts that show a sequence of results—
results alternate up and down in value. indicated by the highlighting—that violate
Rule 6. An analysis is no longer under statistical control if there is any (a) rule 3; (b) rule 4; and (c) rule 5.
obvious nonrandom pattern to the results.
Practice Exercise 15.4
Figure 15.6 shows three examples of control charts in which the results in-
dicate that an analysis no longer is under statistical control. The same rules In Practice Exercise 15.3 you cre-
apply to precision control charts with the exception that there are no lower ated a property control chart for a
warning limits and lower control limits. glucometer. Examine your property
control chart and evaluate the glu-
Using Control Charts for Quality Assurance cometer’s performance. Does your
conclusion change if the next three
Control charts play an important role in a performance-based program of results are 255.6, 253.9, and 255.8
quality assurance because they provide an easy to interpret picture of the mg/100 mL?
statistical state of an analysis. Quality assessment samples such as blanks,
Click here to review your answer to
this exercise.
5 Mullins, E. Analyst, 1994, 119, 369–375.
972 Analytical Chemistry 2.1

standards, and spike recoveries are monitored with property control charts.
A precision control chart is used to monitor duplicate samples.
The first step in using a control chart is to determine the mean value
and the standard deviation (or range) for the property being measured
while the analysis is under statistical control. These values are established
using the same conditions that will be present during subsequent analyses.
Preliminary data is collected both throughout the day and over several days
to account for short-term and for long-term variability. An initial control
chart is prepared using this preliminary data and discrepant points identi-
fied using the rules discussed in the previous section. After eliminating
questionable points, the control chart is replotted. Once the control chart is
in use, the original limits are adjusted if the number of new data points is at
least equivalent to the amount of data used to construct the original control
chart. For example, if the original control chart includes 15 points, new
limits are calculated after collecting 15 additional points. The 30 points are
pooled together to calculate the new limits. A second modification is made
after collecting an additional 30 points. Another indication that a control
chart needs to be modified is when points rarely exceed the warning limits.
In this case the new limits are recalculated using the last 20 points.
Once a control chart is in use, new quality assessment data is added at
a rate sufficient to ensure that the analysis remains in statistical control. As
with prescriptive approaches to quality assurance, when the analysis falls
out of statistical control, all samples analyzed since the last successful verifi-
cation of statistical control are reanalyzed. The advantage of a performance-
based approach to quality assurance is that a laboratory may use its experi-
ence, guided by control charts, to determine the frequency for collecting
quality assessment samples. When the system is stable, quality assessment
samples can be acquired less frequently.

15E  Key Terms

control chart duplicate samples field blank
good laboratory practices good measurement method blank
practices
proficiency standard protocol for a specific quality assessment
purpose
quality assurance program quality control reagent blank
spike recovery standard operations statistical control
procedure
trip blank

15F Chapter Summary

Few analyses are so straightforward that high quality results are obtained
with ease. Good analytical work requires careful planning and an attention
to detail. Creating and maintaining a quality assurance program is one way
 Chapter 15 Quality Assurance 973

to help ensure the quality of analytical results. Quality assurance programs

usually include elements of quality control and quality assessment.
Quality control encompasses all activities used to bring a system into
statistical control. The most important facet of quality control is written
documentation, including statements of good laboratory practices, good
measurement practices, standard operating procedures, and protocols for
a specific purpose.
Quality assessment includes the statistical tools used to determine
whether an analysis is in a state of statistical control, and, if possible, to
suggest why an analysis has drifted out of statistical control. Among the
tools included in quality assessment are the analysis of duplicate samples,
the analysis of blanks, the analysis of standards, and the analysis of spike
recoveries.
Another important quality assessment tool, which provides an ongoing
evaluation of an analysis, is a control chart. A control chart plots a property,
such as a spike recovery, as a function of time. Results that exceed warning
and control limits, or unusual patterns of results indicate that an analysis is
no longer under statistical control.

15G Problems
1. Make a list of good laboratory practices for the lab that accompanies
this course, or another lab if this course does not have an associated
laboratory. Explain the rationale for each item on your list.

2. Write directives outlining good measurement practices for (a) a buret,

for (b) a pH meter, and for (c) a spectrophotometer.

3. A atomic absorption method for the analysis of lead in an industrial

wastewater has a method detection limit of 10 ppb. The relationship
between the absorbance and the concentration of lead, as determined
from a calibration curve, is

A = 0.349 # (ppm Pb)

Analysis of a sample in duplicate gives absorbance values of 0.554 and
0.516. Is the precision between these two duplicates acceptable based
on the limits in Table 15.1?

4. The following data were obtained for the duplicate analysis of a 5.00
ppm NO -3 standard.
sample X1 (ppm) X2 (ppm)
1 5.02 4.90
2 5.10 5.18
3 5.07 4.95
974 Analytical Chemistry 2.1

sample X1 (ppm) X2 (ppm)

4 4.96 5.01
5 4.88 4.98
6 5.04 4.97

Calculate the standard deviation for these duplicate samples. If the

maximum limit for the relative standard deviation is 1.5%, are these
results acceptable?

5. Gonzalez and colleagues developed a voltammetric method for the

determination of tert-butylhydroxyanisole (BHA) in chewing gum.6
Analysis of a commercial chewing gum gave a result of 0.20 mg/g. To
evaluate the accuracy of this results, the authors performed five spike
recoveries, adding an amount of BHA equivalent to 0.135 mg/g to
each sample. The experimentally determined concentrations of BHA
in these samples were reported as 0.342, 0.340, 0.340, 0.324, and
0.322 mg/g. Determine the percent recovery for each sample and the
mean percent recovery.

6. A sample is analyzed following the protocol shown in Figure 15.2, using

a method with a detection limit of 0.05 ppm. The relationship between
the analytical signal, Smeas, and the concentration of the analyte in parts
per million, CA, as determined from a calibration curve, is
S meas = 0.273 # C A
Answer the following questions if the limit for a successful spike recov-
ery is ±10%.
(a) A field blank is spiked with the analyte to a concentration of
2.00 ppm and returned to the lab. Analysis of the spiked field blank
gives a signal of 0.573. Is the spike recovery for the field blank ac-
ceptable?
(b) The analysis of a spiked field blank is unacceptable. To determine
the source of the problem, a spiked method blank is prepared
by spiking distilled water with the analyte to a concentration of
2.00 ppm. Analysis of the spiked method blank gives a signal of
0.464. Is the source of the problem in the laboratory or in the field?
(c) The analysis for a spiked field sample, BSF, is unacceptable. To de-
termine the source of the problem, the sample is spiked in the labo-
ratory by adding sufficient analyte to increase the concentration by
2.00 ppm. Analysis of the sample before and after the spike gives
signals of 0.456 for B and a signal of 1.03 for BSL. Considering this
data, what is the most likely source of the systematic error?
6 Gonzalez, A.; Ruiz, M. A.; Yanez-Sedeno, P.; Pingarron, J. M. Anal. Chim. Acta 1994, 285,
63–71.
 Chapter 15 Quality Assurance 975

7. The following data were obtained for the repetitive analysis of a stable
standard.7
sample Xi (ppm) sample Xi (ppm) sample Xi (ppm)
1 35.1 10 35.0 18 36.4
2 33.2 11 31.4 19 32.1
3 33.7 12 35.6 20 38.2
4 35.9 13 30.2 21 33.1
5 33.5 14 32.7 22 34.9
6 34.5 15 31.1 23 36.2
7 34.4 16 34.8 24 34.0
8 34.3 17 34.3 25 33.8
9 31.8

Construct a property control chart for these data and evaluate the state
of statistical control.

8. The following data were obtained for the repetitive spike recoveries of
field samples.8
sample % recovery sample % recovery sample % recovery
1 94.6 10 104.6 18 104.6
2 93.1 11 123.8 19 91.5
3 100.0 12 93.8 20 83.1
4 122.3 13 80.0 21 100.8
5 120.8 14 99.2 22 123.1
6 93.1 15 101.5 23 96.2
7 117.7 16 74.6 24 96.9
8 96.2 17 108.5 25 102.3
9 73.8

Construct a property control chart for these data and evaluate the state
of statistical control.

9. The following data were obtained for the duplicate analysis of a stable
standard.9
sample X1 (ppm) X2 (ppm) sample X1 (ppm) X2 (ppm)
1 50 46 14 36 36
2 37 36 15 47 45
7 Standard Methods for the Analysis of Waters and Wastewaters, American Public Health Association:
Washington, D. C., 18th Ed., 1992. The data is from Table 1030:I.
8 Standard Methods for the Analysis of Waters and Wastewaters, American Public Health Association:
Washington, D. C., 18th Ed., 1992. The data is from Table 1030:II.
9 Standard Methods for the Analysis of Waters and Wastewaters, American Public Health Association:
Washington, D. C., 18th Ed., 1992. The data is from Table 1030:I.
976 Analytical Chemistry 2.1

sample X1 (ppm) X2 (ppm) sample X1 (ppm) X2 (ppm)

3 22 19 16 16 20
4 17 20 17 18 21
5 32 34 18 26 22
6 46 46 19 35 36
7 26 28 20 26 25
8 26 30 21 49 51
9 61 58 22 33 32
10 44 45 23 40 38
11 40 44 24 16 13
12 36 35 25 39 42
13 29 31

Construct a precision control chart for these data and evaluate the state
of statistical control.

15H  Solutions to Practice Exercises

Practice Exercise 15.1
To estimate the standard deviation we first calculate the difference, d, and
the squared difference, d 2, for each duplicate. The results of these calcula-
tions are summarized in the following table.
duplicate d = X1 – X2 d 2
1 –0.6 0.36
2 –2.3 5.29
3 0.4 0.16
4 –0.8 0.64
5 2.3 5.29
Finally, we calculate the standard deviation.

s= 0.36 + 5.29 + 0.16 + 0.64 + 5.29 = 1.08

2#5
Click here to return to the chapter.

Practice Exercise 15.2

Adding a 10.0-µL spike to a 10.0-mL sample is a 1000-fold dilution; thus,
the concentration of added glucose is 25.0 mg/100 mL and the spike
recovery is
%R = 110.325- .0
86.7 # 100 = 94.4%

Click here to return to the chapter.

 Chapter 15 Quality Assurance 977

Practice Exercise 15.3

The UCL is 256.9, the UWL is 254.4, the CL is 249.4, the LWL is 244.4, 260
and the LCL is 241.9 mg glucose/100 mL. Figure 15.7 shows the resulting UCL
property control plot. 255 UWL

mg glucose/100 mL
Click here to return to the chapter. 250 CL

245
Practice Exercise 15.4 LWL
LCL
Although the variation in the results appears to be greater for the second 240

10 samples, the results do not violate any of the six rules. There is no evi- 235
dence in Figure 15.7 that the analysis is out of statistical control. The next 0 5 10 15 20
sample number
three results, in which two of the three results are between the UWL and
the UCL, violates the second rule. Because the analysis is no longer under
Figure 15.7 Property control plot for Prac-
statistical control, we must stop using the glucometer until we determine
tice Exercise 15.3 and for Practice Exercise
the source of the problem. 15.4.
Click here to return to the chapter.
978 Analytical Chemistry 2.1