EPA-650/4-74-005-m
NOVEMBER 1974
Environmental Monitoring Series
•X'X'X'X-XvX'X'X'X'X'X'X'X'XvX'X'
•X'XvX'X'X'X'XvX'X'X'XvX'XvX'X'
i
UJ
O
-------
-------
EPA-650/4-74-005-m
GUIDELINES FOR DEVELOPMENT
OF A QUALITY ASSURANCE PROGRAM:
VOLUME XIII-TEST FOR LEAD
IN GASOLINE BY ATOMIC
ABSORPTION SPECTROMETRY
by
Denny E. Wagoner, Franklin Smith, and D. Gilbert
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, N. C. 27709
Contract No. 68-02-1234
ROAP 26BGC
Program Element No. 1HA327
EPA Project Officer: Steven M. Bromberg
Quality Assurance and Environmental Monitoring Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
November 1974
-------
This report has been reviewed by the Environmental Protection Agency
and approved for publication. Approval does not signify that the
contents necessarily rei'ect the views and policies of the Agency,
nor does mention of traide names or commercial products constitute
endorsement or recommendation for use.
11
-------
TABLE OF CONTENTS
SECTION
I
II
III
IV
APPENDIX
A
B
C
D
INTRODUCTION
OPERATIONS MANUAL
2.0 GENERAL
2.1 EQUIPMENT SELECTION
2.2 REAGENTS
2.3 CALIBRATION
2.4 SAMPLE ANALYSIS
2.5 CALCULATIONS AND DATA REPORTING
QUALITY ASSURANCE PROCEDURES
3.0 GENERAL
3.1 FUNCTIONAL ANALYSIS OF THE MEASUREMENT METHOD
3.2 COLLECTION OF INFORMATION TO IDENTIFY TROUBLE
3.3 INDEPENDENT PREFORMANCE AUDIT
3.4 DATA QUALITY ASSESSMENT
REFERENCES
REFERENCE METHOD FOR THE DETERMINATION OF LEAD IN
GASOLINE BY ATOMIC ABSORPTION SPECTROMETRY
GLOSSARY OF SYMBOLS
GLOSSARY OF TERMS
CONVERSION FACTORS
PAGE
1
3
3
7
9
11
17
21
24
24
25
31
35
36
44
45
46
48
50
-------
LIST OF FIGURES
FIGURE NO. PAGE
1. Operational flow chart of the analytical method. 4
2. Basic components of an AA spectrometer. 5
3. Absorption process. 5
4. Sample AA spectrometer calibration curve. 16
5. Sample receiving record. 20
6. Calculation data record. 22
7. Laboratory test record. 23
8. Sample control chart for duplicate measurements of
gasoline samples. 33
9. Sample control chart for the measurement of standard
samples. 34
10. Example illustrating p < 0.10 and satisfactory data
quality. 42
11. Example illustrating p > 0.10 and unsatisfactory data
quality. 42
IV
-------
LIST OF TABLES
TABLE NO. PAGE
1 Lead determination checklist 6
2 Estimate means and standard deviations of important
error sources 28
3 Computation of mean difference, d, and standard
deviation of differences, s, 39
d
4 Sample plan constants, k for P{not detecting a lot
with proportion p outside limits L and U} < 0.1 43
-------
ABSTRACT
This document presents guidelines for a quality control program for
the determination of the total lead content of gasoline within the
concentration range of 0.010 to 0.10 g of lead/U.S. gals. These guide-
lines include:
1. Good operating practices,
2. Directions on how to assess performance and quality data,
3. Directions on how to identify trouble and improve data quality,
4. Directions to permit design of auditing activities.
This document is not a research report. It is designed for utilization
by laboratory personnel.
This work was submitted in partial fulfillment of contract 68-02-1234
by Research Triangle Institute under the sponsorship of the Environmental
Protection Agency. Work was completed as of December 1974.
vi
-------
SECTION I INTRODUCTION
This document presents guidelines for developing a quality assurance
program for the test for lead in gasoline by atomic absorption spectrometry.
For convenience of reference, this method as published by the Environmental
Protection Agency in the Federal Register, July 8, 1974, is reproduced as
appendix A of this report.
This document is divided into three sections.
Section I, Introduction. The Introduction lists the overall objectives
of a quality assurance program and delineates the program components neces-
sary to accomplish the given objectives.
Section II, Operations Manual. This manual sets forth recommended
operating procedures to insure the collection of data of high quality, and
instructions for performing quality control checks designed to give an
indication or warning that the measurement method is out of control and that
invalid data or data of poor quality are being collected, allowing for cor-
rective action to b^ taken before future measurements are made.
Section III, Quality Assurance Procedures. Suggested quality control
procedures, summation of the analytical method, and a statistical treatment
of existing data resulting from the published analytical techniques are
given. Based upon the existing knowledge of the method, quality control
procedures are recommended to achieve the highest quality data for the
determination of lead in gasoline by atomic absorption spectrometry.
The objectives of this quality assurance program for the method are to:
1. Minimize systematic errors (biases) and to monitor and control
precision within acceptable limits in the measurement process;
2. Provide routine indications for operating purposes of satisfactory
performance of personnel and/or equipment;
3. Provide for prompt detection and correction of conditions, either
instrumental or operational, that contribute to the collection
of poor quality data; and
&. Collect and record information necessary to describe the quality
of the data.
To accomplish the above objectives a quality assurance porgram must
contain the following components:
-------
1. Recommended operating procedures,
2. Training of personnel and evaluation of performance of personnel
and equipment,
3. Routine monitoring of the variables and parameters that may have
a significant effect on data quality,
4. Development of statements and evidence to qualify data and detect
defects, and
5. Action strategies to increase the level of precision/accuracy in
the reported data.
Implementation of a properly designed quality assurance program should
enable laboratory technicians to achieve and maintain an acceptable level
of precision and accuracy in their testing for total lead in gasoline. It
will allow a laboratory to report an estimate of the precision of its
measurements for each lead analysis.
Quality assurance guidelines for the analysis of lead in gasoline by
atomic absorption spectrometry as presented here are designed to insure
the collection of valid daita by prevention, detection, and quantification
of equipment and personnel variations in the laboratory through:
1. Recommended operating procedures as a preventive measure.
2. Quality control checks for rapid detection of undesirable
performance, and
3. A quality audit to verify independently the quality of the data.
The scope of this document has been purposely limited to that of a
laboratory manual.
-------
SECTION 11 OPERATIONS MANUAL
2.0 GENERAL
This manual sets forth recommended procedures for determination of
lead in gasoline by atomic absorption spectrometry. [Appendix B (Test for
Lead in Gasoline by Atomic Absorption Spectrometry) is reproduced from the
Federal Register, and is included as appendix A of this document.] This
method is virtually identical to "Standard Method of Test for Lead in
Gasoline by Atomic Absorption Spectrometry," No. 3237-73 published by the
ASTM (ref. 1). Quality control procedures and checks which are designed to
give an indication or warning that invalid or poor quality laboratory data
are being collected, are written as part of the operating procedures and
are to be performed by the operator on a routine basis. In addition, the
performance of special quality control procedures and/or analysis of stan-
dard samples as prescribed by the supervisor for assurance of data quality
may be required of the operator on a routine basis.
The sequence of operations to be performed for each laboratory analysis
is given in figure 1. Each operation or step in the method is identified
by a block. Quality checkpoints in the analysis process, for which appro-
priate quality control limits are assigned, are represented by blocks
enclosed by heavy lines.
The precision/accuracy of data obtained from this method depends upon
equipment performance and on the proficiency and conscientiousness with
which the analyst performs the various steps. Detailed instructions are
given for minimizing or controlling equipment error, and procedures are
recommended to minimize analyst error. Before using this document, the
analyst should study the method as reproduced in appendix A in detail. In
addition, the analyst should study the instrument operations manual fur-
nished by the manufacturer of the atomic absorption spectrometer. Formal
instruction (training course) in the operation of an atomic absorption spec-
trometer is recommended for the analyst.
It is assumed that all types of apparatus satisfy the reference-method
specifications and that the manufacturer's recommendations will be followed
when using a particular piece of equipment.
-------
EQUIPMENT SELECTION
1. SELECT THE EQUIPMENT ACCORDING TO THE
SPECIFICATIONS GIVEN IN THE REFERENCE
METHOD (APPENDIX A) AND ACCORDING
TO SUBSECTION 2.1
2. PERFORM VISUAL AND OPERATIONAL
CHECKS OF EQUIPMENT AND RECORD
NEW EQUIPMENT IN A
RECORD FILE (2.1.1)
REAGENT PREPARATION
3. PREPARE AND STORE REAGENTS AS
DIRECTED IN SUBSECTION 2.2
CALIBRATION
A. CALIBRATE THE EQUIPMENT AND
DEVELOP A WORKINS ANALYTICAL CURVE
ACCORDING TO SUBSECTION 2.3
SAMPLE ANALYSIS
5. IDENTIFY THE SAMPLE AND DOCUMENT
ACCORDING TO SUBSECTION 2.4.1
6. ANALYZE SAMPLE ACCORDING TO
SUBSECTION 2.4.3..
DATA PROCESSING
7. PLOT WORKING ANALYTICAL CURVE AND
DETERMINE THE LE/J) CONTENT BY
READING FROM THE GRAPH ACCORDING
TO SUBSECTION 2.5
8. VALIDATE DATA BY COMPARATIVE OR MEASURED
VALUES OF STANDABD SOLUTION AND REFERENCE
SAMPLES
9. REPORT DATA ACCORDING TO
SUBSECTION 2.5
EQUIPMENT
SELECTION
INSPECTION AND
DOCUMENTATION
REAGENT
PREPARATION
CALIBRATION
SAMPLE RECEIVING
AND IDENTIFICATION
SAMPLE
ANALYSIS
PLOT CURVE
DETERMINE
CONCENTRATIONS
VALIDATE
DATA
REPORT
DATA
Figure 1. Operational flow chart of the determination process.
-------
To conclude this Introduction, let us describe the components of an
atomic absorption (AA) spectrometer and the analytical determination of
lead in gasoline. The basic components of an AA spectrometer are a primary
source of radiation, usually a hollow cathode lamp (HCL), a method of pro-
ducing neutral atoms (in this method, an air acetylene flame is utilized),
a monochromator to isolate the resonant line, a detector (a photomultiplier,
PM, tube), and a means of amplification and readout of the analytical sig-
nal. The basic measurement method consists of radiation from a light source
(HCL) passing through the absorption medium. The neutral (ground state)
o °
atoms, Pb , absorb the resonant energy (2833 A); the dispersion unit isolates
the resonance line to be measured from the other lines emitted by the light
source, and the resulting radiation flux is measured by the detection sys-
tem. A schematic of the instrument and the absorption process are presented
in figures 2 and 3, respectively.
FLAME
Figure 2. Basic components of an AA spectrometer.
GROUND STATE
CHEMICAL
SPECIES
NUCLEUS
FLAME
^^£*. ^ LIGHT
+\ • ) + ENERGY
\—/ (2833«
NEUTRAL ATOM
EXCITED STATE
'/^^
•0
EXCITED ATOM
Figure 3. Absorption process.
-------
u
Of
c
o
41
4->
-s
W
a
»
•
*"">
si
O 4-*
1-1 U
1- 0
«C OL
^J qj
y5
O IB
a
^"^
to
1
I
t—
CO
1-4
s
o
fee
Ulcf
t3 oc
UICX
CC Ul
ce
ex
at
Z 0
o c
1— 4-> O
i ai _c
o o o
10 o
'P
ut
C3
ui -a
OL l-
• 0
0
Ul
3£L O>
S: c
-^ ^
cr -r-
ui • •" • ' •.
•:''••
;!;'." •••#•
A. A. Spectrometer
•—
'?)
. •' i
'
>
::':•.':•
Volumetric Flasks
(SI
?Tv^f
... ,"
*:1L-
V
a.
IX
u
i
ro
1
^^?f
^\\
,. +
..
"T
;.
h
[
:x&>.
•>
Micropipet
*r
T/.1
' ' >
j
L
;
X
f
.-...-• i
-•?-
^
Acid Wash/Solutiol
LT)
..•?!?.
!
1
c
<
:
J
to
ro
4-*
fO
3
cr
to
'^
$*% •"'
#' ''v-'
s
OD
ro
ro
cr
f~-
^;y|-
.•'' j
• .• 1
"
J
3
^
A_j
.
•
j
s
m
a:
to
CO
CO
o-
OC5
t
•.
ro
...j
t
^
i
, t
"!X':'
;
i
C
O i—
4J C
3 --
O O.
1/5
C
•o
ro C
01
m
i- JD
(-> Q_
3
— O
o
a o
ID
nro in
, c
t> C
4-*
n^ C^
a c:
a
0) C
A
••;;:|
•':
• •;'* ..
.! ' •£
••; ••:•••••"
i
1
i
Working Standard
1 a*A ^nlntinn
p
to
•'
!' • " •
':-:
,'f
} V
f
0 >
•f- S-
4-> 3
>>(-!
C 0
J <£••-
C t-
•r- J3
j "£^
j o re
3=t_
t
^
.. ,|
% 1
j
u^«
•'
"' :
-_,.
•
'
^^
•a
u.
:.;
***• ::'p^
- _ .:•
:•
. v
_
: "• '.
Reference Samples
t
01
-------
In the determination of lead in gasoline, the chemical pretreatment
(iodine and a quaternary ammonium salt) in MIBK allows for the determination
of the total lead content independent of the alkyl lead species and without
sample (burner) memory effects (ref. 2).
2.1 EQUIPMENT SELECTION
A listing of the required apparatus along with certain miscellaneous
equipment is given in table 1 in this section. Additional specifications,
criteria, and/or design features as applicable are given here to aid in the
selection of apparatus to insure the collection of data of consistent
quality. Procedures and, where applicable, checks for acceptance limits of
variability are given.
2.1.1 Atomic Absorption Spectrometer
2.1.1.1 Design Characteristics. The instrument used must be capable of
scale expansion, nebulizer adjustment equipped with a slot burner and premix
chamber for use with an air-acetylene flame. The spectrometer may be of
the single-beam variety, double-beam, or dual double-beam optics. The
double-beam system minimizes the effects of variation in lamp intensity,
detector sensitivity, optics, and electronics. A double-beam spectrometer
readily offers a stable baseline and virtually eliminates periods of stabil-
ization when changing light sources to another element. It is also recom-
mended that burners be adjustable on three axes. The instrument should have
features for monitoring and adjustment of the burner gas-flow rates and the
normal safety features to protect the operator.
Versatile burner position adjustment allows the instrument to be
optimized for a particular analysis.
Note 1; Instances have been reported of the drain tube collapsing
and causing flashback. It is recommended that this tube be
of corrugated polytetrafluorethylene (Teflon).
For a more thorough discussion of atomic absorption spectrometers and
their commercial suitability, a recent treatise such as in reference 3
should be consulted.
The wavelength mechanism should read directly in wavelength units (A),
be accurate, repeatable, and not drift with minor changes in temperature.
-------
In general, to avoid problems when dealing with sources with complex spectra»
o
the wavelength setting should be accurate to +2 A and reproducible to
o
HKL A. Readout systems may be digital or an analog system with meter and
recorder. A recorder gives a permanent recording of the raw data with time
and offers the advantage of rechecking data. A digital system on the other
hand is advantageous in handling large quantities of data.
Safety features which control the on-off sequence of the fuel and the
oxidant, minimum flow rates, and ignition of the flame are recommended
although they are not required to obtain a precise and accurate analysis.
3
A good ventilation system (_>_ 300 ft /min) is of the utmost importance.
Safety procedures in handling of cylinders, compressors, etc., must be
adhered to (ref. 4).
2.1.1.2 Acceptance Check. The spectrometer should be checked out according
to the manufacturer's instructions and calibrated according to the proce-
dures in subsection 2.3.
2.1.1.3 Documentation. Record in the receiving-record log book a descrip-
tion of the spectrometer (basic design, single- or double-beam, etc.), its
serial number, and the results of the acceptance check. Sign and date the
entry.
2.1.2 Volumetric Flasks and Pipets
2.1.2.1 Design Characteristics. All glassware shall be Class A, volumetric
glassware (ref. 5). An ample supply of volumetric flasks of the 50 m£,
100 m£, 250 m£ and 1.0 £ sizes should be available. The following volumet-
ric pipets are required: 2 m£, 5 m£, 10 m£, 20 m£, and 50 m£. It is recom-
mended that a sufficient supply be available to perform a complete set of
analyses. All pipetting should be done by a transfer device. Pipetting
by mouth is not recommended under any circumstances!
2.1.2.2 Acceptance Check. The volumetric glassware is inspected for
cracks, scratches, and damage and calibrated (optional) according to sub-
section 2.3.
2.1.2.3 Documentation. Record in the receiving-record log book a descrip-
tion of the glassware, its catalog or serial number, and the results of the
acceptance check. Sign and date the entry.
-------
2.1.3 Mlcropipet
2.1.3.1 Design Characteristics. A 100 y£ pipet of the Eppendorf type or
equivalent is required (+1 percent accuracy and reproducibility).
2.1.3.2 Acceptance Check. The micropipet is inspected for damage and
checked for accuracy and reproducibility according to subsection 2.3.
2.1.3.3 Documentation. Record in the receiving-record log book a descrip-
tion of the pipet, its catalog or serial number, and the results of the
acceptance check. Sign and date the entry.
2.1.4 Wrist Action Shaker (Optional)
2.1.4.1 Design Characteristics. A shaker (wrist action) equipped with
a timer should be used.
2.1.4.2 Acceptance Check. Check performance according to manufacturer's
specifications.
2.1.4.3 Documentat ion. Record in the receiving-record log book a descrip-
tion of the shaker, its serial number, and the results of the acceptance
check. Sign and date the entry.
2.2 REAGENTS
A listing of the required reagents, along with additional specifica-
tions, criteria, and complete description of their preparation is given.
In the event new reagents (new prepared solutions) are introduced during
the analysis procedure, verification of the calibration curve will be con-
sidered necessary. Liquid stock standards (PbCl_) will not be considered
as a substitute for preparation of stock solutions from the solid-reagent-
grade lead chloride.
In reagent preparation all chemicals, when available, should be ACS
analytical reagent grade or better. Unless otherwise indicated, it is
intended that all reagents shall conform to the specifications of the com-
mittee on analytical reagents of the American Chemical Society, where such
specifications are available. Other grades may be used, provided it is
first ascertained that the reagent is of sufficiently high purity to permit
its use without lessening the accuracy of the determination. This can be
established by analysis of a quality control sample (i.e., a sample whose Pb
-------
content is accurately known). Each lot of reagents must be checked by
establishing or checking tha working analytical curve before acceptance.
2.2.1 Distilled Water
Unless otherwise indicated, references to water shall be understood
to mean distilled water or tfater of equal purity. It is recommended that
distilled water be obtained crom an all-glass distillation apparatus. Water
of high purity can be obtained by distillation, followed by passing the
distilled water through a mixed-bed ion exchange resin (ref. 6).
2.2.2 Acid Wash
All glassware should be washed in 1/1 (vol to vol) nitric acid-water
solution prior to use in the analysis of samples. This acid wash should be
followed by several rinses with distilled water.
2.2.3 Aliquat 336 (tricapryl methyl ammonium chloride)
2.2.4 Aliquat 336/MIBK Solution (10 Percent v/v)
Dissolve and dilute 100 m£ (88.0 g) of Aliquat 336 with MIBK to one
i. Store in an amber bottle sealed with a polyethylene-lined cap.
2.2.5 Aliquat 336/MIBK Solution (1 Percent v/v)
Dissolve and dilute 10 m£ (8.8 g) of Aliquat 336 with MIBK to one
H. Store in an amber bottle.
2.2.6 Iodine Solution
Dissolve 3.0 g iodine crystals in toluene and dilute to 100 m£ with
the same solvent. This solution should be stored in an amber (brown)
bottle and sealed with a polyethylene-lined cap.
2.2.7 Lead-Sterile Gasoline
Gasoline containing less than 0.005 g Pb/gal.
2.2.8 Lead, Standard Solution (5.0 g, Pb/gal.)
The analytical balance should be checked before preparing the stan-
dard lead solution by weighing a standard weight between zero and 1 g. If
the measured and actual weights are within +0.4 mg, proceed with the reagent
preparation. Record the actual and measured weights in the laboratory log
book. If the weights differ by more than +0.4 mg, report to the lab super-
visor that the balance is out of calibration. Calibration of a balance
should be carried out by a certified manufacturer's representative.
10
-------
Dissolve 0.4433 g of lead chloride (PbCl ) previously dried at 105°C
for 3 hours in approximately 200 mH of 10-percent liquid 336/MIBK solution
(2.2.4) in a 250 m£ volumetric flask. The transfer of PbCl- and Aliquat
336 is facilitated by a powder funnel or its equivalent. Dilute to the
mark with the 10 percent Aliquat solution, mix, and store in a brown bottle
having a polyethylene-lined cap. This solution contains 1,321 yg Pb/m£,
which is equivalent to 5.0 g Pb/gal.
Note 2: In certain cases, difficulty has been experienced in dissolv-
ing the PbCl^. If this problem is experienced (j> 3 hours),
it is recommended that a new lot of reagent be purchased.
2.2.9 Lead, Standard Solution (1.0 g Pb/gal.)
By means of a pipet and a rubber bulb, accurately transfer 50.0 m£
of the 5.0 g Pb/gal (2.2.8) solution to a 250-mi volumetric flask. Dilute
to volume (to the graduation mark) with 1 percent Aliquat 336/MIBK solution.
Store in a brown bottle having a polyethylene-lined cap.
2.2.10 Lead, Standard Solution (0.02, 0.05, and 0.10 g Pb/gal.)
Transfer accurately by means of pipets 2.0, 5.0, and 10.0 m£ of the
1.0 g Pb/gal solution (2.2.9) to three 100 me volumetric flasks. Add 5.0
m£ of 1 percent Aliquat 336 solution to each flask; dilute to the mark with
MIBK (2.2.11). Mix well and store in dark bottles having polyethylene-
lined caps.
2.2.11 Methyl Isobutyl Ketone (MIBK)
4-methyl-2-pentanone; ACS analytical reagent grade, should be stored
in its original container and kept tightly sealed.
2.3 CALIBRATION
A listing of the required steps for preparing the working standard
solution is given in detail. A general (specific, when possible) descrip-
tion is rendered for the setup and calibration of an atomic absorption
spectrometer. The establishment of a working analytical curve is given in
a step-by-step procedure so that in later analyses one can apply a sequence
of steps to determine the concentration of lead in gasoline, be it an actual
field sample or a prepared standard. The working analytical curve should
never be omitted or assumed to be the same for subsequent analyses.
11
-------
2.3.1 Glassware Cleaning
An initial wash for all glassware should consist of a 2-hour soak
in the acid wash solution, followed by several rinses with distilled water.
2.3.2 Glassware Calibration (ref. 7) (Optional)
All volumetric glassware should be freed of water streaks before
being calibrated. Pipets ne^d not be dried; volumetric flasks should be
allowed to drain. The water used for calibration should be allowed to
reach thermal equilibrium with its surroundings. This can be ascertained
by monitoring the temperature of the water until no further change is noted.
An analytical balance is used for calibrations involving volumes of 50 m£
or less; a top-loader electronic balance is sufficient for volumes greater
than 50 m£.
2.3.2.1 Calibration of a Volumetric Pipet. Determine the weight of a
clean, dry receiver, usually a weighing bottle. Pipet an aliquot into the
receiver and weigh the receiver and contents to the nearest 0.1 milligram.
At 25°C the volume in mi delivered by the pipet is calculated by dividing
the weight of the difference in the receiver's empty and full weights in
grams by 0.9961 g/m£. (A variation of +5°C about 25°C results in only about
0.1 percent change in the density of water.)
The calibration should be performed at least three times and averaged.
2.3.2.2 Calibration of a Volumetric Flask. Place a clean, well-drained
flask on a top loader or laboratory balance and weigh to the nearest milli-
gram. Fill the flask to the mark and weigh to the nearest milligram.
Repeat this determination several times. Calculate the volume of the flask
from the difference in weights divided by 0.9961 g/m£ (average of three
determinations) and etch the volume of the flask on its surface.
2.3.3 Preparation of Working Standards
At least three working standards and a blank will be prepared using
the 0.02, 0.05, and 0.10 g Pb/gal standard lead solutions described in
subsection 2.2.10.
2.3.3.1 Step 1. To each of four 50-m£ volumetric flasks, the following
will be added:
12
-------
1. 30 mH of MIBK (2.2.11),
2. 5,0 m£ of the respective low-lead standard solution—i.e.,
5 m£ of either 0.02, 0.05, or 0.10 g Pb/gal (2.2.10),
3. 5.0 m£ of lead-free gasoline to the three working standards
and to the blank. The blank will contain only 5.0 m£ of lead-
free gasoline (<0.005 g Pb/gal.),
Stopper the flasks after each addition to avoid evaporation.
2.3.3.2 Step 2. Add immediately 0.1 m£ (100 u£) of iodine/toluene solu-
tion (2.2.6) by utilizing a 100 \iH Eppendorf pipet and mix well.
Note 3: EPA practice will be to mix well by shaking vigorously for
approximately 1 minute. It is suggested that this time of
mixing and degree of agitation be consistent. A wrist-action
shaker with timer would be advantageous for this procedure.
If a shaker is used, a small piece of parafilm across the
stopper will prevent the possibility of loss of sample.
2.3.3.3 Step 3. Add 5 m£ of 1 percent Aliquat 336 solution (2.2.5) and
mix.
2.3.3.4 Step 4. Dilute to the mark with MIBK and mix well. A general
mixing procedure would be to invert and shake for at least 12 times. An
alternative would be to employ the wrist-action shaker.
2.3.4 Preparation and Operations of an Atomic Absorption Spectrometer
o
The optimization of the atomic absorption instrument at 2833 A
(analytical line for determination of lead), adjustment of the gas mixture
(lean blue oxidizing flame), and control of the sample aspiration rate are
required for the instrument to be at optimum conditions and to give sensi-
tivity and stability for a precise and accurate analysis (refs. 8, 9). The
operation of an atomic absorption instrument should be by trained laboratory
technicians who are thoroughly versed in the operation of an atomic absorp-
tion instrument. The procedure for peaking (optimization) of the monochrom-
o
ator at 2833 A should be done according to the manufacturer's instructions.
The control of burner position, aspiration rate, and gas mixture should be
set following the manufacturer's instructions. This is only a guideline,
and each spectrometer must be adjusted on an individual basis.
13
-------
The instrument should be placed on a solid surface with enough work
area to allow manipulation of samples during the measurement phase. The
area should be free of dust and acid fumes. A vent is required to remove
fumes and vapors created in the flame. This will protect the analyst,
protect the instrument, and aid in stabilizing the flame.
The air supplied must be dry, and free from oil and dust. A small,
oil-less air compressor with a reservoir and a gas drying unit is recommended.
Note 4; Cylinder air Ls required, not oxygen. Burners not specifical-
ly designed for the use of oxygen are hazardous when oxygen
is used. Acetylene as supplied from the manufacturer is dis-
solved in acetone. Acetylene tanks should be replaced when
2 2
the tank pressure is below 5 kg/cm (71 Ib/in ). The tank
should always be mounted vertically.
Note 5: Acetylene and copper form an explosive mixture. Copper
hardware (tubing or fittings) must not be used with acetylene.
Cylinder valves must be closed to avoid the possibility of a pressure
regulator failing while unattended. A regulator will always fail in the
open position, thereby pressurizing the lines and creating a hazardous
situation. Cylinders should be strapped down while in use and removed from
the work area when empty to avoid danger to the laboratory operators.
A vessel (2-gallon capacity, wide mouth) in which the level of liquid
can be observed is required to gather the effluent from the burner drain.
The tip of the drain tube should always be below the level of liquid (water).
The vessel should be an inert material, preferably not glass.
2.3.4.1 Optimization of Spectrometer. After installation of the instrument,
a lead (Pb) hollow cathode lamp is inserted in the holder, switched on, and
adjusted to the recommended current. Turn on the instrument and allow it
to warm up. It is recommended that instruments, except for hollow cathode
lamps, be left on continuously. Allow the hollow cathode lamp to warm at
its operating current for 15 to 30 minutes. The best operating current is
usually the lowest current that requires only moderate instrument gain and
produces the most stability. The lamp position and the optical system are
adjusted to make the light beam fall on the entrance slit of the monochro-
meter (in most cases, the optical system and lamps are prealigned prior to
14
-------
customer delivery). The wavelength is set for the Pb line at 2833 A by
adjusting the monochrometer until a maximum response is attained.
2.3.4.2 Flame Ignition. Ignite the flame and allow the burner to reach
thermal equilibrium by giving it an initial 10-minute warmup. In igniting
the flame, certain rules are essential. The oxidant (air) is turned on
first and the fuel (CJA.^) second during ignition. Shut down requires the
fuel (C2H2) to be turned off first, followed by the oxidant (air).
Using reagent blank (2.3.3.1, 2.3.3.2) adjust the gas mixture and the
sample aspiration rate until a lean (blue) oxidizing flame is obtained. Do
not adjust flame conditions on pure MIBK. This will result in the blank
solution and subsequent analyses having an inner cone (fingers) of unburned
hydrocarbons. In most cases, it is necessary to increase the oxidant (air)
flow rate or to reduce the sample flow rate (aspiration rate) to obtain a
sharply defined blue cone. Failure to follow this procedure can result in
poor precision.
2.3.4.3 Burner Adjustment. While aspirating the 0.10 g Pb/gal working
standard (2.3.3.1, 2.3.3.2), adjust burner height for maximum absorption.
Also adjust the burner forward and back to achieve maximum sensitivity.
Some instruments require the use of scale expansion to produce a response
of 0.150 to 0.170 absorbance for the 0.10 g Pb/gal standard. If an
instrument requires scale expansion, the 0.10 g Pb/gal standard should be
adjusted to read 0.150 absorbance units.
2.3.4.4 Development of a Working Curve.
Step 1. Aspirate the reagent blank (2.3.3.1, 2.3.3.2) and adjust the
instrument to read zero absorbance.
Step 2. Aspirate the three working standards. The blank should be
aspirated after each standard.
Step 3. Check for linearity. For example, if the 0.10 g Pb/gal stan-
dard reads 0.150 absorbance units then the 0.05 and 0.02 standards should
read 0.075 and 0.030 absorbance units, respectively (fig. 4).
Step 4. A working calibration curve should be prepared with absorbance
plotted versus the three working standard concentrations (fig. 4). The
curve should be linear within this concentration range. If a linear curve
is not obtained, additional working standards between 0.01 and 0.10 g Pb/gal
15
-------
W
et!
Q
CO
M
f j , t ^ • .- ->-j, - .,- ^-(-4-4.1^,4-144^ [J i+4j-t-it- - -t-t"p<4-K-i--
- -, _:_•!.' .?--^HJf+Rfl.^4J4i4g;l4J5SS±
-------
should be prepared by appropriate dilutions according to this section
(2.3.3.1, 2.3.3.2). If the problem is ascertained to be instrumental, the
instrument must be repaired. (Data points deviating more than about +0.002
g Pb/gal from the best-fit curve should not be expected.) Chemical stan-
dards and new reagents should be prepared, and the working analytical curve
repeated before calling the instrument manufacturer.
2.4 SAMPLE ANALYSIS
A listing of the required steps in sample analysis is given. Sample
documentation, including receiving and analysis, in the laboratory are
rendered to prevent confusion and loss of identity of the sample. Sample
handling procedures are recommended to prevent unwanted bias in the method.
2.4.1 Recommended Quality Control Practices
As part of the analysis procedures, certain checks are recommended
for quality control and for data-quality-assessment purposes. Procedures
for performing the checks are given along with recommendations for a minimum
acceptable frequency for performing the checks. If the checks indicate
problems in the analysis of gasoline samples, the frequency of the checks
should be increased until the problem is identified and corrective action
taken.
To aid in understanding the following discussion, five types of samples
referred to in the analysis procedure are defined here. They are:
1. Gasoline sample; A gasoline sample collected in the field for
analysis of its Pb content,
2. Blank sample: A working standard sample containing 5.0 m£ of
lead-free gasoline (< 0.005 g Pb/gal) and no standard lead solu-
tion (see 2.3.3.1) used to establish the baseline for the AA
spectrometer.
3. Standard reference sample (SRS): A gasoline sample taken from a
stock of gasoline the Pb content of which has been certified by
the NBS or other suitable agencies such as the Quality Assurance
and Environmental Monitoring Laboratory of the EPA. An SRS can
serve as a quality control sample when its concentration is known
by the analyst. However, its primary use will be as a blind
17
-------
sample (i.e., the analyst does not know the concentration) for
performance audits,
4. Quality control sample; A gasoline sample for which the Pb con-
tent is known to the analyst and analyzed periodically for
quality-control purposes. Preferably this would be an SRS;
however, if such samples are not available, duplicate results from
gasoline samples can be used.
5. Working standard samples: Samples prepared by the analyst from
stock reagents and used as calibration standards.
The following sequence of analyses is recommended as the minimum level
acceptable for quality control and data-quality-assessment purposes.
1. The spectrometer calibration curve should be verified by analyzing
the working standard samples a) before analyzing gasoline samples,
b) at the end of a 4-hour working period or at the end of a mea-
surement cycle if less than 4 hours. If all points fall within
the +0.004 g Pb/gal limits given in figure 4 and are randomly
distributed about the curve, the calibration is acceptable. If
any point falls outside the limits, the calibration curve should
be reconstructed before continuing. Reanalysis of gasoline sam-
ples is necessary when the results of the previous analysis com-
bined with the apparent drift of the spectrometer indicate that
the actual Pb content in the samples could be near the standard:
i.e., 0.05 g Pb/gal.
2. Every tenth sample analyzed should be a quality-control sample.
If SRS's are available for this purpose the results should be
plotted on a control chart as illustrated in figure 8. (Note
the action limit of +0.005 g Pb/gal can be used until sufficient
data are obtained (e.g., 20 to 30 measurements) to compute new
limits which should be less than those for duplicates.) Other-
wise, a second aliquot should be taken from the first, eleventh,
twenty-first gasoline samples and analyzed as the tenth,
twentieth, and thirtieth gasoline samples respectively. Results
should be plotted on a control chart as directed in 3.2.1.
3. A blind sample (i.e., an SRS provided by the supervisor) should
be analyzed a) after every 20 gasoline samples or for every
18
-------
measurement cycle if less than 20 samples are to be analyzed, and
b) any time a gasoline sample is found to exceed the standard.
2.4.1 Sample Receiving Procedure
Samples as received from the field must be labeled with at least
the following information:
1. Time of station test,
2. Date of station test,
3. Location,
4. Sample number,
5. Inspector name and title,
6. Inspector signature.
This data should be recorded in a sample receiving log book as shown in
figure 5. At this time, each sample should be assigned a laboratory analy-
sis number. This should be recorded in the sample receiving log book and
on the chain-of-custody label.
2.4.3 Sample Handling Procedures
The samples should be handled and stored in accordance with safety
procedures for flammable liquids. Since light fractions of a gasoline
sample evaporate easily, it is recommended that the sample be stored under
refrigeration. Before analysis, the sample should reach room temperature.
The sample must be tightly covered at all times except when removing aliquots
for analysis.
2.4.4 Analysis Procedure
Step 1. For each gasoline working standard sample, or SRS sample,
transfer 30 m£ MIBK into a 50 m£ volumetric flask and add 5.0 m£ (pipet
with rubber bulb or equivalent transfer device) of the sample and mix. The
gasoline should be allowed to reach room temperature (25°C) before the
aliquot is removed.
Step 2. Refer to subsection 2.3 and perform the analysis according
to (2.3.3.2-2.3.3.4).
Step 3. Aspirate (suction capillary is immersed in the solution) the
samples and working standards (2.3), and record the absorbance values with
frequent checks of the zero. It is recommended that the blank be aspirated
19
-------
SAMPLE RECEIVING RECORD
1. Test Station Sample [dentiflcation
Test Station Sample Number:
Test Station Location;
Region:
Date of Station Test: / /
MO DA YR
Time of Station Test: :
HR MIN
Inspector's Name:
Inspector's Title:
2. Laboratory Identification
Laboratory Name:
Laboratory Address:
Laboratory Analysis Number:
3. Laboratory Technician
Name:
Signature:
Date:
Figure 5. Sample receiving record.
20
-------
between each sample. It is recommended that the quality-control procedures
given in section 2.4.1 be followed during the analysis of samples.
During the course of analysis (measurement phase), difficulties may
arise. An indication of problems can be detected by changes in absorbance
of the blank and/or a standard sample (working standard). Another recom-
mended check is to analyze an SRS with known concentration for every 10
gasoline samples or to aspirate the initial tenth gasoline sample with each
succeeding group of 10 samples. If the absorbance is low and unsteady, the
atomizer may be blocked. The atomizer must then be cleaned or replaced.
If the flame looks unsteady or colored when spraying the blank, the burner
should be cleaned. The instrument operation manual (manufacturer) should
be consulted for cleaning procedures.
The standard solutions (working analytical curve) should be aspirated
at the end of a 4-hour work period or at the end of each measurement cycle.
2.5 CALCULATIONS AND DATA REPORTING
Plot the absorbance values versus concentrations for the working stan-
dards as illustrated in figure 4. Determine the concentrations of the
samples from the calibration curve by reading the g Pb/gal corresponding to
the sample absorbance. The calculations are complete when the calibration
data record, shown in figure 6, is recorded in the laboratory test record
log book.
Concentrations of lead should be reported to the nearest 0.001 g Pb/gal.
The total lead in gasoline is to be recorded in the laboratory test record
log book in the format shown in figure 7. The data report is not complete
until the laboratory test record is filled out.
21
-------
CALCULATION DATA RECORD
1. Test Station Sample Identification
Test Station Sample Number:
Region:
Date of Station Test: / /
Mo Da Yr
Time of Station Test: /
Hr Min
2. Laboratory Identification
Laboratory Name:
Laboratory Address:
Laboratory Analysis Number:
3. Calculation Data
Lead Calibration Curve' Graph Number:
Absorbance of the Sample:
Absorbance of the Blank:
Grains of Lead Read from Calibration Curve:
Milliliters of Gasoline Sample:
Temperature of Gasoline Sample:
4. Test Information
Date of Laboratory Teist: / /
Mo Da Yr
Time of Laboratory Test: /
Hr Min
Pb, g/gal:
5. Laboratory Technician
Name:
Signature:
Figure 6. Calculation data record.
22
-------
LABORATORY TEST RECORD
^- Test Station Sample Identification
Test Station Sample Number:
Region:
Date of Station Test: / /
MO DA YR
Time of Station Test: :
HR MIN
2. Laboratory Identification
Laboratory Name:
Laboratory Address:
Laboratory Analysis Number:
3. Test Information
Date of Lab Test: / /
MO DA YR
Time of Lab Test: :
HR MIN
Pb, g/gal
4. Laboratory Technician
Name:
Signature:
Date:
Figure 7. Laboratory test record.
23
-------
SECTION III QUALITY ASSURANCE PROCEDURES
3.0 GENERAL
The control of data quality is a function of two related activities
of the quality assurance program: 1) development of standard operating
procedures including centre1 limits, and 2) assurance of conformance to the
procedures and control limits. Standard operating procedures and control
limits are recommended in the operations manual of this document. It is
emphasized that if the analyst conscientiously adheres to the procedures
and checks of section II, then the precision and accuracy of the lead deter-
minations should be within acceptable limits. Assurance of data quality
basically involves collectLng the information necessary to document and
demonstrate the quality of the measured data. This section of the document
will discuss the activities necessary to document and demonstrate data
quality.
Verification of data quality is important in this instance because the
data generated by this method are to be used to determine if the standard
for lead in gasoline is being exceeded. If results indicate that the stan-
dard is being exceeded, the appropriate enforcement group will be required
to take action. Thus, the professional competence of the analyst, the
operating procedures used, and the measured values that he reports may be
challenged in a court of law.
The quality assurance procedures presented in this section should be
carried out or closely monitored by the individual directly responsible for
the quality of the reported data. In each laboratory, one individual should
be assigned the responsibility for quality assurance. With the exception
of the independent audit, euLl functions could be performed by the analyst
if properly trained.
The purposes of this Election are to:
1. Present information relative to the measurement method (i.e., a
functional analysis) to identify the important operations and
factors,
2. Present techniques for the collection of information to identify
trouble,
24
-------
3. Present an independent performance audit procedure for use in
quantifying data quality on an interlaboratory basis,
4. Present techniques for data-quality assessment.
These four purposes will be discussed in the order stated in the subsections
that follow. The first subsection (3.1) will contain a functional analysis
of the measurement method with the objective of identifying the most impor-
tant factors that affect the quality of the reported data and of estimating
the expected variation and biases in the measurements resulting from equip-
ment and analyst errors.
Subsection 3.2 will contain suggestions for the collection and analy-
sis of information to identify troluble. This will involve the use of
control charts for duplicate measurements of gasoline samples and for mea-
surement of working standard samples with appropriate criteria for decision-
making concerning whether the operation is in control and should be left
alone or if it is out of control and corrective action required.
Subsection 3.3 contains a discussion of an independent performance
audit. Such an audit involves randomly inserting standard reference samples
(i.e., NBS or otherwise certified samples) into the measurement process.
Such an audit, if feasible, could serve as an independent check of the mea-
surement process from sample handling through the final calculations. It
would provide a means of assessing data quality as a function of bias and
precision and serve as an independent verification of data quality for
future users of the data.
Data quality assessment is discussed in subsection 3.4. A method for
estimating the precision and accuracy of the reported data using the results
from the independent performance audit is given. Also, a method of testing
the quality against given standards using sampling by variables is given.
3.1 FUNCTIONAL ANALYSIS OF THE MEASUREMENT METHOD
The determination of lead in gasoline requires a sequence of operations
and measurements that yields as an end result a number that serves to repre-
sent the mass of lead in a unit volume of gasoline. The degree of agreement
between the measured and the true value of a sample can be estimated from
the agreement between measured and known or accepted values of reference
samples. Precision and accuracy of the measurement process are reduced to
25
-------
and/or maintained within acceptable limits by identifying and, where feasi-
ble, by eliminating systematic errors. The importance of a variable on the
precision and/or accuracy of a measurement process is a function of the
variable's mean value and variance, how it is related to the dependent
variable, and its probability of occurance under normal operating conditions.
The objectives of this subsection are to:
1. Evaluate variables and estimate error ranges,
2. Determine through a variance analysis the expected variability
in the lead in gasoline measurements,
3. Estimate through a bias analysis the expected bias, if any, in
lead in gasoline measurements.
A functional analysis of the measurement process is performed to deter-
mine all the major operations and variables that may affect the quality of
the reported measurements. Data quality is characterized by measures of
precision and bias. In subsection 3.1.1, variables believed to be important
to the measurement method are discussed. Estimates of the mean, variance,
and probability distribution are made using data from published reports when
available, and using engineering judgments when documented data are not
available. These data are then used in a variance analysis (subsection
3.1.2) to determine the resulting variability of the measured value, i.e.,
the mass of lead per unit volume of gasoline. The data from subsection
3.1.1 are also used in subsection 3.1.3 to estimate the potential bias of
the measurement process.
3.1.1 Variables Evaluation and Error Range Estimates
A sample, at the time it is collected, will contain a specific but
unknown quantity of lead per unit volume of gasoline. The difference in
the lead-per-volume ratio existing at the time of sample collection and
that measured at some latter time is due to errors in the measurement
process. Following a hypothetical sample of gasoline from collection
through analysis, the following operations could, if not properly carried
out, adversely influence the measurement results.
1. Sample handling. Exposure to the atmosphere at any time between
sample collection and analysis could result in the loss of the
light fractions o:: the gasoline sample due to evaporation. This
26
-------
would effectively increase the lead content per unit volume,
resulting in a higher-than-true measured value.
2. Failure to properly tune the instrument. Failure to use a
definitely lean (oxidizing) flame can reduce the precision of the
measurement method. The flame should be adjusted to give a
sharply defined blue cone with no trace of unburned hydrocarbons
visible above the cone.
3. Preparation of standards. Preparation of standards from weighing
of dried reagent-grade lead chloride through the dilutions re-
quired to obtain the standard solutions could be a source of
errors if careful technique is not followed.
The above operations are rated 40, 40, and 20 for their contribution
to the measurement-method variability. This rating is an estimate and not
derived from actual data.
Sample handling would act as a positive bias and would probably be
randomly distributed on a between-laboratories basis. It would then be
expected to have a greater influence on the. reproducibility of the method
where several laboratories are involved than on the repeatability where
only one laboratory is involved. Using the values of a , the repeatability
standard deviation, and a , the reproducibility standard deviation, as
K
obtained in subsection 3.1.2 and, assuming that sample handling accounts
for 40 percent of the total variance, the error term for sample handling
(SH), will be modeled as a normal variate with a mean of 0.001 g Pb/gal and
standard deviations of 0.0014 g Pb/gal for repeatability and 0.0023 g Pb/gal
for reproducibility. These will be written as N(0.001, 0.0014) and N(0.001,
0.0023) for repeatability and reproducibility respectively.
There are no actual measurement data to support the positive mean
value. However, it is felt that, in general, evaporation and sample con-
tamination combined will result in a positive bias.
Failure to properly tune the instrument (i.e. , to properly adjust the
flame) will result in reducing the precision of the measurement method. The
error due to flame adjustment (FA) would be expected to be positive, espe-
cially with gasoline that has a high aromatic content. It is therefore
modeled by a normal distribution with a mean of 0.001 g Pb/gal and standard
27
-------
deviations of 0.0014 g Pb/gal and 0.0023 g Pb/gal for repeatability and
reproducibility, respectively.
Variability in the Preparation of Standards (PS) is assumed to account
for only 20 percent of the total variability. On a reproducibility bases,
the purity of the reagent-grade lead chloride, the accuracy of weighing out
0.4433 g of the reagent, and subsequent dilutions to obtain the required stan-
dard solutions are sources of error. For repeatability the variability of
the standards should be very small. The error due to preparation of stan-
dards is modeled as N(0, 0.0008 g Pb/gal) and N(0, 0.0016 g Pb/gal) for
repeatability and reproducibility respectively.
Table 2 summarizes the error sources discussed above with their esti-
mated mean values and standard deviations.
Table 2. Estimate means and standard deviations of
important error sources
Repeatability
Reproducibility
Operation
Mean Std. dev. Mean Std. dev.
(g Pb/gal) (g Pb/gal) (g Pb/gal) (g Pb/gal)
Sample Handling (SH) 0.001 0.0014 0.001 0.0023
Instrument tune up (FA)
(Flame Adjustment) 0.001 0.0014 0.001 0.0023
Preparation of
Standards (PS) 0.0 0.0008 0.0 0.0016
3.1.2 Variance Analysis
Many different factors may contribute to the variability of a measure-
ment method, for example:
1. The analyst,
2. Apparatus and reagents used,
3. Equipment calibration,
4. The environment (temperature, humidity, pollutant concentration,
etc.).
The variability will be larger when the measurements to be compared
are performed by different analysts and/or with different equipment, than
when they are carried out by a single analyst using the same equipment. Many
28
-------
different measures of variability are conceivable according to the circum-
stances under which the measurements are performed.
Only two extreme situations will be discussed here. They are:
1. Repeatability, r, is the value below which the absolute difference
between duplicate results, i.e., two measurements made on the same
sample by the same analyst using the same equipment over a short
interval of time, may be expected to fall with a 95 percent
probability.
2. Reproducibility, R, is the value below which the absolute difference
between two measurements made on the same sample by different
analysts in different laboratories using different equipment may
be expected to fall with a 95 percent probability.
The above definitions are based on a statistical model according to
which each measurement is the sum of three components:
CPb = £Pb + b + 6
where
C , = the measured concentration of Pb in gasoline, g Pb/gal
C = the general average, g Pb/gal
b = an error representing the differences between laboratories,
g Pb/gal.
e = a random error occurring in each measurement, g Pb/gal.
In general b can be considered as the sum
b - b + b (2)
r s
where b is a random component and b a systematic component. The term b
L S
is considered to be constant during any series of measurements performed
under repeatability conditions, but to behave as a random variate in a
series of measurements performed under reproducibility conditions. Its
variance will be denoted as
var b = a, (3)
29
-------
the between-laboratory variance including the between-analyst and between-
cquipment variabilities.
The term e represent a random error occurring in each measurement. Its
variance
2
var e = a
r
will be called the repeatability variance.
For the above model the repeatability, r, and the reproducibility, R,
are given by
r - 1.96 i/2 a - 2.77 a (4)
R = 2.77*/a_ + of = 2.77 n . (5)
2
where a will be referred to as the reproducibility variance.
K
Values of a and a can be obtained from the values of repeatability
r K
and reproducibility respectively as given for the reference method in the
Federal Register (see appendix A). The repeatability, r, is given as 0.005
g Pb/gal. Using this value in equation (4) gives a = 0.0018 g Pb/gal.
Reproducibility is given as R = 0.01 g Pb/gal; then from equation (5) a =
K
0.0036 g Pb/gal.
As can be seen the reproducibility standard deviation is larger by
almost a factor of 2 than the repeatability standard deviation. It is felt
this large difference is due primarily to differences in analyst techniques
for performing operations such as sample handling, proper flame adjustment,
and preparation of standards.
For this analysis it is assumed that the total variance is equal to
the sum of the variances of the three operations previously discussed.
The variance (see table 2) of C , then is given by
o2 {Cpb} = a2 {SH} + a2 {FA} + a2 {PS}. (6)
Using the relationship in equation (5) the between-laboratory variance
cr can be determined.
JL
30
-------
222
°L = °R - °r
and a = 0.0031 g Pb/gal.
Li
3.1.3 Bias Analysis
There are no data available for estimating the bias of the measure-
ment process. However, the assumption made concerning instrument tune up
in the previous subsection would result in a positive bias of about 4 per-
cent at C , = 0.05 g Pb/gal. The assumption concerning bias could be eval-
uated by measuring reference samples, if and when they become available.
Assuming that the true or acceptable value, C* , of a gasoline sample
is known then from equation (1)
SPb - CPb - ; <7)
represents an estimate of the bias of the measurement method. An estimate
of the bias can be obtained from audit results as discussed in section 3.3.
3.2 COLLECTION OF INFORMATION TO IDENTIFY TROUBLE
In a quality assurance program, one of the most effective means of pre-
venting trouble is to respond immediately to indications of suspicious data
or equipment malfunctions. Certain visual and operational checks can be
performed by the analyst while the measurements are being made to help in-
sure the generation of data of acceptable quality. These checks are written
as part of the routine operating procedures in section II.
The use of control charts is recommended as a method for monitoring
and documenting the performance level of the measurement process. Recom-
mended control charts are:
1. A control chart for duplicate results (or measurements of SRS's)
to monitor analyst technique and equipment stability.
2. A control chart for the measurement of working standard solutions
to monitor calibration stability.
A sample control chart for each of the above parameters with suggested
limits is given in the following subsections. Control charts are discussed
in textbooks such as references 10 and 11.
31
-------
3.2.1 Control Chart for Duplicates
A sample control chart for the difference in duplicate measurements
of a gasoline sample is given in figure 8. It is recommended that a dupli-
cate measurement be made on every tenth sample (e.g., sample number 1 would
be reanalyzed as sample number 11, and sample number 10 reanalyzed as sample
number 21). The second measurement should be separated by at least 9 sam-
ples from the original measurement.
From control chart theory, the average range, R, for duplicates is
equal to 1.128 x a , or 0.002 in this case. The upper control limit, UCL,
is equal to 3.27 R or 0.0065 as shown in figure 8. (Note that this is the
3-sigma value, whereas reproducibility was defined as the 2-sigma value.)
Take the absolute difference in the duplicates, and plot the point on
the graph of figure 8. Connect each point to the previously plotted point
by a straight line. A point falling outside the UCL indicates that an
actual change has occurred in the measurement process. Such a change could
be due to poor technique, equipment change, and/or sample deterioration.
As long as the plotted points remain within the UCL, the measurement process
is considered in control, and no action is required.
When a point falls outside the UCL, a quick check would be to prepare
and measure a working standard sample of about the same absorbance to check
the calibration curve. If the calibration curve has not changed, the most
likely cause of the excess variability is poor technique. The sample
(another aliquot from the same gasoline sample or SRS) should be measured
again after reviewing the analysis procedures. Plot the difference in the
original and third measurements and, if it is below the UCL, continue the
operation. Document the cause and corrective action taken on the control
chart of figure 8.
3.2.2 Control Chart for Working Standard Samples
Duplicate measurements of a standard sample should agree closer than
duplicate measurements of a gasoline sample, because a portion of variabil-
ity due to analyst technique is eliminated. It is estimated that the cali-
bration curve should be suspect if the measured value of a standard sample
differs more than 0.004 g Pb/gal from its known value.
Figure 9 is a sample control chart for the differences in the measured
32
-------
w
H
U
M
_J
Pi
Q
Z
M
td
OS
w
0.007
0.006
0.005
0.004
0.003
0.002
0.001
t
ACTION LIMIT = 0.0065
WARNING LIMIT = 0.005
.UCL
CHECK NO.
DATE OR
TIME
ANALYST
10
PROBLEM
AND
_CORRECTIVE
ACTION
DUPLICATE NUMBER
Figure 8. Sample control chart for duplicate measurements
of gasoline samples.
33
-------
_ 0.005T-
« &
3~C 0.004
pd oo
Eo « 0.003
CHECK NO.
DATE
ANALYST
PROBLEM AND
CORRECTIVE
ACTION
ACTION LIMIT
WARNING LIMITS
^ — ^
"" M ^^^^^k
f *^*
IB-^ .^r
»*»«^ ^-H CL
y^v^ * ^^••^^
^r ^^•'•^^^^
_ ^^ WARNING LIMITS
— — i— — — — — • — — — — — — — — — — — — — — — • —
~
ACTION LIMIT LCL
—
l\\\\\\\\\\^
I
2
3
4
5
6
7
8
9
10
^
MEASUREMENT NUMBER
Figure 9. Sample control chart for the measurement
of working standard samples.
34
-------
value and the known value of standard samples. The UCL is taken as the
3-sigma value9 and it is assumed that three standard deviations is 0.004 g
Pb/gal. As each standard is measured, the difference should be computed
and plotted on the graph. Each point is connected to the previously plotted
point with a straight line. Corrective action, such as preparing new stan-
dards and/or recalibrating the spectrometer, should be taken anytime:
1. One point falls outside the action limits.
2. Two out of three points fall between the warning and action lines,
and
3. Seven consecutive points fall on the same side of the average or
zero line.
Analysis of gasoline samples should not be attempted until a standard sample
(i.e., a sample prepared from the standard lead solution) can be measured
to within +0.004 g Pb/gal.
3.3 INDEPENDENT PERFORMANCE AUDIT
If implemented properly, an independent audit can be used to evaluate
the total measurement process through the use of standard reference samples
(SRS). An SRS is defined as a gasoline sample whose lead content is accu-
rately known (preferably NBS certified) to the auditors, but unknown to the
analyst being audited. Results from an audit provide an independent assess-
ment of data quality by providing a means of estimating the precision and
bias of the reported results. If SRS's are not feasible, then other NBS-
certified lead samples (e.g., lead in iso-octane), even though it will be
obvious to the analyst that they are special samples, can be used. The
results may not estimate the true variability of the measurement method
because of the recognizible special samples.
3.3.1 Procedure for Performing a Quality Audit
The individual or organization responsible for performing the audit
should obtain a supply of gasoline samples with known lead concentrations.
Samples should be placed in sample containers identical to the containers
used in the field. If possible the SRS should not be distinguishable from
regular field samples. These reference samples should be shipped or deliv-
ered to the analysis laboratory in the same manner that is used for field
samples.
35
-------
3.3.2 Frequency of Audit
The optimum frequency of audit is a function of certain costs and
the desired level of confidence in the data quality assessment. Also,
another consideration would have to be the quality of the data presently
being reported. The repeatability of this method appears small enough to
indicate that within-laborafory precision of the method will not be a big
problem. However, reproducibility or between-laboratory biases could be
important to the accuracy (bias) of the reported data.
Initially an auditing level of once every 20 samples or once for every
measurement cycle if less than 20 samples are to be analyzed is recommended.
If the data for the first audit period (calendar quarter) indicate that
only good quality data are being reported, the auditing level could be re-
duced. However, the lower the audit level, the higher becomes the risk of
declaring good data bad. The risk of declaring bad data good is constant
for all audit levels for the proposed sampling plan.
3.4 DATA QUALITY ASSESSMENT
Two aspects of data quality assessment are considered in this section.
The first considers a means of estimating the precision and accuracy of the
reported data, e.g., reporting the bias, if any, and the standard deviation
associated with the measurements. The second consideration is that of
testing the data quality against given standards using sampling by variables.
For example, lower and upper limits, L and U, may be selected to include a
large percentage of the measurements and outside of which it is desired to
control the percentage of measurements to, say, less than 10 percent. If
the data quality is not consistent with these limits, L and U, then action
is taken to correct the possible deficiency as quickly as possible and to
correct the previous data when possible and/or feasible.
3.4.1 Estimating the Precision/Accuracy of the Reported Data
This section will indicate how the audit data collected in accordance
with the procedure described in Section 3.3.1 will be utilized to estimate
the precision and accuracy of the measure of interest. Similar techniques
can also be used by a specific firm or laboratory to assess their or its own
measurements. The audit data collected as a result of following the
36
-------
procedures in the previous section are the measured and known values of P
i
and the difference:
. - CPb.
C = measured value of lead in the reference sample, g Pb/gal,
j
C' = known value of lead in the reference sample, g Pb/gal, and
- the audit number, j - 1, . . . n.
Let the mean and standard deviation of the differences d., j = 1, .
audits be denoted by d and s,, respectively. Thus,
n
and
Sd =
n
/(n - 1)
1/2
3.4.2 Statistical Tests
The mean d is an estimate of the relative bias in the measurements
(i.e., relative to the known or accepted value). Assuming the audit value
to be unbiased, the existence of a bias in the laboratory data can be
checked by the appropriate t-test, i.e.,
d - 0
1
See reference 12 for a discussion of the t-test. If t is significantly
large in absolute values, i.e., greater than the tabulated value of t with
n - 1 degrees of freedom, which is exceeded by chance only 5 percent of the
time, then the bias is considered to be real and some check should be made
for a possible cause of the bias. If t is not significantly large, then
the bias should be considered zero or negligible. However, its calculated
value will be reported with the laboratory data for that audit period.
37
-------
The standard deviation, a,, is a function of both the standard devia-
tion of the laboratory measurements and of precision with which the refer-
ence-sample value ia known. Assuming the reference-sample values are known
with much greater precision than the laboratory measurements, then the cal-
culated a, is an estimate of the standard deviation of the laboratory mea-
surement. Table 3 contains an example calculation of d and s,, starting
d
with the differences for a sample size of n = 12.
The calculated standard deviation can then be utilized to check the
reasonableness of the assumption made in subsection 3.1.2 concerning
a {C } « a - 0.0036 g Pb/gal, under reproducibility conditions. The cal-
culated standard deviation, s , may be directly checked against the assumed
value, o , by using the statistical test procedure
K
2 **
£- . d
f 9 '
a«W
2
where x /f is the value of a random variable having the chi-square distri-
2
bution with f » n - 1 degrees of freedom. If x /f is larger than the tabu-
lated value exceeded only 5 percent of the time, then it would be concluded
that the test procedure is yielding results with more variability than is
acceptable, due to some assignable cause of large variation.
The measured values should be reported along with the estimate bias,
d, standard deviation, s,, the number of audits, n, and the total number of
determination periods (number of days analyses were performed) N, sample
(n _< N). Estimates, i.e., s, and d, which are significantly different from
the assumed population parameters should be identified on the data sheet.
For example, based on the data of table 3, if the analyst reported a value
of C = 0.05 g Pb/gal for one of the N field tests not audited, then that
measurement would be reported as
1. Measured value, C = 0.05 g Pb/gal
2. Calculated bias, d = T = -0.00075 g Pb/gal
3. Calculated standard deviation, a {C_, } = s, = 0.0025 g Pb/gal
R Pb d
4. Auditing level, n = 12, N = 65 (N = 5 days per week x 13 weeks
per quarter).
From the above data, users of the data can calculate confidence limits
appropriate to what the data are to be used for.
38
-------
Table 3. Computation of mean difference, d, and
standard deviation of differences, s
General formulas
d = CPbj " CPb.
dl dl
d2
d3 d3
d4 d4
4 4
,2
a d_
d6
d7 d?
dft d«
o o
d9 d9
dio d?o
dll dll
d!2 d!2
Ed. Zd2
d = Zd /n
2 Sd2 - (Zd )2/n
s , - n
d n - 1
s , = /T
d s ,
Specific
Data
-0.005
-0.002
0.000
-0.003
-0.006
0.001
0.002
0.003
0.004
0.001
-0.003
0.000
-0.008
d = -0.
s, = 0.
d
s, = 0.
d
example
(g/gal)
0.000025
0.000004
0.000000
0.000009
0.000036
0.000001
0.000004
0.000009
0.000016
0.000001
0.000009
0.000000
0.000114
0007
0000065
0025 g/gal
39
-------
2
The t-test and x -test described above are used to check on the biases
and standard deviations separately. In order to check on the overall data
quality as measured by the percent of measurement deviations outside pre-
scribed limits, it is necessary to use the approach described below.
3.4.3 jSampling by Variables
Because the lot size ''i.e., the number of determination periods
during a particular period, normally a calendar quarter) is small, N < 150
(estimated), and consequently, the sample size is small on the order of
n = 7 to 13, it is important to consider a sampling by variables approach
to assess the data quality with respect to prescribed limits. That is, it
is desired to make as much use of the data as possible. In the variables
approach, the means and standard deviations of the sample of n audits are
used in making a decision concerning the data quality.
Some background concerning the assumptions and the methodology is re-
peated below for convenience. However, one is referred to one of a number
of publications having information on sampling by variables; e.g., see
references 12-16. The discussion below will be given in regard to the
specific problem herein, whLch has some unique features as compared with
the usual variable sampling plans.
The difference between the analyst-measured and the known value of C
Pb
is designated as d., and the mean difference over n audits by d, that is,
11
* E
Theoretically, C , and C' should be measures of the same lead con-
centration, and their difference-'should have a mean of zero on the average.
In addition, their differences should have a standard deviation approximately
equal to that associated with measurements of C , separately.
Assuming three standard deviation limits (using the assumed a{C , } «
0.0036 g Pb/gal as derived in the variance analysis of subsection 3.1.2),
the values -3(0.0036) = -0.011 g Pb/gal and 3(0.0036) = 0.011 g Pb/gal define
lower and upper limits, L and U, respectively, outside of which it is
desired to control the proportion of differences, d . Following the method
40
-------
given in reference 13, a procedure, for applying the variables sampling plan
is described below. Figures J.O and 11 illustrate examples of .satisfactory
and unsatisfactory data quality with respect to the prescribed limits L and U,
The variables sampling plan requires the sample mean difference, d; the
standard deviation of chese differences,, s,; and a constant;, k, which is
u
determined by the value of p, the proportion of the differences outside the
limits of L and U. For example, if it is desired to control at 0.10 the
probability of not detecting lots with data quality p equal to 0.10 (or 10
percent of the individual differences outside L and U) and if the sample
size n = 12, then the value of k can be obtained from Table II of reference
13. The values of d and s are computed in the usual manner; see table 3
for formulas and a specific example. Given the above information, the test
procedure is applied and subsequent action Is taken in accordance with the
following criteria:
1. If both of the following conditions are satisfied:
d-ksd^L=~ 0.011 g Pb/gal
d 4- k sd £ U = 0.011 g Pb/gal
the individual differences are considered to be consistent with
the prescribed data quality limits and no corrective action is
required.
2. If one or both of these inequalities is violated, possible defi-
ciencies exist in the measurement process as carried out for that
particular lot (group) of determination periods. These deficien-
cies should be identified and corrected as soon as possible to
prevent future measurements of unacceptable quality. Data correc-
tions should be made when possible, i.e., if a quantitative basis
is determined for correction.
Table 4 contains a few selected values of n, p, and k for convenient
reference.
Using the values of d and s in table 3, k = 2.045 for a sample size
of n = 12, and p = 0.10 (table 4), the test criteria can be checked; i.e.,
d - k sd = -0.00075 - (2.045) (0.0025) = -0.0059 > L = -0.011 g Pb/gal
d + k sd = -0.00075 + (2.045) (0.0025) = 0.0044 < U = 0.011 g Pb/gal
Therefore, both inequalities are satisfied; hence, the data are consistent
41
-------
Figure 10. Example illustrating p < 0.10 and satisfactory data
quality.
p (percent of measured
differences outside
limits L and U) > 0.10
Figure 11. Example illustrating p < 0.10 and unsatisfactory data quality.
42
-------
with the limits. The laboratory responsible for generating these data
should be notified that its quality control activities are satisfactory.
Even if the limits had been exceeded, the calendar quarter of data or a por-
tion of that quarter of data should be invalidated only if one or more of
the reported measurements approached, within measurement error, or exceeded,
the standard.
The above plan provides a 90-percent probability of detecting lots
with 10 percent or more defects (i.e., deviations falling outside the desig-
nated limits L and U).
Table 4. Sample plan constants, k for P{not detecting a lot
with proportion p outside limits L and U; < 0.1
Sample Size n k_(g --_0._2_)
3 3.039 4,258
5 1.976 2.742
7 1.721 2.334
10 1.595 2.112
12 1.550 2.045
13 1.533 2.02
14 1.519 1.999
15 1.506 1.981
43
-------
SECTION IV
1. "Standard Method of Tesit for Lead in Gasoline by Atomic Absorption
Spectrotnetry," ASTM Method No. 3237, Part 17, 1180, 1973.
2. M. Kashiki, S. Yamozoe, and S. Oshima, Anal. Chem. Acta 53, 95, 1971.
3. C. Veillon, Handbook of Commercial Scientific Instruments Atomic
Absorption, Vol. 1, Marcel Dekker, Inc., New York, 1972.
4. H. H. Fawcett and W. S. Wood, Safety and Accident Prevention in Chemi-
cal Operations, Interscience Publishers, John Wiley and Sons, New York,
1965.
5. Natl. Bus. Std. Circ. 4-34, 1941; 602, 1959.
6. Standard Methods for the Examination of Water and Wastewater, American
Public Health Association, Washington, D.C., 1971.
7. D. A. Skoog and D. M. West, Fundamentals of Analytical Chemistry, Holt,
Rinehart and Winston, New York, 1963.
8. Walter Slavin, Atomic Absorption Spectroscopy, Interscience Publishers,
John Wiley and Sons, New York, 1968.
9. B. V. L'vov, Atomic Absorption Spectrochemical Analysis, Adam Hilger,
London, 1970.
10. E. L. Grant and R. S. Leavenworth, Statistical Quality Control, 4th
ed., McGraw-Hill, St. Louis, 1972.
11. D. A. Simons, Practical Quality Control, Addison-Wesley Publishing
Company, Reading, Mass., 1970, pp. 131-150.
12. A. Hald, Statistical Theory with Engineering Applications, John Wiley
and Sons, New York, 1952.
13. D. B. Owen, "Variables Sampling Plans Based on the Normal Distribution,"
Technometrics 9, No. 3, August 1967.
14. D. B. Owen, "Summary of Recent Work on Variables Acceptance Sampling
with Emphasis on Non-normality," Technometrics 11, 1969, pp. 631-37.
15. Kinji Takogi, "On Designing Unknown Sigma Sampling Plans Based on a
Wide Class of Non-Normal Distributions," Technometrics 14, 1972,
pp. 669-78.
16. C. Eisenhart, M. Hastay, and W. A. Wallis, eds., Techniques of Statis-
tical Analysis, Statistical Research Group, Columbia University,
McGraw-Hill, New York, 1947.
44
-------
TEST FDR TOE DETERMINATION
OF LEAD IN GASCUNE
1 Kcur i 1 This m.Thod ,w>-s ilie ue'ir-
rr.ln.'s.ti.-'n of the t •! u Joad cc ;^,u cl rric,i>line
wJt'lUl tho co'l.T Hf'V.fnu .,'l.;e of 0010 to
fl 10 sj of lead'US (,'iu The method couipen-
*«tog for variations in gasoline composition
*-nrl is in.ieponci :it or lead :<'Kyi type
2 Rii:xlf alisoipuon
Came spcctioii,<-i \\ .11 21'..ill A •..' iii/ standards
prepared, from rca^oisi trradf lead chloride
By the use of this irc'itmcm, .\'.( alkyl !c:«)
compo.mds ^ivc identical re^nonse
3 ApflarflMi.i 3 1 Atorr.Io Absorption Spec-
tomtt<-r. capabit- of scale expansion mid neb-
ulizer adjustment and equipped \vitn a slot
burner and pnnux chamber lor use with an
»lr-ac"(y!cnc Ila.iv
32 Vtil'imvti < n.'iks "iO'ivl lOO-ml, 250
unl. and one litre size .
33 Plpels 2--nl, 5-rnl, ]0-;i.l. I") nil nncl
6C-rnl sizes
34 Microplpe' i:)0-/.l, Fppcndoi,' tviv or
•fHiivalenr
4. Reagent ^ 4 1 Puntc ol Ke.igen:s---Iiea-
genl grade chemical"; ihall he u.^ed In all
tests. Unli";s others i.se indlcnled. it Is in-
tended that all reipents siiall conform to the
specifications of the Cummltteo on Analytical
Reagents of the Amencau Cnemical Society,
where such spci'lficni Ions are available Other
grades rnnv be used provided it is tirst as-
certained that the reagent is of sufficiently
high purity to permit Its use without less-
ening the accuracy of the determination
4 2 Purity of Water— Unless otherwise In-
dicated, rc-rcrences to aater shall be under-
etxxxi to mean distilled water or water of
equal purity.
43 Aliquat 33C (tricapryl rnribvl ammo-
nium chloride).
44 Aliquat 33G/MIBK Solution (10 per-
cent v/v)—Dissolve and dilute 100 ml i860
g) of Aliquat 3JO with MIBK to one liter.
46 Aliquat 33G/MIBK Solution (1 percent
T/v)—Dissolve and dilute 10 ml (8.8 g) of
Aliquat 330 with MIBK to one liter
4.6 Iodine Solution—Dissolve and dilute
8.0 g iodine rryst.ils with Toluene to 100 ml.
4.7 I.*adChloMJe
48 I-ead-Sterile OaviHne—Ga^oime con-
fining less than 0 005 tr Pb-pa).
49 Ijead, S^nnti.ird Solution (50 g Pb/
gal)—Dissolve 11441;* g of lead chloride
(PbCl ) previously dried at love" .'or 3 h in
alvjtit 200 ml ol ID perron', Aliquat 33U MIBK
solution In a 250-ml volumetric ila.sk. Dilute
to the murk vviih the 10 pel cent Alli(iiat
«olut'on, inu.. aid store in a bn\>. 11 bottle
having a polyeluviene-llned cap This solu-
tion contains 1,321 ug Pb 'ml, which Is equiv-
alent to 5 0 R Pb'gal.
4.10 Ix:ad, Standard Solution (10 g Pb/
ual)—Bv means of a pipet, accurately trans-
fer bu u iiu oi u;t Ci J K i^i ,'tfl s^l^Mon to
a 250-ml volumetric flr-sk dilute to volume
with 1 percent. Aliqual/MTRK soluuoii Store
In a brown bottle liavmg a polyethylene-
lined cap.
4 1! i-j-nd, St»!j!iard Solutions (002. 003.
»jiid 010 g i'B/gai) — Transfer accurauny by
nieans of plpets 20, 5 G, and 100 ml of tli«
1 0-g Pb/sa) nol-illon to iOO-ml volumetric
flasks: Bite 50 mi of ! pcvoejit Aliquat 336
Dilution IG »n.rr, fl.isk. dilute to the mark
with MIBK Mix well ar,d si ore In bottles
Jjavtng polypthvlt i)p-)Uioil c;if..s
4)2 Mc'thvl u.obutyl Ketone (MIBK).
(4-methy!-/-pent?none).
5. Calfhrantii. 6.1 Preparation of 'Working
Stund;xrti:i—l'fej>are three wtfrkiny Ktaadards
and a blank using the 002. 006, n.id 0 10-j;
I'b.'^al Btamia-rd U.ad solutions described in
* il.
51 1 To each <>f four Sifcnii volunietilc
fla.sK6 containing 30 ml of MIBK, add 5 0 ml
of low lead standard solution and 50 ml of
load-frre gasoline In the case of the blank,
add only 5 0 ml of lend-free gasoluie _
5 1 2 Add Im-nedifttely 01 tnl of 'inline/
toluene soHitioi, i-y moans ol the 100-;ii Kp-
pendorf plpet M;x well '
5 1 '1 Arid S n-,1 of 1 percent Alt )U;U 336
solution .tiid rn'.A
5 1 4 D'lute to '.oliiino with MIBK and mix
vi ell
$ ?. v-'i eparat son t','. riu-.trument—Optimise
tJie .i'u)ivilc ao.sorplloii equipmeut for lead at,
281:< A Using the reagent bl.ink. Mljust. the
gas mixture and the sample susptraUoiv rate
to oblair, an oxldixin^ flame.
521 Aspirate the 01-g Pb't;al wcnklni;
standard and adjiut the burner position to
give maximum response Some instruments
require the u.se of scale expansion to produce
a readliiK of 0 160 to 0 170 for this standard
522 Aspirate the reagent blank to zero
the instrument and check the absorbnnces of
the turpt workli,;; standards, for linearity
6 Procedure 6 1 To a 50 ml volumetric
n.isk containing 30 ml MIBK, add 5 0 ml of
ga-solme sample and mix •
6 1 1 Add 0 10 ml (100 ul) of Iodine/
toluene soluUoii and allow the mixture to re-
act about 1 iiiin l
612 Add 5.0 mi of 1 percent Aliquot 3367
MIBK solution and mix
6.1.3 Dilute to volume with MIBK and
mix
6 2 Aspirate the- samples and working
standards and record the absorbtoice values
with frequent checks of the zero.
7 Calcvlauont 7.1 Plot the abaci-bailee
v&lut%£ versus concentration reprftss'iit^ni bv
the working standards and read the concen-
trations of the samples from the graph.
0. Precision 8! The following criteria
should be used for judging the acceptability
of results (95 percent confidence) :
8 1 1 Repeatability—Duplicate results by
the same operator should be considered sus-
pect If they differ by more than 0006 g/gal.
81.2 Keproductlblllt?—The results sub-
mitted by ea<:lj of two laboratories should
not b» considered suspect unless the two re-
mit! differ by more than. 0 01 g/gal.
[FK ix/c.w u44'i ema i-u f*,ti -liinmi
-------
APPBDIXB
GIDSSARY OF SWBOLS
SVMBQL
A
o
A
cm
gal
g
I
1°
mSL
Pb
£
/
nm
T
Vi
N
n
r
a{X}
R
R
DEFINITION
Absorbance
Angstrom
Centimeter
Gallon
Gram
Transmitted Light
Incident Light
Microliter
Milliter
Lead
Liter
Per
Nanometer
Transmittance
Micron
Lot size—i.e., the number of determination
periods to be treated as a group
Sample size for the quality audit (section 3.3)
Repeatability of the measurement method at the
95-percent confidence level
Assumed standard deviation of the parameter X
(population standard deviation)
Assumed mean value of the parameter X (popu-
lation mean)
Computed average of a finite sample of measure-
ments (sample mean)
Reproducibility of the measurement method
(section 3.0)
Range; i.e., the difference in duplicate
Average range or difference for duplicate
results (fig. 6)
46
-------
APPENDIX B GLOSSARY OF SWBOLSHWINUED
SYMBOL DEFINITION
d The difference in the audit value and the
measured value arrived at by the analyst
for the j^h_ audit
d Mean difference between known and measured
values of reference samples for n audits
s. Computed standard deviation of difference
d
between known and measured values
P Percent of measurements outside specified
limits L and U (section 3.4)
k Constant used in sampling variables (section
3.4)
p{Y} Probability of event Y occurring
a Repeatability standard deviation
a Reproducibility standard deviation
K.
t 1 Statistic used to determine if the sample
n-1 _
bias, d, is significantly different from
zero (t-test)
2
•^j- Statistic used to determine if the sample
2
variance, s , is significantly different
2
from the assumed variance, a , of the parent
distribution (chi-square test)
L Lower quality limit used in sampling by vari-
ables
U Upper quality limit used in sampling by
variables
CL Center line of a quality control chart
LCL Lower control limit of a quality control chart
UCL Upper control limit of a quality control chart
Measured lead in a gasoline sample in g/gal
C' The known concentration of lead in reference
sample, g/gal
C , The average of several measurements on the
same or identical samples (eq. 1)
47
-------
APPBDIXC
GLOSSARY OF TERMS
The following glossary lists and defines the terms as used in this document.
Absorbance
Accuracy
Bias
Chain of Custody
Label
Determination
Method
Determination
Process
Lot
Population
Precision
Quality Audit
Quality Control
Check
Quality Control
Sample
Repeatability
The logarithm to the base 10 of the reciprocal of
transmittance.
A measure of the error of a process expressed as a com-
parison between the measured value and the true value.
The systematic or nonrandom component of system error.
The seal placed on the container which contains the
gasoline sample from the test station.
A set of procedures for making a determination.
The process of making a determination including method,
personnel, equipment, and environmental conditions.
A specified number of objects to be treated as a group.
A very large number of like objects (i.e., measurements,
checks, etc.) from which the true mean and standard
deviation can be deduced with a high degree of accuracy.
The degree of variation among measurements on a homogene-
ous material under controlled conditions, and usually
expressed as a standard deviation or as a coefficient
of variation.
A management tool for independently assessing data quality.
Checks made by the operator on certain items of equipment
and procedures to assure data of good quality.
Sample whose Pb concentration is known to the analyst.
The value below which the absolute difference between
duplicate results, i.e., two measurements made on the
same sample by the same analyst in the same laboratory
using the same equipment over a short interval of time,
may be expected to fall with a 95 percent probability.
48
-------
APPENDIX C
GLOSSARY OF TERMS-CONTINUED
Reproducibility
Sample
Standard Reference
Sample (SRS)
Transmittance
Working Standard
Sample
The value below which the absolute difference between
two measurements made on the same sample by different
analysts in different laboratories using different
equipment may be expected to fall with 95 percent
probability.
Objects drawn usually at random from the lot for checking.
Certified sample submitted to laboratory for quality con-
trol check as part of an independent performance audit.
Transmittance is the ratio of the transmitted light to
the incident light.
Sample prepared by the analyst from the standard lead
solution to check the lead calibration curve.
-------
APPENDIX D
TO CONVERT FROM
Nanometer (nm)
Gal
Transmittance (T)
CONVERSION FACTORS
JQ
Micron (p)
o
Angstom (A)
Centimeter (cm)
Liter (£)
Absorbance (A)
MULTIPLY BY
0.001
0.1
io-7
3.785
10
-A
50
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1 REPORT NO,
EFA-650A-7U-005-m
2.
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
5. REPORT DATE
Guidelines for Development of a Quality Assurance
Program: Volume XIII - Test for Lead in Gasoline by
Atmospheric Absorption Spectrometry.
November
6. PERFORMING ORGANIZATION CODE
7. AUTHORiSS
Denny E. Wagoner, Franklin Smith, D. Gilbert
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Research Triangle Institute
P.O. Box 1219li
Research Triangle Park, Worth Carolina 27709
10. PROGRAM ELEMENT NO.
11. CONTR ACT/GRANT
1HA3P7
68-02-1231*
12. SPONSORING AGENCY NAME AND ADDRESS
Office of Research and Development
U.S. Environmental Protection Agency
Washington, D. C. 20U60
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This document presents guidelines for a quality control program for the
determination of the total lead content of gasoline within the concentration range
of 0.010 to 0.10 g of lead/U.S. gals. These guidelines include:
1. Good operating practices
2. Directions on how to assess performance and quality data
3. Directions on how to identify trouble and improve data quality
H. Directions to permit design of auditing activities
The document is not a research report. It is designed for utilization by
laboratory personnel.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Quality Assurance
Quality Control
Air Pollution
Gasoline
Lead
b.IDENTIFIERS/OPEN ENDED TERMS
c. COS AT I Field/Group
13H
13B
21D
7B
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
unclassified
21. NO. OF PAGES
57
20 SECURITY CLASS (Thispage)
unclassified
22. PRICE
EPA Form 2220-1 (9-73)
51
-------
-------
-------
i
n
>
9 r> s
0) * O
lill
-to S3
2. » n> 5
jr -• -• m
-£20
X ~ 2 H
^ 0) Cfl —
»J
-J
OO
o
-Pa.
I
•^J
-P»
O
O
cn
z
o
NO 33
=5==!
CD
CD
CO (D
Q. —
Q. -.
Q.
Di
3 ;j
o cp
— o
O CD
Si <
CD CD
io 3
CD g
- CD
3 ^
O. (U
O-o
IS
m -< O
"O > m
-• -n
2 m
O m
-4 in
o
2
n
------- |