&EFA
United States
Environmental Protection
Agency
Atmospheric Research and Exposure
Assessment Laboratory
Research Triangle Park, NC 27711
December 1993
Research and Development
EPA 600/R-93/235
Preparation of Lead-Containing
Paint and Dust Method Evaluation
Materials and Verification of the
Preparation Protocol by
Round-Robin Analysis
r
-------
December 1993
PREPARATION OF LEAD-CONTAINING PAINT AND DUST
METHOD EVALUATION MATERIALS AND VERIFICATION
OF THE PREPARATION PROTOCOL
BY ROUND-ROBIN ANALYSIS
Prepared by
E. E. Williams
D, A. Binstock
W. F. Gutknecht
Center for Environmental Measurements and Quality Assurance
Research Triangle Institute
Research Triangle Park, North Carolina
EPA Contract No. 68-D1-0009
RTI Project No. 4960-141
Mrs. Sharon Harper, Work Assignment Manager
Atmospheric Research and Exposure Assessment Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC
-------
DISCLAIMER
The information in this document has been funded wholly or in part by the
United States Environmental Protection Agency (USEPA) under EPA Contract
No.68-Dl-0009 to the Research Triangle Institute. It has been subjected to the
Agency's peer and administrative review, and it has been approved for publication
as an EPA document. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
11
-------
ACKNOWLEDGEMENTS
This document was prepared under the direction of Drs. Joseph J. Breen
and Benjamin S. Lim of the Office of Pollution Prevention and Toxics (OPPT), U.S.
Environmental Protection Agency (USEPA), Washington, DC, and Mr. Michael E.
Beard and Ms. Sharon L. Harper of the Atmospheric Research and Exposure
Assessment Laboratory (AREAL), U.S. Environmental Protection Agency,
Research Triangle Park, NC.
The authors acknowledge the efforts of statisticians Dr. Larry Myers of the
Research Triangle Institute, and Mr. Jack Suggs of AREAL/USEPA, Research
Triangle Park, NC.
Special acknowledgement is given to Dr. Joseph Walling, AREAL/USEPA,
Research Triangle Park, NC; Dr. Benjamin Lim, OPPT/USEPA, Washington, DC;
and Dr. James DeVoe, Inorganic Analytical Research Division, National Institute
of Standards and Technology (NIST), Gaithersburg, MD for their careful review.
in
-------
EXECUTIVE SUMMARY
The determination of lead in paint, dust, soil and other matrices is receiving
increased attention because of the adverse health effects associated with exposure
to low levels of this environmental contaminant. Because exposure to lead
hazards may be minimized or prevented by appropriate detection, abatement or
containment, the accurate and precise identification of lead levels in paint, dust
and soil is an important environmental concern. The concentration of lead in
paint, dust and soil samples may be determined either in the laboratory or in the
field. In order for concentration data to be reliable, it is important to also
calibrate instruments and benchmark analytical performance with the use of
reference materials. These materials are homogeneous, well-characterized, and
have a known concentration of the analyte(s) of interest. However, the availability
of reference materials for the routine analysis of environmental lead samples is
limited, and there are no standard protocols for the production of these materials.
This study was carried out to prepare a series of lead-containing paint and
dust reference materials according to criteria established at a Lead Reference
Materials Workshop sponsored by the U.S. Environmental Protection Agency. The
criteria for the production of the materials, called Method Evaluation Materials
(MEMs) included the following;
• lead concentration,
• material homogeneity, and
• characteristics of the matrix.
After the materials were prepared, the protocol for the preparation was validated
by analysis of the materials for the following:
• measured lead concentrations within 20% of the target
concentrations, and
IV
-------
• sample to sample variations (homogeneity) of the materials
statistically non-significant relative to overall standard deviations.
The analyses were carried out by:
• the Research Triangle Institute, and
• 33 external laboratories.
Because a sufficient number of laboratories analyzed the MEMs using
different selected extraction/analytical methods, statistical analysis of the data
also allowed a comparison of laboratory performance using these proven methods.
Four MEMs were prepared at the following targeted lead concentrations:
• 100 /ig/g in dust,
• 1500 fjLg/g in paint,
• 4000 ng/g in dust, and
40000 fjig/g in paint,
from "real-world" lead-containing paint and dust, collected from households in
North Carolina and California, abatement sites in Pennsylvania and a vacant
hospital in Ohio.
The paint materials were collected as chips scraped from walls, woodwork
and other surfaces. Aliquots were taken from each bag of chips, ground by hand
using a mortar and pestle, and then analyzed to obtain estimates of the lead
levels. Analysis was performed using microwave/acid extraction and measuring
the lead levels by inductively coupled plasma emission spectrometry. Specific
paint materials were chosen on the basis of these results to meet target
concentrations. The paint materials chosen were then mechanically ground to a
fine powder (•& 120 microns) and each batch prepared mixed thoroughly.
The dust was collected in home vacuum cleaners and also high efficiency
particulate collection vacuum cleaners. The dust was sent to a commercial firm
-------
for sterilization and then sieved to a particle size «;250 microns. The sieved dust
samples were each thoroughly mixed and were then subjected to preliminary
analysis as described for paint, and batches selected relative to the target
concentrations.
Prior to a round robin analysis of the selected, prepared materials
verification analyses were performed.
The concentrations of the MEMs, determined by RTI to be acceptable
relative to the target concentrations, were the following:
• 84.2 ± 11.9 fjig/g - low lead-containing dust,
* 1410 ± 44.5 fig/g - low lead-containing paint,
• 4670 ± 330 Atg/g - high lead-containing dust,
• 37900 ± 500 ju,g/g - high lead-containing paint, and
These samples were submitted in duplicate to laboratories for round-robin
analysis.
The sample set submitted to round-robin analysis also included Standard
Reference Materials (SRMs) of paint and "dust" (a soil SRM was used as a
surrogate for dust) prepared and certified by the National Institute of Standards
and Technology (NIST). The following Standard Reference Materials were
included as single blind samples:
1162 ± 31 ftg/g - NIST SRM 2711, Montana Soil, used as a
surrogate dust sample
118700 ± 400 ng/g - NIST SRM 1579, Powdered Lead-based
Paint.
The complete sample set included 2 bottles of each paint MEM, 2 bottles of
each dust MEM, one bottle of paint SRM, and one bottle of "dust" SRM for a total
of 10 bottles of samples. Each laboratory was asked to analyze two aliquots of
VI
-------
each sample for a total of 20 analyses. Laboratories were recruited for
participation in the round robin on the basis of their experience and willingness to
carry out the analyses fay methods commonly used to analyze environmental lead
samples:
• hotplate (HP) or microwave (MW) extraction followed by analysis by
atomic absorption spectrometry (AAS) or inductively coupled plasma
emission spectrometry (ICP) , and/or
• energy dispersive laboratory X-ray fluorescence (Lab XRF).
A total of 33 laboratories performed 42 different sets of analyses, as follows:
Methodoloev
MW/AAS
HP/AAS
MW/ICP
HP/ICP
Laboratory XRF
Number of Performances
7
9
9
10
7
The number of laboratories analyzing by each method (a minimum of seven
(7) performances were required) was sufficient for a statistical comparison of
methods. Results of the statistical analysis provided data for determination of the
method mean, consensus value, repeatability and reproducibility of methods for
each test sample. The method means and consensus values indicated that the
protocol produced samples having acceptable concentrations relative to the target
concentrations. Precision data indicated that the average sampling coefficient of
variance (cv) was 1.37%; the 95% upper confidence limit of the cv was 2.5%; and
therefore, 95% of all test samples were found to have a concentration within 5% of
the consensus value (95% to 105 % of the consensus value). Therefore, the
homogeneity of the materials was considered to be acceptable.
vii
-------
A comparison of data by method showed that the MW/AAS method gave
results with the highest concentrations for all six test samples. Laboratory XRF
gave the lowest results for 5 out of 6 test samples. A pairwise comparison of
method means indicated that these two methods also showed the most statistically
significant differences. When the data for matrices was pooled, the repeatability
(within-lab variation) of the laboratory XRF method was shown to be best (4.8%)
for all methods tested (range of methods: 4.8% - 12.9%); but the reproducibility
(between-lab variation) of this method (19.4%) was poor (range of methods: 11.7% -
21.0%). The reproducibility of the MW/ICP method was the best (11.7%) across all
concentrations of the test samples.
The poor reproducibility of the Lab XRF method was attributed to:
• failure to request that laboratories follow the same protocol for the
analyses, and/or
• the provision of an inadequate number of calibration standards for
the instrumental analysis. (This is suggested by the quadratic
appearance of log recovery plots for the Lab XRF method.)
Results also indicated that recoveries for analyses by AAS showed a positive
bias relative to ICP results. This bias was believed to result from the lack of
background correction by a number of laboratories analyzing by AAS. It is also
possible that the concentrations were suppressed in the ICP measurements, but
laboratories analyzing by ICP were warned about signal suppression arising from
matrix effects, and were instructed to dilute solutions for analysis into a 1 - 10
/itg/mL range to minimize these effects. It is suggested that further studies be
performed to investigate the bias observed in results reported by the analytical
methods, and the poor reproducibility shown by Laboratory XRF.
vm
-------
TABLE OF CONTENTS
1.0 INTRODUCTION 1-1
1.1 OVERVIEW 1-1
1.2 REPORT 1-4
2.0 DESIGN OF THE METHOD EVALUATION MATERIALS 2-1
2.1 CRITERIA ESTABLISHED BY THE LEAD REFERENCE
MATERIALS WORKSHOP 2-1
2.1.1 Paint 2-1
2.1.2 Dust 2-1
2.2 CONCENTRATIONS PROPOSED FOR THE METHOD
EVALUATION MATERIALS 2-2
3.0 PREPARATION OF THE METHOD EVALUATION MATERIALS 3-1
3.1 PAINT 3-1
3.1.1 Collection of Materials 3-1
3.1.2 Selection of Bulk Materials 3-2
3.1.3 Grinding 3-2
3.1.4 Blending 3-2
3.1.5 Determining the Effect of Aliquot Weight
on Analytical Results 3-3
3:1.6 Production of Target 0.15% Material 3-5
3.1.7 Preliminary Verification of Concentration and Homogeneity 3-5
3.2 DUST 3-7
3.2.1 Collection of Materials 3-7
IX
-------
3.2.2 Sterilization 3-7
3.2.3 Removal of Debris 3-7
3.2.4 Selection of Bulk Materials 3-8
3.2.5 Blending 3-8
3.2.6 Determining the Effect of Aliquot Weight
on Analytical Results . 3-8
3.2.7 Preliminary Verification of Concentration
and Homogeneity 3-9
3.3 BOTTLING THE TEST SAMPLES 3-9
3.4 FINAL VERIFICATION OF CONCENTRATIONS OF THE METHOD
EVALUATION MATERIALS 3-10
4.0 ROUND-ROBIN ANALYSIS OF THE METHOD EVALUATION
MATERIALS 4-1
4.1 ROUND-ROBIN DESIGN 4-1
4.2 RECRUITMENT OF LABORATORIES 4-2
4.3 ROUND-ROBIN ANALYSIS 4-3
4.3.1 Standard Operating Procedures 4-3
4.3.2 Letter of Instructions 4-4
4.3.3 Data Reporting Form 4-4
4.3.4 Instrument Parameter Forms 4-5
4.3.5 Responses from Participating Laboratories 4-5
4.3.6 Notification of Results 4-6
5.0 STATISTICAL ANALYSIS OF RESULTS 5-1
5.1 CENSORED, MISSING DATA 5-1
x
-------
5.2 OUTLYING DATA 5-2
5.3 METHOD MEANS 5-3
5.4 CONSENSUS VALUES 5-3
5.5 REPEATABILITY AND REPRODUCIBILITY 5-7
5.6 SAMPLE HOMOGENEITY 5-12
5.7 PAIRWISE COMPARISON OF METHOD MEANS 5-17
5.8 COMPARISON OF MEASUREMENTS BY ATOMIC ABSORPTION
SPECTROMETRY AND INDUCTIVELY COUPLED PLASMA
EMISSION SPECTROMETRY 5-17
6.0 SUMMARY AND CONCLUSIONS 6-1
7.0 RECOMMENDATIONS FOR FURTHER STUDY 7-1
8.0 REFERENCES 8-1
XI
-------
LIST OF TABLES
Table 1.
Table 2.
Table 3.
Table 4.
Table 5.
Table 6.
Table 7.
Table 8.
Table 9.
Table 10.
Table 11.
Concentrations of Lead Measured in Paint and Dust Method
Evaluation Materials Relative to Changes in Aliquot Weight for
Extraction 3-4
The Concentration and Homogeneity (RSD) of Paint and Dust Method
Evaluation Materials Determined at RTI by Microwave Extraction with
Measurement by Inductively Coupled Plasma
Emission Spectrometry 3-6
Test Sample Set for Round-Robin Analysis. Source of Bulk Materials,
Targeted Concentration and Final Concentration of Bottled Materials
Determined at RTI by Microwave Extraction with Measurement by
Inductively Coupled Plasma Emission Spectrometry 3-11
Consensus Values and Method Means for Paint Samples Submitted to
Round-Robin Analysis 5-4
Consensus Values and Method Means for Dust Samples Submitted to
Round-Robin Analysis 5-5
Recovery (%) by Method (Relative to Round-Robin Consensus Values)
of Paint and Dust Samples Submitted to Round-Robin Analysis . . . 5-6
Estimates of Sample-to-Sample Variation (Sample RSD), Repeatability
(Within-Lab Variation), and Reproducibility (Between-Lab Variation) of
Paint and Dust Samples Submitted to Round-Robin Analysis .... 5-8
Instrumental Detection Limits for Lead by Methods in the Round-
Robin 5-13
Repeatability and Reproducibility (%) by Method Averaged across
Matrices for Paint and Dust Samples Submitted to Round-Robin
Analysis 5-14
Method Evaluation Materials and Standard Reference Materials
Identified to Differ Significantly by Sample-Specific, Pairwise
Comparison of Method Means Determined by Round-Robin
Analysis
5-18
Comparison of Method Means of Test Samples Submitted to Microwave
Extraction Procedure Used in the Round-Robin with Concentrations
Determined by a Total Microwave Digestion at RTI 5-20
xn
-------
LIST OF FIGURES
Figure 1. Repeatability Versus Lead Concentration by Method 5-10
Figure 2. Reproducibility Versus Lead Concentration by Method . 5-11
Figure 3. 95% Confidence Interval for the Geometric Mean Recovery (%)
by Method 5-15
xm
-------
LIST OF APPENDICES
Appendix A. Statistical Approach
A-l Statistical Design of the Round-Robin
A-2 ISO Guide 35
Appendix B. Participating Laboratories
Appendix C. Standard Operating Procedures
C-l AAS/ICP SOP - Standard Operating Procedures for Lead in
Paint by Hotplate- or Microwave-based Acid Digestion and
Atomic Absorption or Inductively Coupled Emission
Spectrometry
C-2 Laboratory XRF SOP - Standard Operating Procedures for
Energy Dispersive X-ray Fluorescence Analysis of Lead in
Urban Soil and Dust Audit Samples
Appendix D. Instructions to Laboratories
D-l Letter of Instruction to AAS/ICP Laboratories
D-2 Letter of Instruction to Laboratory XRF Laboratories
D-3 RTI Copy of Data Reporting Form with Sequence Tracking
Appendix E. Reported Results
E-l MW/AAS Laboratories
E-2 HP/AAS Laboratories
E-3 MW/ICP Laboratories
E-4 HP/ICP Laboratories
E-5 Laboratory XRF Laboratories
xiv
-------
Appendix F. Letter Sent to Laboratories Reporting Preliminary Results
of Round-Robin
Appendix G. Statistical Analysis of Results
G-l Report by Larry Myers
G-2 Review of Statistical Analysis by Jack Suggs
G-3 Raw Data File
G-4 Missing/Censored Observations
G-5 Candidate Outlying Observations
G-6 Method Means, Consensus Values, Repeatability and
Reproducibility
G-7 Recovery and Log Recovery Plots by Laboratory Operation
G-7-1 MW/AAS
G-7-2 HP/AAS
G-7-3 MW/ICP
G-7-4 HP/ICP
G-7-5 Laboratory XRF
G-8 Plots of Repeatability/Reproducibility versus Lead
Concentration
G-9 Geometric Mean Recovery by Method
G-10 Method Effects and Pairwise Comparison of Method Means
Appendix H. Total Microwave Digestion Method
xv
-------
SECTION 1.0
INTRODUCTION
1.1 OVERVIEW
As a result of the growing concern about the adverse health effects
associated with exposure to lead in the environment, the identification and
assessment of hazards from lead-based paint (LBP) and LBP-containing dust and
soil have become critical environmental issues. Because the identification of LBP
hazards requires either field or laboratory analysis, an increasing number of lead-
containing matrices are being submitted to analysis. Unfortunately, there is a
lack of reference materials, materials of known concentrations, to support the
reliability of the results. Regulations in support of the establishment of lead
tester certification programs (Title X1) and a National Lead Laboratory
Accreditation Program2 (NLLAP) have been promulgated to ensure that these
decisions are based upon analytical data that is accurate, reproducible and
representative.
The analysis of reference materials, well-characterized, homogeneous
materials of known concentration, is necessary for the accurate calibration of
instruments and essential to the evaluation of laboratory performance in the
preparation and analysis of samples. Two types of reference materials are
important in analytical chemistry quality assurance:
• standard reference materials (SRMs) produced and certified by the
National Institute of Standards and Technology, and
• performance evaluation materials (PEMs).
Of the two types of reference materials, SRMs are more homogeneous and
more stringently characterized. The analytical uncertainty for SRMs is less than
or equal to 10 percent, as compared to 10 - 25 percent for PEMs3. Thus, SRMs
are more costly and less available for routine quality assurance/quality control
(QA/QC) activities. PEMs are more easily prepared, less costly than SRMs, and
1-1
-------
are therefore better suited for routine QC checks.
The purpose of this study was to develop and test a protocol for the
production of homogeneous performance evaluation materials, hereafter called
Method Evaluation Materials (MEMs), as prescribed by the U. S. Environmental
Protection Agency (U.S. EPA)-sponsored Lead Reference Materials Workshop4
(LRMW) held in May, 1991. The protocol was tested by round-robin analysis of
the concentration and homogeneity of the MEMs produced following the protocol.
In addition to the provision of concentration and homogeneity data for the series
of MEMs, the results of the round-robin allowed a comparison to be made of
proven extraction/analytical methods used by the participating laboratories.
The preparation and verification of the protocol was designed relative to
the following:
• establishment of target concentrations and homogeneity for the
method evaluation materials, consistent with proposals at the Lead
Reference Materials Workshop,4
* collection of real-world paint and dust,
• preparation of materials at the targeted concentrations,
• verification of the concentration and homogeneity of the MEMs by
analyses at RTI,
• designation of methods for analysis in the round-robin,
• recruitment of laboratories for measurement by select
extraction/analysis methods,
• statistical design of the round-robin
- identification of replicates,
- identification of Standard Reference Materials to be
submitted as blinds, and
- identification of a minimum number of laboratories analyzing
by a particular extraction/analysis,
• round-robin analysis of MEMs and SRMs,
1-2
-------
• statistical analysis of results, and
• conclusions and recommendations for further study.
The results of the round-robin study were expected to provide the following
data:
* method mean - a concentration for a test sample determined from
averaging the results reported by a specified method of analysis,
• consensus value - a concentration for a test sample determined by
averaging the method means determined by different laboratories
and/or methods,
• recovery by method - a ratio of the method mean to the consensus
value, expressed as percentage,
• repeatability - within-lab variation, the relative standard deviation
(%) determined for replicate samples analyzed in one laboratory,
• reproducibility - between-lab variation, the relative standard
deviation (%) determined for replicate samples analyzed by
laboratories using the same method, and
• sample-to-sample variation - the homogeneity of the material
determined from a test of the hypothesis that the variation between
replicate aliquots is zero.
The interpretation of data was applied to examine the following:
• protocol for MEM preparation by comparing the consensus values
with the targeted concentrations, with the expectation that the
targeted concentrations and consensus values agreed within 20%;
• sample-to-sample variation by comparing repeatability and
reproducibility of replicate samples analyzed by the same method;
and
• comparison of methods by determining
- the 95% confidence interval of method means, and
- the statistically significant differences by pairwise comparison
of method means.
1-3
-------
1.2 REPORT
This report describes the preparation of paint and dust method
evaluation materials and verification of the preparation protocol. The reader may
refer to the following sections for specific information:
• design and preparation of the materials - Sections 2 and 3,
• round-robin analysis - Section 4,
• statistical analysis of results of the round-robin - Section 5.
• summary and conclusions - Section 6, and
• suggestions for further study - Section 7.
1-4
-------
SECTION 2.0
DESIGN OF THE METHOD EVALUATION MATERIALS
2.1 CRITERIA ESTABLISHED BY THE LEAD REFERENCE
MATERIALS WORKSHOP
The design for MEMs was developed in a reference materials workshop held
May 13-14, 1991 in Washington, DC.4 The nature of "real-world" samples, health
effects, and regulations were considered to be the principal driving forces for the
preparation of MEMs. Subsequently it was decided that the matrices of the
reference materials match the matrices of the samples typically submitted to the
laboratory for analysis. Matrix-matching is critical because the nature of the
matrix is a significant factor in the effectiveness of extracting lead from paint and
dust samples; i.e., old dried paint samples extract differently from newly-prepared
paint films.5 Matching the matrix of reference materials and samples, i.e.,
binders, particle size, is also important for accurate analysis by Laboratory XRF.
2.1.1 Paint
It was decided in the workshop that paint be collected from dwellings at
least 40 years old. Assuming an aliquot of 0.25 g for atomic spectroscopic
analysis, it was proposed that the material be ground to a particle size of s200
microns in order for the aliquot to be representative of the bulk sample. A
concentration range of 500 to 50,000 /tg/g (0.05% to 5%) was proposed to cover the
current regulations.
2.1.2 Dust
It was suggested in the Workshop that "real-world" dust be collected for
preparation of reference material. No decisions were made about particle size,
although it was decided that an appropriate concentration range for reference
materials for lead in bulk dust of 50 to 10,000 /u,g/g be established to encompass a
concentration range inclusive of lead in hand wipes to post-abatement lead levels.4
2-1
-------
2.2 CONCENTRATIONS PROPOSED FOR METHOD EVALUATION
MATERIALS
It was decided that, practically, only a limited number of MEMs could be
analyzed as a means of evaluating the preparation protocol. Therefore, in order to
verify the preparation protocol by a determination of concentration and
homogeneity, it was decided that paint and dust MEMs be prepared only at two
different concentrations, and that each of the two concentrations be split into two
replicates and bottled as two separate samples. This would provide a total of four
samples of paint, and four samples of dust for testing.
For dust samples, a low level sample (approximating household dust) and a
high level sample (approximating post-abatement dust), were proposed. For paint
samples, a low level paint sample (having a concentration between the Consumer
Product Safety Commission (CPSC) action limit6 of 600 ftg/g and the Department
of Housing and Urban development (HUD) action level7 of 5000 jig/g), and a high
level sample (approximating a concentration commonly detected on the exterior of
older dwellings) were targeted. The following concentrations were proposed for
the MEMs:
• 100 fjig/g - low level dust (household),
• 1500 fj,g/g - low level paint,
• 4000 jtg/g - high level dust (post-abatement), and
• 40000 /ig/g - high level paint (exterior).
2-2
-------
SECTION 3.0
PREPARATION OF THE METHOD EVALUATION MATERIALS
As noted, an important consideration for the preparation of reference
materials is matching the matrix of the reference material to the matrix of the
samples typically submitted to analysis. Therefore, the preparation of the method
evaluation materials used in this study required the collection of "real-world"
paint and dust samples.
3.1 PAINT
Paint samples submitted to laboratory analysis are often multiple layers of
different kinds of paint that have embrittled from age and weathering. In order
to emulate samples submitted to a laboratory, the method evaluation materials in
this study were prepared from "real-world," multi-layered paint.
3.1.1 Collection of Materials
The collection of real-world samples was facilitated by contacts acquired
through RTI tasks in support of EPA programs for lead-based paint and lead-
based paint-containing matrices. The tasks performed for the EPA included
coordination of a preliminary round-robin8 for the evaluation of spectroscopic
methods for the analysis of lead in paint, dust and soil; coordination of Lead
Reference Materials Workshop4; and collection of lead-based paint for standard
reference materials (SRMs) prepared by the National Institute of Standards and
Technology (NIST). As a result of these tasks, RTI established an extensive
repository of lead-based paint containing matrices. This repository contains paints
from interior walls, interior woodwork, and exterior trim collected from
abatement and demolition projects across the country. The specific paint
materials used to prepare the test MEMs for this study were collected from a
vacant hospital in Athens, Ohio, The paint collected from this site was old, and
multi-layered from regular repaintings since the establishment of the hospital in
3-1
-------
the late 19th century. It was peeling from the substrate to such an extent that
the firm of Osborne and Assoc.,9 an abatement contractor, was able to collect the
chips by sweeping the floors and cold-scraping the walls and woodwork with
squeeges.
3.1.2 Selection of Bulk Materials
Preliminary screening analyses of paint samples were carried out at the
time of sample custody. Aliquots of several grams each were removed from each
of the bulk samples and ground by hand with a mortar and pestle. Aliquots were
then removed from the ground material and extracted by a microwave (MW)
method10 utilizing a combination of nitric acid (HNO3) and hydrochloric acid (HC1).
The concentration of lead in the extracts was measured by inductively coupled
plasma emission spectrometry (ICP).
The majority of the samples collected at the Athens site contained lead at
concentrations in the range of 5% to 40%, but two bulk paint samples having
concentrations of 3.8% and 0.36% were also identified. The 3.8% and 0.36%
materials were chosen for the preparation of the MEMs; though well above the
target of 0.15%, the 0.36% material was the lowest level available in the
repository.
3.1.3 Grinding
Both bulk paint samples were ground to a particle size of s250 microns (/mi)
in a crossbeater mill11, and then ground to a particle size «;120 /xm in a Retsch12
grinder.
3.1.4 Blending
The ground paints were individually mixed for 30 minutes in a Turbula13
blender.
3-2
-------
3.1.5 Determining the Effect of Aliquot Weight on Analytical Results
One of the concerns in development of a reference material is the effect of
aliquot weight on the analytical results. It is desirable to maximize an aliquot
size in order to minimize errors associated with lack of homogeneity in the
sample, while still achieving acceptable analyte recovery, i.e., ) 90%. Maximizing
aliquot size is particularly important for samples having lead concentrations near
the detection limit of the analytical method used. Therefore, the effect of the
aliquot weight on the analytical results was investigated by removing aliquots
from the high-lead and low-lead paint bulk materials, and analyzing the aliquots
by the MW/ICP method10.
Aliquot sizes of 50 mg, 100 mg, and 250 mg were selected for investigation
because these aliquot weights are commonly used in the analysis of environmental
samples with lead concentrations in a normal to high range (>10 fig/g to 120,000
/ig/g). For the determination, samples at the three different aliquot weights were
removed in duplicate from each bulk material. For example, two 50 mg aliquots,
two 100 mg aliquots, and two 250 mg aliquots were removed from the prepared
low and high lead-containing paint materials, yielding a total of 12 samples for
analysis. The results of the analyses are given in Table 1. A statistical evaluation
showed all of the measured concentrations to be equivalent at the 95% confidence
interval, except for the 250 mg aliquot of low paint. A review of the analytical
data indicated that this sample was measured at an instrumental (ICP)
concentration of 41.5 /ig/mL, well above the measured concentrations of the other
paint samples (and an instrumental range concentration later prescribed for the
round-robin evaluation of these materials). Because of the high instrumental
concentration of the 250 mg aliquot, ICP signal suppression was considered a
source of the depressed concentration of this sample relative to the 50 and 100 mg
aliquots. (The difference in AAS and ICP results will be discussed in Section 5.)
An aliquot weight of 100 mg was selected for the paint materials because this
3-3
-------
Table 1. Concentrations of Lead Measured in Paint and Dust Method Evaluation
Materials Relative to Changes in the Aliquot Weight for Extraction
Sample
Low Paint
High Paint
Low Dust
High Dust
Mean (jag/g) + SD (% RSD)
(n=2)
Aliquot Size
50 mg 100 mg 250 mg
3600 ± 7.06 (0.196)
36800 ± 1203 (3.27)
97.4 ± 29.2 (29.9)
4340 ± 503 (11.6)
3530 ± 42.4 (1.20)
36200 ± 283 (0.781)
79.8 ± 0.42 (0.53)
4160 ± 84.9 (2.04)
3310 ± 28.3 (0.854)
36000 ±425 (1.18)
81.2 ±0.71 (0.87)
4100 ±6.97 (0.17)
Legend:
% RSD = Percent Relative Standard Deviation
3-4
-------
weight gave consistently high recoveries. Increasing the weight to 250 mg would
not improve precision.
3'!-6 Production of Target 0.15% Material
As stated earlier, a bulk paint material having a lead concentration of
about 1500 jtg/g (0.15%) could not be located. Achieving this target concentration
was considered important to the evaluation process, and therefore, when a source
of bulk paint having a lead concentration lower than 0.36% could not be found, an
attempt was made to determine if separation of the layers of the multi-layered
chips would yield layers containing lead at different concentrations. It was
believed that the most recently applied layers, i.e., the outermost layers, would
contain lead at the lowest levels.
The 0.36% paint material was found to be a combination of multi-colored
layers of paint; therefore, it was possible to identify and separate (by hand) chips
that appeared to have the same colored layers, and were believed to have an
identical painting history. From these selected chips, the outermost layers were
removed with a scalpel to yield a paint sample representing the most recent
painting. This method was used to isolate a material that, upon analysis, showed
a concentration of 0.15%. The 0.15% material was carried through all the
preparation steps (grinding, blending) described for the preparation of the 0.36%
material. The previously prepared 3.8% material, and the 0.15% material were
designated as "high paint" and "low paint," respectively.
3.1.7 Preliminary Verification of Concentration and Homogeneity
The concentrations of both the low and high paint materials were
determined by analyzing 100 mg replicate aliquots (except the low paint material,
where n =1) of the prepared materials by the MW/ICP method.10 Results of the
concentration verification, given in Table 2, indicated that the targeted
concentrations for the selected samples were achieved. Acceptable homogeneity
was achieved as indicated by a relative standard deviation (RSD) of 1.87% for the
3-5
-------
Table 2. The Concentration and Homogeneity (RSD)
of Paint and Dust Method Evaluation Materials
Determined at RTI by Microwave Extraction
with Measurement by Inductively Coupled Plasma Emission Spectrometry
J S f ff f J f f
High Paint
Low Paint
High Dust
Low Dust
Concentration (ftg/g) +_ SD
36300 ± 679 (n=6)
1400 (n=l)
4130 ± 61.8 (n=4)
80.5 ± 0.938 (n=4)
RSD (%)
1.87
—
1.50
1.17
3-6
-------
high paint. Only one sample was analyzed for the preliminary verification of
concentration of the low paint; therefore, precision data were not available.
3.2 DUST
3.2.1 Collection of Materials
The RTI repository of lead-contaminated dust materials includes household,
hotel, street, and post-abatement dust. Household and hotel dust samples were
collected as vacuum cleaner bags; post-abatement dust was supplied to RTI as
High Efficiency Particulate Air (HEPA) vacuum cleaner bags from abatement sites
in the Midwestern and Eastern United States. Street dust was collected from
street sweepers in Durham, North Carolina.
Household dust, collected from local households and from households in
California, was used to prepare the low dust MEM for this evaluation. The high
dust MEM was prepared from HEPA-vacuumed dust collected from abatement
sites in Pennsylvania.
3.2.2 Sterilization
Because dust samples contain large amounts of debris, animal protein and
microbiological organisms, all bulk dust samples were sterilized by irradiation
prior to handling. Upon receipt at RTI, the bulk dust was shipped to Neutron
Products, Inc.,14 and gamma-irradiated for 12 hours for a total minimum dose of
2.5 MRads.
Although the samples were only visually examined for the growth of
microbiological organisms, it did not appear that the dust samples were
recontaminated from the post-sterilization opening of containers or from
atmospheric moisture. The bulk dust appeared to be stable after sterilization.
3.2.3 Removal of Debris
The sterilized bags of dust were returned to RTI and individually sieved to
remove debris and hair. The dust was sieved through a coarse (2.00 mm) and fine
3-7
-------
(250 /um) screen using a Ro-Tap15 apparatus.
3.2.4 Selection of Bulk Materials
Aliquots of 100 mg were removed from individual bags of sieved dust and
analyzed by the MW/ICP method10 in order to identify materials with appropriate
lead concentrations for the preparation of the MEMs.
3.2.5 Blending
Because the weight of sieved dust from one vacuum cleaner bag was
insufficient to provide enough material for the low dust sample, batches of sieved
household dust with concentrations approximately equal to 100 fig/g were blended
for 30 minutes in a Turbula13 blender to achieve an adequate weight of dust at the
targeted concentration. The concentration of lead in the blended material was
determined by removing four 100 mg aliquots and analyzing each by the MW/ICP
method.10 The results of the analysis for the blended household dust indicated a
concentration of 80 /ug/g, as targeted for the low dust sample.
It was not necessary to blend bulk samples of post-abatement dust because
the weight of the sieved sample was sufficient for the round-robin test samples.
The concentration of the post-abatement dust was found to be 4100 /-tg/g, as
targeted for the high dust sample.
3-2-6 Determining the Effect of Aliquot Weight on Analytical Results
The effect of aliquot weight on analytical results was also investigated for
the prepared dust samples. Aliquots of 50 mg, 100 mg, and 250 mg were removed
in duplicate from each of the low and high dust samples. The aliquotting was
analogous to that carried out for the paint materials; a total of 12 aliquots were
removed for analysis by the MW/ICP method10. Results of the analyses, given in
Table 1, indicated that the measured concentrations were consistent over the 50 t<
100 mg range of aliquot weights. Improvements in precision were observed with
increases in aliquot weight. An aliquot size of 100 mg was prescribed for the
3-8
-------
analyses because this weight gave acceptably precise results, and was consistent
with the aliquot size prescribed for the analysis of paint samples. The 95%
confidence intervals for the concentrations of the 50, 100, and 250 mg aliquots
were equivalent.
3.2.7 Preliminary Verification of Concentration and Homogeneity
The concentrations of the high and low dust samples were determined by
taking replicate 100 mg aliquots of the prepared materials and analyzing by the
MW/ICP method.7 The results of the analyses are given in Table 2. Acceptable
target concentrations and homogeneity (RSD z 1.50%) were achieved.
3.3 BOTTLING THE TEST SAMPLES
The method evaluation materials and the standard reference materials were
bottled by direct weighing of prepared materials into screw-cap bottles.
Approximately 150 bottles of each matrix were prepared by accurately weighing 5
grams each of the high and low paint, and 2 grams each of the high and low dust
into 20 mL plastic screw-cap bottles. During the transfers, the four stock
containers of the bulk high and low paint and dust materials were tumbled in all
directions several times after the removal of every 5 to 7 samples. The bottles
containing the MEMs were numbered sequentially to track the loading from the
bulk material. The sequence number was recorded by RTI.
The NIST Standard Reference Materials were bottled using the same
procedure as the method evaluation materials, i.e., 5 grams of NIST SRM 1579,
and 2 grams of NIST SRM 2711 were weighed into 20 mL plastic screw cap
bottles. The bottles of bulk SRMs were also tumbled through all directions after
every 5 to 7 aliquots were taken, and SRM samples were sequentially numbered
to track the loading from the stock material into the 20 mL bottles. The sequence
number was recorded by RTI.
3-9
-------
3.4 FINAL VERIFICATION OF CONCENTRATION OF THE METHOD
EVALUATION MATERIALS
Five bottles were removed at random from each of the four prepared sets of
method evaluation materials (high paint, low paint, high dust, and low dust).
From each bottle, five 100 mg aliquots were removed. (Bottles were tumbled
through all axes between the removal of each aliquot.) The aliquots were analyzed
by the MW/ICP method10. The final concentrations of the bottled materials
yielded samples with concentrations within 20 percent of the targeted range (100
- 100,000 /ig/g):
84.2 ± 11.9 /ig/g -low dust,
1060 ± 21.2 jig/g -NISTSRM2711
(1162 ± 31 fig/g - certified value),
• 1410 ± 44.5 /u.g/g - low paint,
4670 ± 330 j*g/g -high dust,
37900 ± 500 ftg/g - high paint, and
116000 ± 3500 jug/g - NIST SRM 1579
(118700 ± 400 /ig/g - certified value).
The targeted concentrations for the paint and dust samples, the sources of the
samples, and the final verified concentrations are presented in Table 3.
3-10
-------
Table 3. Test Sample Set for Round-Robin Analysis. Source of Bulk Materials,
Targeted Concentration and Final Concentration of Bottled Materials
Determined at RTI by Microwave Extraction
with Measurement by Inductively Coupled Plasma Emission Spectrometry
Samples
Low Paint
(P-l, P-4)
High Paint
(P-3, P-5)
Paint SRM
(P-2)
Low Dust
(D-2, D-4)
High Dust
(D-l, D-5)
Dust SRM
(D-3)
Source
Athens, Ohio
Athens, Ohio
NIST SRM 1579
Household dust,
NC&CA
Post-abatement
dust, PA
NIST SRM 2711
Targeted
Concentration
fcg/g)
1500
40,000
120,000
100
4000
1000
Concentration
(MW/ICP)
Mean (jug/g) ±
SD(%RSD) n=25
1,410 ± 44.5 (3.16)
37,900 ± 500 (1.35)
118,700 ±400 (0.34)
(certified value)
84.2 ± 11.9 (14.1)
4,670 ± 330 (7.07)
1162 ± 31 (2.67)
(certified value)
Legend:
MW = Microwave Digestion Method
ICP = Inductively Coupled Plasma Emission Spectrometery
3-11
-------
SECTION 4.0
ROUND-ROBIN ANALYSIS
OF THE METHOD EVALUATION MATERIALS
Following preparation of the MEM materials and their analysis within RTI,
the materials were farther evaluated by round-robin analysis. A statistical design
for the round robin was developed by the U.S. EPA and is presented in Appendix
A-l.
4.1 ROUND-ROBIN DESIGN
The design called for each laboratory to receive as blind samples two bottles
of each of the four MEMs. Each laboratory was also to receive a sample of each
matrix at a third concentration. This third material, a standard reference
material (SRM), provided one additional sample per matrix, and was also
submitted as a blind sample. A suggestion was made to include two blind samples
of the same SRM, consistent with the submission of two MEM samples of the
same concentration, but this suggestion was rejected because of the increased
number of analyses, and thus cost incurred, for the participating laboratories. As
a result, a total of ten samples were planned for submission to round-robin
analysis.
Each laboratory was requested to remove two aliquots from each sample,
thereby preparing and analyzing each sample in duplicate. As a result, a total of
twenty (20) results were to be reported for each laboratory operation.
The samples were to be either extracted using a specified hotplate or
microwave method, and analyzed by atomic absorption spectrometry (AAS)
inductively coupled plasma emission spectrometry (ICP); or to be analyzed by
Laboratory X-ray Fluorescence (Lab XRF). These methods were chosen because of
their relevance to analyses carried out for environmental lead samples.
Laboratory XRF was included because it had performed successfully using the
protocols outlined in the EPA Urban Soil Lead Abatement Demonstration Project
(Three City Study)16. The methods of analysis (extraction/analytical and
4-1
-------
Laboratory XRF) resulted in a total of five candidate methods:
Method 1 - MW/AAS,
Method 2 - HP/AAS,
Method 3 - MW/ICP,
Method 4 - HP/ICP, and
Method 5 - Laboratory XRF.
ISO Guide 3517 (Appendix A-2) provided a reference for the statistical
evaluation, and for expressing the results of the homogeneity testing. (See Section
5.6.)
4.2 RECRUITMENT OF LABORATORIES
A number of laboratories were recruited on the basis of their participation
in a previous round robin8, or as contacts facilitated through other tasks carried
out by the Research Triangle Institute (RTI) in support of EPA lead programs.
The goal was the recruitment of a minimum of eight to ten laboratories for
analysis of the samples by each of the five operations. A total of 36 laboratories
were recruited for participation in the round-robin; 11 of the 36 laboratories
agreed to analyze samples by two methods, resulting in the potential of 47
analytical operations. Projected participation by operation was a follows:
MW/AAS - 9 operations
HP/AAS - 9 operations
MW/ICP - 9 operations
HP/ICP - 12 operations
Laboratory XRF - 8 operations
At the completion of the round, results for 42 operations were reported by 33
4-2
-------
laboratories. A list of participating laboratories is provided in Appendix B.
4.3 ROUND-ROBIN ANALYSIS
4.3.1 Standard Operating Procedures
Standard Operating Procedures (SOPs) were sent to all participating
laboratories prior to the submission of the test samples. The protocols provided
to laboratories are given in Appendix C.
4.3.1.1 Analysis by Atomic Absorption Spectrometry or Inductively
Coupled Plasma Emission Spectrometry -
The EPA/AREAL report, "Standard Operating Procedures for Lead in Paint
by Hotplate- or Microwave-based Acid Digestion and Atomic Absorption or
Inductively Coupled Plasma Emission Spectrometry,"10 was sent to laboratories
analyzing by AAS or ICP. Laboratories analyzing by these methods were
instructed to follow the protocols provided in the SOP. The SOP is provided in
Appendix C-l.
4.3.1.2 Analysis by Laboratory X-ray Fluorescence
A reference draft protocol from the US EPA Environmental Monitoring
Systems Laboratory (EMSL)/Las Vegas, "Standard Operating Procedures for
Energy-Dispersive X-ray Fluorescence Analysis of Lead in Urban Soil and Dust
Audit Samples,"18 was provided to laboratories analyzing by laboratory X-ray
fluorescence. Laboratories were asked to follow the protocol specified in the
EMSL/Las Vegas document only if the laboratory did not have a protocol for the
analysis of dust. The draft SOP is included in Appendix C-2 to provide a record
of the information sent to participating XRF laboratories.
Two dust audit samples prepared by the EMSL/Las Vegas for the EPA
Urban Soil Lead Abatement Demonstration Project16 were provided to the
laboratories analyzing by Laboratory X-ray fluorescence. These audit materials,
BAL-1 and CIN-1, had lead concentrations of 58 pg/g and 2275 /ig/g, respectively.
4-3
-------
The audit samples were provided to all laboratories because some of the
participating laboratories did not have suitable calibration standards for the
analysis of dust. In order to establish a consistency in the instrument calibration,
all laboratories using the XRF method were asked to use BAL-1 and CIN-1 to set
up a calibration curve for the analysis of the dust samples.
4.3.2 Letter of Instructions
A letter of instructions was submitted to the laboratories along with the set
of test samples. Exemplary letters sent to AAS/ICP and Laboratory XRF
participants are provided in Appendix D.
Laboratories were requested to tumble every sample bottle prior to analysis,
and to carry out analyses in duplicate. If an extraction technique was used, the
laboratory was asked to remove two 100 mg aliquots, carry each aliquot through
the extraction procedure, and analyze the extract. XRF laboratories were
instructed to remove two sufficiently large aliquots to prepare "infinitely thick"
samples for analysis.
4.3.3 Data Reporting Form
Laboratories were requested to report results to RTI in a Data Reporting
Form provided by RTI. The form indicated the name of the laboratory and its
assigned identification number for the round-robin, as well as the extraction
and/or analytical method to be performed for the analyses. A space was available
for the laboratory to indicate its experience (number of years) with the method.
Exemplary Data Reporting Forms are provided in Appendices D-l and D-2, for the
extraction methods and Laboratory X-ray fluorescence, respectively. Sequence
numbers for loading samples shipped to a participating laboratory were recorded
on an RTI copy of the laboratory's Data Reporting Form. Exemplary copies are
provided in Appendix D-3. Completed Data Reporting Forms (coded by laboratory,
and categorized by method) are provided in Appendices E-l through E-5.
4-4
-------
4.3.4 Instrument Parameter Forms
Forms were included with the set of samples for the laboratories to provide
instrumental parameters appropriate to the analyses. AAS/ICP laboratories were
asked to provide information including manufacturer, model number, background
correction, and calibration data. Laboratory XRF parameters, i.e., manufacturer,
sample preparation, X-ray source, were requested of these laboratories.
Laboratories were requested to submit the forms to RTI along with the Data
Reporting Forms. Instrumental parameter forms are provided in Appendices D-l
and D-2 for AAS/ICP and Laboratory XRF analyses, respectively. Results were
due to RTI no later than April 30, 1992.
4.3.5 Responses From Participating Laboratories
A total of 42 sets of results were reported to RTI from 33 participating
laboratories. (Nine laboratories analyzed the test samples by two different
methods.) The final distribution of results by method was as follows:
MW/AAS - 7,
HP/AAS - 9,
MW/ICP - 9,
HP/ICP - 10, and
Laboratory XRF - 7.
Two laboratories did not return MW/AAS data because the laboratories
encountered problems with melted and/or imploded plastic centrifuge tubes. (The
tubes were required for the microwave extraction procedure,10 and were supplied
by RTI. One laboratory carried out subsequent analyses using a total digestion by
a HP/ICP method; the results from the total digestion were not included in the
statistical analysis. Two laboratories encountered problems believed to be
attributed to the homogeneity and/or prescribed aliquot size for the low dust
4-5
-------
material. One laboratory found that repeated analyses of the same extract of the
low dust sample gave repeatable results, yet poor repeatability was achieved when
replicate aliquots were removed, and each was extracted and analyzed.
4.3.6 Notification Of Results
Following the statistical analysis of results (presented in Section 5), letters
were sent to participating laboratories summarizing the results of the preliminary
statistical analysis. The letter included tables from a draft paper to be published
in the proceedings of the American Chemical Society Symposium, "Lead Poisoning
in Children: Exposure, Abatement and Program Issues/'19 held in August, 1992.
This letter is provided in Appendix F.
4-6
-------
SECTION 5.0
STATISTICAL ANALYSIS OF RESULTS
A statistical analysis20"22 of the data submitted by the participating
laboratories was performed to determine the following:
• mean concentration by method for each of the six test samples,
• consensus value for each of the six test samples,
• statistically significant differences between method means,
determined for each of the six test samples,
* homogeneity (sample-to-sample variation of the material),
• repeatability (within-lab variance) by method, and
• reproducibility (between-laboratory variance) by method.
The report of the statistical analysis by RTI statistician Dr. Larry Myers is
provided in Appendix G-l. The statistical analysis was reviewed by EPA
statistician Mr. Jack Suggs. This review is provided in Appendix G-2.
5.1 CENSORED, MISSING DATA
A total of 33 laboratories reported results for 42 combinations of
extraction/analysis methods. Analyses of 10 test samples (blind duplicate high
and low paint and dust samples, and single blind samples of SRMs 1579 and
2711) were carried out in duplicate for a total of 20 reported results per
extraction/analysis. One laboratory reported triplicate results; two results were
not reported. Therefore, a total of 848 results were examined statistically. The
original data entries for statistical analysis (raw data) is provided in Appendix
G-3; missing and censored observations are provided in Appendix G-4.
5-1
-------
5.2 OUTLYING DATA
At the outset, results that were reported non-quantitatively, i.e., less than
a specific concentration (primarily for the low dust sample), were excluded from
the statistical analysis, yielding 820 results to be examined for outliers.
For each of the six combinations of matrix (dust, paint) and level (high, low,
and SRM), a nominal concentration was calculated as the median of all reported
results from the extraction methods. Laboratory XRF data were excluded because
of the following factors:
• a preliminary statistical examination of the data indicated a negative
bias relative to data for the extraction methods, and
• XRF analyses were not carried out using a standardized SOP, as in
the case of the AAS/ICP analyses.
A recovery for each extraction method result was calculated as the ratio of the
reported concentration divided by the nominal concentration. Using recoveries
between 0.35 and 2.00, the average and standard deviation of the recovery was
calculated for each of the method (5) by matrix (2) by level (3) combinations (a
total of 30 combinations). (The restriction to recoveries between 0.35 and 2.00 was
a prescreen intended to remove grosser outliers having the potential of distorting
the final means and standard deviations.) For each of the 820 reported results, a
score for the recovery was calculated by subtracting the average recovery from the
individual calculated recovery and dividing by the standard deviation of recovery
for the given combination. Any measurement whose absolute recovery score
exceeded 2.76 was excluded as an outlier. (Candidate outlying observations are
provided in Appendix G-5.) This corresponded to the upper and lower one-half of
one percent of a normal distribution. As a result of this screening, an additional
28 reported results were excluded, allowing a total of 792 results for statistical
analysis.
5-2
-------
5.3 METHOD MEANS
The method mean for each of the six samples (low paint, high paint, paint
SRM, low dust, high dust, and dust SRM) was determined as the average of all
reported results, excluding censored results and outliers. Standard deviations and
relative standard deviations (RSDs) were determined. RSDs were found to be in
the ranges of 1.8% to 11.8% for the paint samples, and 2.2% to 9.2% for the dust
samples. These results are presented in Tables 4 and 5, and in Appendix G-6.
5.4 CONSENSUS VALUES
Consensus values for each of the six samples were calculated as an average
of the method means for the four extraction methods. The standard deviation of
the consensus value for a given sample was determined as the pooled standard
deviation of the mean by method. These values are provided in Tables 4 and 5,
and in Appendix G-6. (The standard deviations calculated and provided to the
laboratories in the notification letter .differ from the standard deviations given in
Tables 4 and 5 because the data reported to laboratories were based upon
preliminary calculations of simple standard deviations of the means. After the
notification letter was sent, it was decided that pooled standard deviations were
more statistically appropriate. Pooled standard deviations for the consensus
values were then determined and are given in Tables 4 and 5.)
For the reasons given for the exclusion of Laboratory XRF data from the
determination of a recovery score, Laboratory XRF values were also excluded from
determination of the consensus values. Method recoveries were calculated as a
ratio of method means to the consensus values, and are presented as percentages
in Table 6.
5-3
-------
Table 4. Consensus Values and Method Means
for Paint Samples Submitted to Round-Robin Analysis
Matrix/
Sample No.
High Paint
(P-3, P-5)
Low Paint
(P-l, P-4)
Paint SRM
(P-2)
NIST 1579
Certified Value:
118,700 ± 400
Consensus
Value" Otg/g) ±
SDb (%RSD)
37,632 ± 861
(2.3)
1690 ± 63
(3.8)
109,859 +
6521
(6.0)
Method
MW/AAS
HP/AAS
MW/ICP
HP/ICP
LabXRF
MW/AAS
HP/AAS
MW/ICP
HP/ICP
Lab XRF
MW/AAS
HP/AAS
MW/ICP
HP/ICP
LabXRF
Method Mean (jug/g)
+ SD (% RSD)
41,281 i 1,274 (3.1)
36,921 + 713 (1.9)
36,654 i 672 (1.8)
35,670 i 796 (2.2)
27,404 ± 1,567 (5.7)
1,896 + 63 (3.3)
1,661 + 74 (4.5)
1,603 + 45 (2.8)
1,600 + 66 (4.1)
1,034 i 76 (7.4)
122,432 ± 6,507 (5.3)
104,340 ± 8,681 (8.3)
118,281 ±2,476 (2.1)
94,382 i 7,021 (7.4)
112,721 + 13,259
(11.8)
8Lab XRF excluded from consensus value determination.
bPooled standard deviations
Legend:
MW =
HP =
ICP =
AAS =
XRF =
SRM =
Microwave Method (EPA/AREAL)
Hotplate Method (NIOSH 7082)
Inductively Coupled Plasma Emission Spectrometry
Atomic Absorption Spectrometry
X-Ray Fluorescence
Standard Reference Material
5-4
-------
Table 5. Consensus Values and Method Means
for Dust Samples Submitted to Round-Robin Analysis
Matrix/
Sample No.
High Dust
(D-l, D-5)
Low Dust
(D-2, D-4)
Dust SRM
(D-2)
NIST 2711
Certified Value:
1162 ± 31
Consensus
Value8 (jjig/g) ±
SDb
4550 ± 120
(2.7)
104 ±6
(5.8)
1186 ± 44
(3.8)
Method
MW/AAS
HP/AAS
MW/ICP
HP/ICP
Lab XRF
MW/AAS
HP/AAS
MW/ICP
HP/ICP
Lab XRF
MW/AAS
HP/AAS
MW/ICP
HP/ICP
Lab XRF
Method Mean Qxg/g)
+ SD
(% RSD)
4,847 ± 127 (2.6)
4,677 ± 103 (2.2)
4,281 ± 113 (2.6)
4,397 ± 133 (3.0)
2,485 ± 117 (4.7)
114 ± 6 (5.3)
108 ± 7 (5.3)
98 ±3 (3.1)
98 ± 9 (9.2)
93 ± 8 (8.6)
1,327 ± 72 (5.4)
1,173 ± 32 (2.7)
1,133 ±24 (2.1)
1,1 12 ±42 (3.8)
1,029 ± 33 (3.2)
"Lab XRF excluded from consensus value determination.
bPooled standard deviation
Legend:
MW =
HP =
ICP =
AAS =
XRF =
SRM =
Microwave Method (EPA/AREAL)
Hotplate Method (NIOSH 7082)
Inductively Coupled Plasma Emission Spectrometry
Atomic Absorption Spectrometry
X-Ray Fluorescence
Standard Reference Material
5-5
-------
Table 6. Recovery (%) by Method8 (Relative to Consensus Values)
of Paint and Dust Samples Submitted to Round-Robin Analysis
Paint
Method
MW/AAS
MW/ICP
HP/AAS
HP/ICP
High
110
97.4
98.1
94.8
Low
112
94.9
98.3
94.7
SRM
111
108
95.0
85.9
Dust
High
107
94.1
103
96.6
Low
110
94.2
104
94.2
SRM
112
95.5
98.9
93.8
aLab XRF recoveries were not determined because these results were excluded from
the determination of consensus values.
5-6
-------
5.5 REPEATABILITY AND REPRODUCIBILITY
Repeatability and reproducibility are expressions of the within-laboratory
and between-laboratory relative standard deviations measured for the six samples
(low paint, high paint, SRM paint, low dust, high dust, and SRM dust),
respectively. The values are based on the one-way analysis of variance of log
recoveries, ignoring sample-to-sample differences (previously determined to be
non-significant, and absorbed in the estimates of repeatability and reproducibility).
Values determined for repeatability and reproducibility are provided in Table 7.
The data in the table indicate that Laboratory XRF gave the most repeatable
results, i.e., lowest percentage of variation for all six samples. The repeatability
of Laboratory XRF is significant, subject to the caveat that the log transformation
may not have sufficiently stabilized the variances in the methods. If the variances
were stabilized by the log transformation, the reduction in within-lab variability
observed for XRF measurements could be attributed to minimal steps required for
sample preparation in XRF analysis.
Reproducibility is the more significant measure of variation in methods
because it reflects both within-laboratory variance and between-laboratory
variance. In general, the data in Table 7 indicate that Laboratory XRF is the
least reproducible method for the analysis of the paint samples, whereas the
MW/ICP method is the most reproducible method for the analysis of this matrix.
The HP/ICP method showed the poorest reproducibility for the analysis of the low
and high dust samples.
The differences in reproducibility of the Laboratory XRF method and the
extraction methods were attributed to the instructions provided for the analyses.
Laboratories using an extraction method were instructed to follow a specific
protocol; whereas, XRF laboratories were provided with a protocol for dust
5-7
-------
Table 7. Estimates of Sample-to-Sample
Variation (Sample RSD), Repeatability (Within-Lab Variation),
and Reproducibility (Between-Lab Variation)
of Paint and Dust Samples Submitted to Round-Robin Analysis
Matrix
Low Paint
High Paint
Low Dust
High Dust
Paint SRM
DustSRM
Parameter
Mean (jtg/g)
Sample RSD (%)
Repeatability (%)
Reproducibility (%)
Mean 0*g/g)
Sample RSD (%)
Repeatability (%)
Reproducibility (%)
Mean Oig/g)
Sample RSD (%)
Repeatability (%)
Reproducibility {%)
Mean (/*g/g)
Sample RSD (%)
Repeatability (%)
Reproducibility (%)
Mean (/ig/g)
Repeatability (%)
Reproducibility {%)
Mean (pg/g)
Repeatability (%)
Reproducibility {%)
Methods
MW/AAS
1896
4.2
11.5
13.3
41281
(0.1
5.6
9.5
114
(0.1
18.3
20.2
4847
(0.1
6.2
8.9
122432
7.2
14.8
1327
3.2
14.2
HP/AAS
1661
(0.1
12.4
17.7
36921
(0.1
4.9
7.1
108
<0.1
12.2
20.6
4677
3.5
6.2
8.9
104340
6.2
30.2
1173
3.7
8.9
MW/ICP
1603
<0.1
11.9
13.3
36654
(0.1
3.8
6.5
98
(0.1
16.0
16.5
4281
(0.1
.9.6
10.6
118281
4.4
7.1
1133
5.1
7.5
HP/ICP
1600
2.2
9.7
16.2
35670
{0.1
4.5
8.2
98
8.9
24.5
35.3
4397
(0.1
11.5
13.7
94382
12.5
29.0
1112
3.2
12.7
LabXRF
1034
(0.1
3.4
18.3
27404
(0.1
3.3
15.7
93
(0.1
8.6
22.2
2485
(0.1
3.7
13.2
112721
1.3
32.4
1029
2.5
8.7
Repeatability =
Reproducibility
Within-Lab Variation
Between-Lab Variation
5-8
-------
analysis only as a reference, and were instructed to follow their own protocol, if
available.
The quadratic tendency observed in lab-specific recovery plots for analysis
by Laboratory XRF suggested that calibrations were made with an inadequate
number of standards. (Recovery plots are provided in Appendix G-7.) XRF
laboratories provided their own paint standards for calibration, but two dust audit
samples, BAL-1 and CIN-1, were provided by RTI for use as calibration standards
for the analysis of dust. It is possible that instructions to generate a dust
calibration curve using only the two audit samples, BAL-1 and CIN-1, resulted in
the poor reproducibility observed for the dust samples. However, it should be
noted that laboratories provided their own standards for the calibration of paint;
and average reproducibility for this matrix was poorer than the average
reproducibility for the analysis of the dust. On the basis of these results, it
appears that the calibration differences, alone, do not explain the high value for
reproducibility by Laboratory XRF.
In order to provide a graphical description of the differences in repeatability
and reproducibility with concentration, the results of the analysis of variance
(expressed in /ng/g) are plotted across a concentration range determined as the
mean concentration by each method of the six samples (low dust, dust SRM, low
paint, high dust, high paint, and paint SRM). The logs of the variance for both
paint and dust matrices were approximately equal, so it was deemed feasible to
generate plots of reproducibility/repeatability for both matrices in the same
regression. Paint and dust matrices were pooled to provide a useful concentration
range for comparisons of repeatability and reproducibility. (This range would have
been limited if paint and dust matrices were examined separately.) Plots for each
method were prepared from a regression of the logs of repeatability/reproducibility
versus the log of the method mean, then exponentiating to generate the plots.
These plots are shown in Figures 1 and 2, and in Appendix G-8. The figures allow
5-9
-------
7
o
X
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
I
0.0
HP/MS
MW/ICP
MW/AAS
HP/MS
Lab XRF
I
0.2 0.4 0.6 0.8 1.0
Method Mean x 10*5 (|ig/g)
Figure 1. Repeatability versus lead concentration by method.
1.2
-------
1.8
b
X
o
i
O
s.
0
cc
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
0.0
0.2
HP/ICP
Lab XRF
MW/AAS
HP/MS
MW/ICP
I
1.0
0.4 0.6 0.8
Method Mean x 10*5 ftig/g)
Figure 2. Reproducibility versus lead concentration by method.
1.2
-------
a visual comparison of reproducibility and repeatability relative to concentration
over the operating range of the methods. The regressions are forced through zero
so that lines have a common origin; and the slopes, the change in repeatability or
reproducibility per unit change in concentration, may be compared. The
representations are a qualitative description, only; they are valid over the
operating range of the method, but do not attempt to model the performance of the
method at minimum detection. (Detection limits for the methods, presented in
Table 8, were provided in the RTI Standard Operating Procedure10 submitted to
participating laboratories.)
Another representation of the variability is to pool the data over the
concentration ranges and matrices to calculate overall repeatability and
reproducibility by method. These data are provided in Table 9.
Figure 3 shows the 95% confidence intervals of the geometric mean
recoveries (method mean/consensus mean) for the five methods examined. The six
horizontal lines associated with each method represent the six samples, and thus,
six concentration levels (SRM 1579, high paint, high dust, low paint, SRM 2711,
and low dust) examined in the round robin. L, M, and U correspond to the low,
mean, and upper limits of the 95% confidence interval, respectively. Plots of the
geometric means by method are provided in Appendix G-9.
5.6 SAMPLE HOMOGENEITY
The round robin was designed to examine sample homogeneity using a two-
way analysis of variance of logs for the blind duplicate MEM samples.
Application of this method to the analysis treated sampling, analysis, and their
interaction as random effects. For example, laboratories using the same method
(MW/AAS, MW/ICP, HP/AAS, HP/ICP, or Laboratory XRF), and replicate samples
selected from the same parent stock (P-l and P-4; P-3 and P-5; D-l and D-5; and
D-2 and D-4; see Tables 4 and 5) were both viewed as random selections from a
normal distribution. The assumption of random effects is appropriate in order to
5-12
-------
Table 8. Instrumental Detection Limits for Lead
by Methods in the Round-Robin
Method j
MW/AAS
HP/AAS
MW/ICP
HP/ICP
Laboratory XRF
IDLa
0.1 ng/mL
0.1 jug/mL
0.05 /ig/mL
0.05 jtig/mL
3^g/g
MDLb
20/tg/g
100 ^g/g
10jLtg/g
50/ig/g
—
instrument Detection Limit - /xg Pb/mL extracted solution
bMethod Detection Limit - /j.g Pb/g matrix
5-13
-------
Table 9. Repeatability and Reproducibility (%) by Method
Averaged across Matrices for Paint and Dust Samples
Submitted to Round-Robin Analysis
Method
MW/AAS
HP/AAS
MW/ICP
HP/ICP
LabXRF
Repeatability
10.7
9.7
10.5
12.9
4.8
Reproducibility
13.7
17.2
11.7
21.0
19.4
5-14
-------
JJ,
Microwave/
AAS
* W'*-' 5
Hotplate/
AAS
SWBWK
Microwave/
ICP
^ pn"1
^ w.. ..w ^A w. vJUIv..^ ><4*lls w. .v... xw v<- XXX. -V%K ««,*
f mm |i
H» iWl W
®
Hotplate/
ICP
:-:->J •»
M
• i _
: l<^ *>
a
b
c
d
e
Lab XRF
^
»* ivi
L
„
1
L
U
Ivi
1
, M , ™
-.-. M
U
-L
1
~ U N - » - - s
, s S ,
V N, X
L A/I U
M
1 1
- nn •«•• •-- --- *••• "«• •---• «> «•• •«» ~c »:< <-^ w t*x
<••.-. vx-: vs» x s% %\ •* *. NV-\ % ^ -.% x*x- :•»>:
•:^: •:•:•:•:• :%•;-:• x-:« «»c •>;-:-:- :•«•:• :->xx -x%-: -:•:•:•> x>» X-K-:
••»> «» »» :«« «« ^ »» w« «« «» xw. x«
1 1 I i
b
• c
d
e
f
50 60 70 80 90 100 110 120 130
Figure 3. 95% Confidence Interval for the Geometric Mean Recovery (%) by Method.
LEGEND a Paint SRM 1579
b High Paint
c High Dust
d Low Paint
e Dust SRM 2711
f Low Dust
108,826 ng/g (Certified Value: 118,700 ± 400 \ig/g)
37,306 n.g/g
4456 ng/g
1676jig/g
1176n.g/g (Certified Value: 1162±31ng/g)
1041
95% Confidence Interval:
L - Lower limit
M - Mean
U - Upper limit
5-15
-------
generalize results to a larger population of laboratories. This model was fit
separately to all 20 combinations of method (5) by matrix (2) by level (2) for all the
method evaluation materials.
A preliminary test for the absence of interaction or interdependence
between sample and laboratory analysis indicated that this assumption was
reasonable. Only one of twenty interaction tests was significant at the 5% level
with this data set (low dust by MW/AAS: 0.025 < p < 0.5). this is the expected
number of rejections by chance alone, under the null hypothesis of no interaction.
Accepting the hypothesis of no interaction means that the contributions of
sampling and analysis to the total variance can be considered to be additive.
The two-way analysis of variance was applied to calculate the relative
standard deviations (RSDs) for the samples. The RSD is equivalent to the
difference in concentration between samples, expressed as percentage. In one case
only (low dust by HP/ICP), the difference between samples was significant (8.9%).
In all other cases, the sample-to-sample differences were less than 0.1% (16 out of
20 cases) or non-significant relative to the variance of the measurement method.
On the average over the 20 cases, the sampling component of variance accounted
for 1.37% of the total variance, with a 95% upper confidence limit for the sampling
coefficient of variance being below 2.5%. It was, therefore, concluded that at the
95% confidence level, the concentrations of samples selected from the bulk
materials were within 5% (between 95% and 105%) of the concentrations given as
the consensus values. The RSD values are shown in Table 7.
The conclusion is that the bulk sample materials prepared by RTI were
homogeneous, and that sample-to-sample variation did not significantly contribute
to the analytical differences measured. According to criteria established in ISO
Guide 3517 (Appendix A-2), the method evaluation materials were considered "very
homogeneous material."
5-16
-------
5.7 PAIRWISE COMPARISON OF METHOD MEANS
Pairwise comparisons of method means within each of the six samples were
performed using ordinary nonsimultaneous t tests at the 95% confidence level.
There were ten possible paired comparisons of methods for each of the six samples
(60 total comparisons), so three (5%) rejections of the null hypothesis were
expected from chance alone. The results of the pairwise comparisons are
presented in Table 10. The statistical comparisons indicated no declared
differences for analysis of the low dust sample, and only two declared differences
for the paint SRM. A total of 28 differences were declared; of these differences, 26
were associated with MW/AAS and Lab XRF, methods that generated extreme
method means for five samples. Lab XRF gave the minimum mean for all samples
except for the paint SRM. MW/AAS gave the maximum mean for all of the
samples. This is a significant finding because the chance of equivalent methods
generating a maximum or minimum result for 6 out of 6 samples is 0.000064. The
statistical interpretation of the method effects is provided in Appendix G-10.
5.8 COMPARISON OF MEASUREMENTS BY ATOMIC ABSORPTION
SPECTROMETRY AND INDUCTIVELY COUPLED PLASMA
EMISSION SPECTROMETRY
As a part of RTFs earlier tasks in support of EPA programs for the analysis
of lead in environmental matrices, RTI carried out method development studies for
the analysis of lead by AAS and ICP. In these studies, low recoveries were found
for the analysis of NIST SRM 1579 by ICP relative to analysis by AAS.8 This
bias was believed to be caused by ICP signal suppression from matrix effects
associated with the paint samples. Because of these observations, RTI instructed
the round-robin laboratories analyzing by ICP to dilute the paint and dust extracts
into the 1 to 10 /ig/mL range prior to analysis, and instructed AAS laboratories to
use background correction, as specified in the SOP10 (Appendix C-l, Sections
1.2.3.1.2, and 4.5.1) sent to the laboratories. Despite these instructions, the data
5-17
-------
Table 10. Method Evaluation Materials and
Standard Reference Materials Identified to Differ Significantly
by Sample-Specific, Pairwise Comparison of Method Means
Determined by Round-Robin Analysis
Method
HP/AAS
MW/ICP
HP/ICP
LabXRF
MW/AAS
Low Paint
High Paint
Dust SRM
Low Paint
High Paint
High Dust
Dust SRM
Low Paint
High Paint
Paint SRM
High Dust
Dust SRM
Low Paint
High Paint
High Dust
Dust SRM
HP/AAS
—
High Dust
None
Low Paint
High Paint
High Dust
Dust SRM
MW/ICP
—
—
Paint SRM
Low Paint
High Paint
High Dust
HP/ICP
—
—
—
Low Paint
High Paint
High Dust
5-18
-------
showed that AAS results were higher than ICP results for paint and dust samples
by 3.5% to 18%, and 4.8 to 17%, respectively.
The difference in MW/AAS and MW/ICP results observed in the round-robin
was investigated by digesting the round-robin test samples using a total digestion
MW method and analyzing by ICP, with the addition of an internal standard. The
method used for the total digestion was a combination of methods used by the U.S.
Fish and Wildlife Service23 and the Institute of Chemical Industry and Metallurgy
of China.24 (The RTI method and the reference methods23'24 are provided in
Appendix H.) The concentrations determined by this extraction/analysis
method23*24 were compared with the results reported for the MW extractions in the
round-robin. The data are provided in Table 11. With the exception of the high
dust sample, the concentrations measured by the total digestion MW/ICP method
agreed closely with the round-robin MW/ICP results, but were consistently lower
than the round-robin MW/AAS results. These data suggest that the difference in
AAS and ICP results observed in the round-robin resulted from AAS signal
enhancement, rather than ICP signal suppression. In fact, a review of
instrumental parameter forms submitted by AAS laboratories indicated that a
number of laboratories did not use background correction, a common source of
positive bias, even though the SOP prescribed background correction for AAS
measurements. This was considered a plausible explanation for the bias
observed.
5-19
-------
Table 11. Comparison of Method Means of Test Samples Submitted to Microwave
Extraction Procedure Used in the Round-Robin with Concentrations Determined
by a Total Microwave Digestion at RTI
Sample
..... .,....„.„
" 5 -,' , "-, v, , -'
"- '/,-> £'""' ' ~ s-
" >,•*'>**',',:',
^ ' ' A v ' m f if
' ' t ' , J , ,'s, ,
' ", '''"',*", ',' '
/ , 'j W « , ^ ',, , ,
High Paint
Low Paint
Paint SRM
High Dust
Low Dust
Dust SRM
Concentration of Lead (/xg/g)
Round-Robin
MW/ICP
(n=36)
36,654 ± 672
1603 ± 45
118,281 ± 2476
4281 + 113
98 ±3
1133 ±24
MW/AAS
(n=28)
41,281 + 1274
1896 ± 63
122,432 ± 6507
4847 ±127
114 ±6
1327 ± 72
Total Digestion
(n=l)
MW/ICP*
36,000
1620
118,700
4960
108
—
MW/AAS
37,000
1715
121,000
4960
136
—
*Concentrations corrected by addition of internal standard
5-20
-------
SECTION 6.0
SUMMARY AND CONCLUSIONS
The round-robin study showed that the protocol used to prepare the paint
and dust method evaluation materials provided homogeneous materials at
targeted concentrations. The hypothesis of homogeneity was accepted in 19 out of
20 cases. (At the 95% confidence level, 1 rejection in 20 is expected by chance
alone.) In 16 of the 20 cases, the sampling component of variance was less than
0.1; in 4 cases the sampling component was less than or equal to 10% of the total
variance. On the average, the sampling component accounted for 1.37% of the
total variance.
The five methods examined as a part of the round-robin study performed
differently, with AAS methods producing results with a positive bias relative to
ICP results. An explanation proposed for the bias was the absence or inadequate
use of background correction by AAS laboratories. Results from analysis by
Laboratory XRF were, in general, negatively biased relative to the results from the
extraction methods. The quadratic tendency of the recovery data (excluding
SRMs) suggested that an inadequate number of standards were provided for
calibration. In addition, no standardized procedures for sample preparation or
analysis were provided.
A pairwise comparison of method means declared the most differences in
method means for the MW/AAS and laboratory XRF methods. The MW/AAS
produced the highest mean for all six samples, whereas the laboratory XRF
method produced the lowest mean for five of the six samples.
Laboratory XRF was the most repeatable of the methods, while HP/ICP
results were the least repeatable. MW/AAS, MW/ICP, and HP/AAS methods
produced results with similar repeatabilities. The MW/ICP method showed the
best reproducibility for five of the six samples.
The results indicate the MW/ICP method to be a method of choice for the
samples analyzed in the round-robin. This method gave good reproducibility
6-1
-------
(total system coefficient of variation <12%), and showed the least variable recovery
across concentrations.
6-2
-------
SECTION 7.0
RECOMMENDATIONS FOR FURTHER STUDY
The study was successful because it provided the following:
• a protocol for the preparation of Method Evaluation Materials for
lead-containing paint and dust,
• a means for validation of the protocol
- at targeted concentrations, and
- of acceptable homogeneity, and
• a means of comparing methods commonly used to analyze lead in
environmental samples.
A number of questions about the differences in analytical methods were
brought to light. Further studies are suggested to resolve questions that include
the differences observed in AAS and ICP results, and the apparent negative bias
observed for Laboratory XRF results.
An investigation of the apparent enhancement of AAS measurements
relative to ICP may include the following:
• comparison of results for paint and dust reference materials by AAS
analysis with and without background correction,
• comparison of ICP results of extractant solutions that are either:
- diluted below concentrations specified in this round-robin (1
10 jug/mL), or
- spiked with a solution of an internal standard, and
• development of a method for minimization of the
enhancement/suppression effects.
7-1
-------
The question of the apparent negative bias observed for Laboratory XRF results
may be examined by the following:
• an investigation of matrix interference,
• the use of standardized protocols,
• the use of standardized materials for instrumental calibrations, and
• the use of internal standards.
7-2
-------
SECTION 8.0
REFERENCES
1. Housing and Community Development Act of 1992 (P.L. 102-550), Title X,
Residential Lead-based Paint Hazard Reduction Act of 1992.
2. U. S. Environmental Protection Agency, Office of Pollution Prevention and
Toxics. EPA national lead laboratory accreditation program. Laboratory
quality system requirements. Revision 1.0. U.S. Environmental Protection
Agency, Washington, D.C., 1993.
3. Task Group on Methods and Standards of the Federal Interagency Lead-
Based Paint Task Force. Laboratory accreditation program guidelines:
Measurement of lead in paint, dust, and soil. EPA 747-R-92-001, U.S.
Environmental Protection Agency, Washington, DC, 1992.
4. Williams, E. E., Grohse, P. M.} Neefus, J. D., and Gutknecht, W. F. A report
on the lead reference materials workshop. EPA 747-R-93-008, U.S.
Environmental Protection Agency, Washington, DC, 1991.
5. Greifer, B., Maienthal, E.J., Rains, T.C., and Rasberry, S.D. Development
of NBS standard reference material No. 1579 powdered lead-based paint.
National Bureau of Standards Special Publication 260-45. U. S.
Government Printing Office, Washington, DC, 1973.
6. Consumer Product Safety Act, "Ban of Lead-Containing Paint and Certain
Consumer Products Bearing Lead-Containing Paint, 15 U.S.C., 2057,2058,
March 1978.
7. Office of Public and Indian Housing, Department of Housing and Urban
Development. Lead based paint: Interim guidelines for hazard
identification and abatement in public and Indian housing. U.S.
Government Printing Office, Washington, DC, 1990.
8. Binstock, D. A., Hardison, D. L., White, J., Grohse, P. M. Evaluation of
atomic spectroscopic methods for determination of lead in paint, dust and
soil. In: Proceedings of the 1991 U.S. EPA/AWMA International
Symposium, Measurement of Toxic and Related Air Pollutants, Durham,
North Carolina, 1991,
9. Osborne, Fred and Assoc, 703 Whitley Ave., Clemmons, North Carolina.
8-1
-------
10. Binstock, D.A., Hardison, D.L., Grohse, P.M., and Gutknecht, W.F.
Standard operating procedures for lead in paint by hotplate- or microwave-
based acid digestion and atomic absorption or inductively coupled plasma
emission spectrometry. EPA 600/8-91/213, U.S. Environmental Protection
Agency, Research Triangle Park, North Carolina, 1991. Available from
NTIS, Springfield, Virginia, PB 92-114172.
11. Cross Beater Mill; Model SKI (Dietz), Serial No. 71475, Glen Mills, Inc.,
395 Allwood Road, Clifton, New Jersey.
12. Retsch Grinder, Model ZM1, Serial No. 33060, Oriden, Brinkman
Instruments Co., Westbury, NY. Also available as Ultra Centrifugal Mill,
Glen Mills, Inc., 395 Allwood Road, Clifton, New Jersey.
13. Turbula Blender, Model T2C, Serial No. 910880, Glen Mills, Inc., 395
Allwood Road, Clifton, NJ.
14. Neutron Products, Inc., 22301 Mt. Ephraim Road, Dickerson, Maryland.
15. Ro-Tap Generator; Model No. 5KH35JN3132T, Rotary Model No. L143, Lid
Model No. L45, Tapper Model No. L42, W.S. Tyler Ro-Tap, Fisher Scientific,
711 Forbes Avenue, Pittsburgh, Pennsylvania.
16. U.S. Environmental Protection Agency, Office of Health and Environmental
Assessment, Environmental Criteria and Assessment Office. Urban soil
lead abatement demonstration project. EPA/600/AS -93-001,Volumes 1-4.
U. S. Environmental Protection Agency, Research Triangle Park, NC, 1993.
17. International Organization for Standardization. Guide 35. Certification of
reference materials - General and statistical principles. International
Organization for Standardization, Geneva, Switzerland, 1989.
18. Boyer, D.M,, and Hillman, B.C. Standard operating procedures for energy
dispersive X-ray fluorescence analysis of lead in urban soil and dust audit
samples. U. S. Environmental Protection Agency, Las Vegas, Nevada, 1992.
19. Williams, E. E., Binstock, D. A., Estes, E.D., Neefus, J.D., Myers, L. E., and
Gutknecht, W. F. Preparation and evaluation of lead-containing paint and
dust method evaluation materials. In: Proceedings of the Symposium on
Lead Poisoning in Children: Exposure, Abatement and Program Issues,
American Chemical Society, Washington, DC, 1992.
8-2
-------
20. Kleinbaum, D.G., and Kupper, L.L. Applied Regression Analysis and Other
Multivariate Methods, Duxbury Press, North Scituate, MA, 1978.
21. Steiner, E.H. Planning and analysis of results of collaborative tests. In:
Youden, N.J., and Steiner, E.H. Statistical Manual of the Association of
Official Analytical Chemists, E.H., AOAC, Arlington, Virginia, 1975.
22. Miller, R. Simultaneous Statistical Interference. Springer Verlag, 1981; also
SAS User's Guide: Statistics, Version 6, Gary, NC 1987.
23. U.S. Fish and Wildlife Service, Method 201, Digestion of Animal Tissue.
24. Bao-hou, Li, Zhong-quan, Yu, and Kai, Han. Determination of Si, Al, Ca,
Mg, Fe, Ti, Mn, Cu, Ci and Ni in Vanadium-Titanium-Iron Ore by
Microwave Oven Digestion, ICP, AA and Chemical Analysis Methods.
Institute of Chemical Industry and Metallurgy. The Academy of Sciences of
China. Beijing, China. 1988.
8-3
-------
Appendix A
Statistical Approach
-------
Appendix A-l
Statistical Design of the Round-Robin
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
ATMOSPHERIC RESEARCH AND EXPOSURE ASSESSMENT LABORATORY
RESEARCH TRIANGLE PARK
NORTH CAROLINA 27711
February 4, 1992
MEMORANDUM
SUBJECT:
FROM:
TO:
Review of RTI's Design of Round Robin for
Lead in Paint and Dust
Jack Suggs
EDAB/EERD/
,AL (MD-77B)
Sharon Harper
According to RTI's design, paint and dust solutions will be prepared at two
different levels each—medium and high concentrations. Each' of the two levels
will be further split into two replicates. Each laboratory will receive 4
aliquots (2 reps x 2 levels) of paint solution and 4 aliquots of dust solution.
In addition, each lab will receive a third "level" or standard reference material
(SRM) of paint and of dust. The SRM's will not be replicated. Each laboratory
will analyze in duplicate each of the aliquots using their method of analysis
(XRF, AA, ICP). Methods AA and ICP also involve two extraction procedures:
microwave and hotplate. The purpose of this study is to:
1. Evaluate the homogeneity of the paint and dust solutions prepared
according to RTI's protocol
2. Estimate and compare between-lab differences
3. Estimate and compare within-lab differences
4. Compare methods of analysis.
A possible solution to these problems may be obtained through the use of
linear models and the analysis of variance. Dust and paint data are treated
separately but with the same model.
To avoid overparameterization in the models, think of the method/extraction-
combinations plus XRF as five different methods:
e.g. XRF AA/M AA/H ICP/M ICP/H
For each of these methods and each level of solution (including SRM), a
separate analysis of variance can be performed.
e.g. Paint, Method = AA/H, Level = high
ANOVA TABLE
Source
Rep
Labs
Rep x Labs
Duplicates
BE
1
7
7
16
EMS
aD
2am.2
2 a,^2
16oR2
4oL2
MS
MS,
MS,
MS3
MS4
MS,/MS3
MS2/MS3
MS3/MS4
-------
All sources of variation are assumed random. The expected mean square
(EMS) column shows the components of variation. These components may be
estimated by equating the EMS to the mean square (MS) column. Also, certain F-
ratios may be calculated (as suggested by the EMS) to test hypotheses
corresponding to objectives in the design.
For example:
1) F = MS,/MSj
is used to test the hypothesis that the variation between replicate aliquots is
zero. This is a test of the homogeneity of the solution,
2) F = MSj/MSj
is used to test the hypothesis that the difference between laboratories is not
significant,
and 3) F = MS3/MS4
is used to test that the difference between replicate aliquots does not differ
(in analysis) from lab-to-lab. This is compared to the variation between
duplicates within each lab represented by aD2.
For the SRM solution, the analysis of variance is less complicated since
there are no replicate aliquots.
e.g. Source DF EMS
Labs 7 aQ2 + 2oL
Duplicates 8 oD2
In addition to tests of hypotheses, estimates of variance components a^
(between reps), aL2 (between labs), oc2 (between dups) can be obtained along with
estimates of reproducibility standard deviations, and repeatability standard
deviations defined by ASTM as
Reprod = (oL2 + aD2)'A
Repeat = OD.
By definition, two measurements made at a given level of solution using a
given method by two different labs should not differ by more than 2.77 (Reprod)
but 1 time in 20 due to chance alone. The value 2.77 (Repeat) applies to two
measurements (duplicates) in the same lab.
These estimates can be obtained along with average values X for each
solution level Med, High, SRM and each Method/extraction to produce the following
table.
-------
Paint Dust
XRF AA/M AA/H ICP/M ICP/H XRF AA/M AA/H ICP/M ICP/H
"X =
crL2 =
Med aD2 = Same
Reprod =
Repeat =
X =
High aD2 = Same
Reprod =
Repeat =
X =
aR2 = not retrievable
2 _
SRM aD2 = Same
Reprod =
Repeat =
Using the entries in the table, the between-lab variances (aL2) , within-lab
variances (oD2), and between-rep variances (aR2) can be examined for homogeneity
across methods and levels. Averages can also be compared. If homogeneity is a
fair assumption, the data may be pooled into a more complex analysis. This is
not really necessary, but a layout of the sources of variation and degrees of
freedom for the full model helps to identify the many different comparisons.
e.g., Paint, medium and high levels (no SRM)
Source df
Homogeneity Levels 1
of solution Reps 1
Levels x Reps 1
Methods 4
Method Level x Methods 4
comparison Rep x Method 4
Level x Rep x Method 4
Between-lab Lab (Method) =35
variation Level x Lab (Method) =35
Level x Rep x Lab (Method) =35
Within-lab Duplicates 152
Total 287
-------
Another possible analysis of the data would involve only the AA and ICP
methods. These methods each have two extraction procedures. The layout of the
analysis of variance for paint (or dust) at two prepared levels (no SRM) would
look something like the following.
Source df
Homogeneity Level 1
of solution Rep 1
Level x Rep 1
Method (i.e. AA vs ICP) 1
Extract, (i.e. Micro vs Hot) 1
Meth x Extract 1
Level x Meth 1
Method Level x Meth x Extract 1
comparisons Rep x Method 1
Rep x Extract 1
Rep x Meth x Extract 1
Level x Rep x Meth 1
Level x Rep x Extract 1
Level x Rep x Meth x Extract 1
Lab (Meth x Extract) =29
Between-lab Level x Lab (Meth x Extract) 29
variation Rep x Lab (Meth x Extract) 29
Level x Rep x Lab (Meth x Extract) 29
Within-lab Duplicates 100
Total 231
Most of the interactions, especially the higher-order interactions, will
probably be zero. In any case, the 3- and 4-way interactions are difficult to
interpret and should probably be combined to provide denominators for F-tests of
single and 2-way interactions.
These are simply suggestions for analysis based on the proposed design.
I'm sure there are other possible approaches. There are two ways that more
balance could be achieved: 1) more labs for AA/microwave, 2) replicate aliquots-
for the SRMs. I know that this last suggestion is prohibited by cost, but it
would provide a comparison of the homogeneity within an SRM as compared to the
prepared materials.
cc: W. J. Mitchell
-------
Appendix A-2
ISO Guide 35
-------
GUIDE 35
Certification of reference
materials —
General and statistical principles
S*cond edition 1S
-------
Contents r*o« Foreword
for*won* a ISO Uhe International OrBanixation foe Standardisation) <• •
worldwide federation of national ttandard* bodie* tISO
Introduction 1 member bodieal. The work of preparing International Stan-
, *'* « normally carried out through ISO technical com-
1 5cop* ' m^'W-^chmwTiwoooVintwwtKJinifyoi^tlofwtweha
* iwfio^vwv , '«<^i«gwwtvoffnataria1t 8 ISO guito «• immxtod asttntiafly for internal uMm ISO com-
- - .......... m««»« or in »omt «$«• for th* fluidanoa of mwnoaf oodi**
f G«oe*.lpnnopl«ofc«rtrficaw>n 11 w»»« dMBnfl with manaa that wouW not ncxmaHy ba tha «T
jact of an Intamational Standard.
7 Certification by a definitiv* method 12
• /•—*: • ^ w w_ ISO GuW*K wa* dflwn UP by tht ISO Comminei on
• Certification by tntariabotatory testing 14 reference material* tREMCOl and was submitted directty to
. _ _ . "SO Councfl for acceptance. Thb second edition cancels and
t Certification based on a metrolocical approach 21 replaces the first edition (ISO Guide 36 : 19851, to which a new
clause 9 ha* been added.
Annex A: Bibliography 33
e iso law
Al ri0hti r*MfV«d. No part of thitpuMcat>OAm*yb«r»produo*dorutib«din»nylormorbyany
means, alactiooic or m*ch«nic*l. incMkng photocopying and microfilm, without pcnraswoit in
writing from ttw pwbr«her.
Intametfonal Organiution fw Stsndarifttition
CAM pofttto W • CH-1211 G«i*v« 2) • Switzerland
Printed in Swftnrtond
-------
ISO GUIDE 36 :198»(E)
Certification of reference materials —
General and statistical principles
Introduction
Th« Committee on reference materials (REMCO) is concerned
with guidelines for the preparation, certification and use of
reference materials. This Guide is intended to describe the
genera) and statistical principles for the certification of
reference materials.
Various sections of this Guide were prepared by different
delegates to REMCO. The project was co-ordinated with
representative* of ISO/TC 69. Applications of ttttatic*!
meinodt.
Acknowledgment is given to J. 0. Co* tBSI. UK) for prep-
aration of the section on the role of reference materials in
measurement systems (clause 31. Much, of clauses 4, 5 and 8 is
based on material contained in three previously published
sources:
al CALL J.P. ere/. The role of standard reference materials
in measurement systems, /V5S Monoynpfi 148, Washing-
ton, DC, National Bureau of Standards. 1975 (especially
Chapter III, by H. H. Kul;
b) UftiAXO. G. A. and GHAVATT. C. C. The role of reference
materials and reference methods in chemical analysts. CrH.
Rev. in An*/. Chtm. • 1977: 381;
c) MAftSCHAi. A. M»ttri»u* tf» ntftf/vnct. Bureau National
de Metrologie. Uboratoire National d'Essais. Paris.
K. R. Eberhardt (ANSI. USA) prepared clause 7 on the use of a
definitive method to certify reference materials. R. Sutamo
and H. Steger (SCC. Canada) prepared clause 8 on the use of
an interlaboratory testing programme to certify reference
materials. H. Merchandise (Community Bureau of Reference,
Commission of the European Communities) prepared clause 9
on a metrological approach to certification, included for the
first time in the second edition of this Guide. G. Uriano (ANSI.
USA) served as editor of the Guide.
Special acknowledgement is given to members of ISO/TC G9/
SC 6 and its Secretary K. Patrick (OIN. Germany, F.R.I, for
their co-operation in preparing those sections of the document
concerned with the statistical analysis of data. In particular the
many contributions of Prof. P. T. WHrich (OIN. Germany. F.R.I
and Or. T. Mryaiu (JISC. Japan) of ISO/TC G9/SC 8 to the
review and editing of the Guide are gntefuffy acknowledged.
Earlier Guides'"1 prepared by REMCO have dealt with the
following aspects of reference matenats:
al mention of reference materials in International Stan-
dards:
b) terms and definitions used in connection with reference
materials;
cl the contents of certificates of reference materials.
The purpose of this Guide is to provide a bask introduction to
concepts and practical aspects related to the certification of
reference materials. ISO Guide 33ia' more fully addresses con-
cepts and practical aspects related to the use of reference
materials.
1 Scope
According to the definition given in 2.1. reference materials
(RMs) may be used in diverse measurement rote* connected
with instrument calibration, method assessment and assign-
ment of property value*. The purpose of clause 3 is to discuss
these measurement roles and to show how traceebaitv" of
measurement may be secured by use of RMs. thus yielding
worldwide compatibility of measurement.
Just as certified reference materials (CRMs) are to be preferred
over other classes of RMs in citations in International Stan-
dards'". so also are CRMs to be preferred over other classes of
RMs in measurement science generaffy. given that CRMs
needed for a particular type of measurement exist. Assistance
in locating the source! s» of supply of CRMs for various tech-
nical fields is afforded by ISO's &r*cton/ of certified nfennc*
It wX be evident that the quality of a measurement based on
use of a CRM will depend in part on the effort and care ex-
pended by the certifying body on determining the property
1) An mtemetiona»v egted definition of "treeeeb*ty" in measurement science • given in reference 161:
vaceeMllty: The ptoperty of a result of e measurement whereby it can be related to appropriate eundarde. geoeretr ntemiuonei or national etan-
derds. through en unbroken chem of compeneone.
-------
ISO GUIDE 36 : 1«8 (E)
valuelat of the candidate CAM. Hence the proceea of cextrfl-
cation'*' should be carried out using wet-cherecterUed
measurement methodi that have high accuracy e* well •*
precision end provide property values traceable to fundamental
unit* of moesuremsnt Furthermore, the methods should yield
value* wrth uncertainties that ere appropriate to th* expected
end use of the CRM. Clauses 4 and 6 deal with two of th* moat
impotent technical consideration! in the certification of
RMs - measurement uncertainties «nd material homogeneity.
Cleuae 6 provide* general principle* for RM certification.
Two commonly uMd general approach** to assuring tech-
nically valid RM certification at* discussed in dauM* 7 and t.
Oeue* 7 describe* the UM of a single method of the highest
accuracy ILe. sometimes referred to ae a "definitive" or
"abeoiute" method) and utusBy employed by • single labora-
tory for RM certification. Clause 8 describes the use of an inter-
laboratory testing approach to RM certification, which might
involve more than one method.
tht World Meefth Orotnlationl and certain technological *U» (tor
eismple rubber Mocks f»r tf» determiAetien of ifr»en.»neei or steel
plstet for the e*erminsMn of hardness). ft « recognaed «*t tfw
dtfMtlon of "iwference mettneT owen MOV* eou*d invoN* sn ov*rl«o
«ith ih* term "mtwrisl rneseure" M tftfirtetf in the Inttmitwttt
*xrt« m«i«o*»i mty be e(%*r»ct»rvr*ndu*ty msnvfecturetf objects which sr«
etoo certified MMduOy. Numerous RMs have properties which,
because they cannot be correltted ¥>*»» an e*tabfi*!>«d chemical struc-
ture or for other r*Mone. csnnot be nweeured in me** or emount of
autwtenoe unfa or determined by •KKtty defined physical or chemical
rmeeuroment metfxxH. Such RM* include certain bio4o0«*l RMt (for
eK«mpte a vaccine to which art intemetionsl unit he* been eMigned by
3 The) role of refaranca mat aria Is in
maaiuramant sciafica
Metrology it the field of knowtodge concerned with measure-
ment. Metrology or measurement science " includes el aspects
both theoretical and practical with reference to measurements.
whatever their level of accuracy, and in whatever fields of
science or technology they occur »l. Thi* clause describe* the
role of reference materials in quantitative measurements.
3.1 The rolt of rafaranca materials In the storage
and transfer of information or property values
By definition 12.11, a reference material has one or more proper-
ties, the values of which are wei established by measurement.
Once the property value's) of a particular RM have been estab-
ished, they are "stored" by the RM (up to its expiration date)
and are transferred when the RM itself is conveyed from one
place to another. To the extent that the property value of an
RM can be determined with a weR-dafined uncertainty, that
properly value can be used as a reference value for intercom-
psrison or transfer purposes. Hence RM* aid in measurement
uansfsr, in time and apace, aimiar to measuring instruments*
and material measures M).
A general scheme for constructing a hierarchical measurement
synem is •ustrtted in section 6.5 of the VocaAotary o/i«$*/
MttrohgyW. The interlinking of various levels and stations
within a measurement systsrn via "reference standards" may,
in principle, be effected by erther measuring instruments or
material measure* or RMs.
An RM must be suitable for the exacting role it performs in stor-
ing and transferring information on measured property values.
The fctowing technical criteria (legal or commercial criteria
11 "Measurement science" It therefore synonymous with "metrotogy" tccording to (he intamstionaf definition of the letter term*; M ihoutd be
noted, however, that current uwge generaty restricts the term "rnet/cteoy » phyifcat measurements tt high eccuracy. The term "inetiefooy" «.
however, being (ncrsesingty used In tie context of chemical, engineering, biological and medical mseiursmsriB.
at So
•ana are not reedty mowbte Cby reason of silt. meet, frsgfiiy. inetabe^orc«*lJ. JnwMehceeetfwmeesurintfrnuMbe
brought to *w Instrument to effect the rneaeuremant transfer. But el ftMt and material measures are readty movabts and thus can be liken to the
nginetru
-------
ISO GUIDE 36:1988 (E)
may be relevant alsol apply to the fitness for purpoM of RMt in
general:
•I the RM itself and the property valuelsl embodied in it
should be stable for en acceptable time-span, under realistic
condition* of storage, transport and use;
b) the RM should be sufficientry homogeneous that the
property valued) measured on one (xtion of the batch
should apply to any other portion of th« batch within
acc«ptabla limits of uncertainty; in cases of inhomogenerty
of the large batch, it may b* necessary to certify each unit
from tha batch separately;
c» the property vaiue(s) of tha RM should hava b««n
established with a precision and an accuracy sufficient to
tha and use(s) of tha RM;
d> clear documentation concerting tha RM and its estab-
lished property velvets) should ba available. Prafarably tha
property value(s) should hava baan certified, so tha
documantation should than includa a certificate, preparad
m accordance with ISO Guide 31 »l.
The word "accuracy" was advisedly us«d in c) to indicata that
wt>enever possible, tha maasuramant of a given property value
should have been made by * method having negligible sys-
tematic error or bin relative to end-use requirements (or where
the result has been corrected for« known bias) and by means
of measuring instruments or material measures which are
traceable to national measurement standards. Subsequent use
of an RM with traceable property values ensures that trace-
ability is propagated to tha user. Since most national measure-
ment standards are themselves hermonited internationally, it
foBows that measurement standards in one country should be
compatible with similar measurements in another country. In
many cases. CRMs are appropriate for the mtercomparisons of
national measurement standards.
3.2 The role) of reference) material* in the
International System of units (Sll
3.2.1 Dependence of the SI base units on substances
and materials
The majority of measurements made in the worid today are
within the framework of the International System of units "I. In
its present form. SI recognizes seven base units, namely the
units of length (metre, symbol ml. mass (kilogram, kg), time
(second, s). electric current (ampere. Ay. thermodynamic
temperature (kervin, K), amount of substance (mole, mol) and
luminous intensity (candeta. cd). The definitions17* of these
base units mention the following substances: krypton-86" (for
defining the metre). piatinum-iridium (for fabricating tha proto-
type kilogram), caesium-133 (for defining the second), water
(for defining the kervin) and carbon-12 (for defining the mole).
Opinions differ as to whether the substances named fall under
the definition of reference material (2.11. The use of these
substances in basic metrology is consistent with tha use of
reference materials in other types of measurement applications.
Certainly such materials have a specie) status as defined
substances on which the St is baaed. The dependency strictry
applies to definition of the unit, since realization of the units
may involve other substances /materials. This is especielry true
in regard to the realization of the mo*i«l end tha k8oorem.
3.2.2 The realization of derived SI units with the aid of
reference materials
From tha seven base units en unlimited number of derived units
of the SI are obtainable by combining base unru as products
and/or quotients. For example, * derived unit of mass concen-
tration is defined as kg-m -» and the derived unit of pressure
(given tha special name pascal, symbol Pal is defined as
m-'-kg-s*1. Formally speaking, the derived units ultimately
depend on the substances on which the base units themselves
depend (see 3.2.1). In practice, the derived units are often
realized not from base units but from RMs with accepted
property values. Thus a variety of substances/materials may be
involved in the realization of derived units (examples 1 and 2
below) or even of base unrts (examples 3 and 4 below).
fxempfe I: The SI unit of dynamic viscosity, the pascal second
(Pa-s • m-'-kg-s-'l may be realized"1 by taking the value
for a wel purified sample of water as 0.001002 Pa-s at 20 'C.
£rampfe 2: The SI unit of molar heat capacity, the joule per
mole-ketvin (J-mof-'-K-i - kg-ml-s-'-mol-'-K-') may
be realized1*01 by taking the value for purified a-alumina as
79.01 J-mol-'-IC-»et25»C.
£rampr* 3: The SI unit of amount of substance, the mole, may
be realized by taking 0.069 72 kg of highly purified gallium
metal I"*.
fxfempfe 4: The SI unit of temperature, the ketvin. may be
realized at any temperature 7, 1273.1$ K < 7", < 903.89 Kl
from measurements of the resistance of a highly pure platinum
wire at 7",. at the triple point of purified water, at the freezing
point of purified on end at the freezing potnt of purified zinc.
coupled wrth use of a specified mathematical relationtui. The
word "thermodynamic" has been deliberately omitted here to
avoid controversy over whether thermodynamic temperatures
are, or are not, the same as International Practical Tempera-
tures of 1968: the intention of the International Committee for
Weights and Measures was to match the two sorts of tempera-
ture exactly, within the framework of knowledge avaajbla
during 1968-197$.
3.2.3 Connection of analytical chemistry to tha
International System of unrts
It w* be noted that purified (often caned "pure") chemical
substances were cited in each of the examples 1 to 4 (3.2.2).
Tha measurement of degree of purity, or more generally of the
chemical composition of materials, is within the realm of
analytical chemistry. In addition to the dependence of SI on
chemical substances, the dependence of aneryticai chemistry
on SI is worthy of examination. Presently, most analytical
II Recently, the General Conference on Weights and Meaeure* redefined the metre et the dietence trsveeed by igM n e vecuum during
1 '299 792 458 of t
-------
ISO GUIDE38:1969 (E)
chemists employ whs wfthln the SI (aft bee* unto except the
candela end also many derived unit*} In their meeeurements.
However, compositional enafysia depend* on en additional
concept, namely th*t pur* chemlcsl specie* axfet to which the
chemical composition* ol other substance* and materiel* «•
referred, by Evoking the law* of chemical change and ttoichio-
metry.
From on* or mot* pur* chemical specie*, Considered to b*
prvnary meeeurement standard*, it it feasible to construct
measurement Werarchiae lor analytical ctomietn/ timter to
Ihoee ueed In physical me*surernentl*>. Example* of euch
measurement standards an;
•1 the electron, to which other species c*n b« connected
by electrochemical ftnaryiitlW;
b) carbon-12. to which other specie* can in principt* b«
connected by mess spsctrometry. (Uourt'f Uw mett$un>-
or volumetric mtMurvments with low-density
•tc.;
th« UM of rcftnmc* nvtwitit for mtatfon of phyticd proptr-
ti«». Tr* folowing IUPAC Commcnion 1.4 puDficttkxv in A^»
trtAppfvd Ct»mittry w* conotrmd with th« c»rtifie«iOf» «nd
UM of r*f«r«no« m«t«n«l« for phytical
cl • NaMy purifwd •Umoni or compound, to which othtr
•DOOM c«n b* conrMcttd by vlecvochemical, Qf«vim«tnc.
thiirrnt/ic, toKtromttric mtthodt, ttc.
TlM "othw •f*c*4" cfttd in tft*M •xjmptw wi in mcny ca*M
b* mod « RlWte. Many tubtuncM ctn Si this rott of !m«r-
madwriM bctwMn prinwy and working anaryticti atarxterd*
using tt* drvarsity oi tachniqoaa and chamical raactiona that an
•nan/at may amptoy. Tha concept of tncMbility appfiea to
•narytical chamittiv a« much at it doaa to other branchea of
moaaurament tdence. Tha quality of the result of a chemical
enatytia w* be enhanced if the returi'i uaceabiGty can be
d*anV ataled in tanrn of the traceabffity of the inttruments.
mneiial measure* and RMt amployad. In mott caaee, the
TracoeNttY wl alto (Inpenrt on the vthiai nf thn rotaThra atnmif
meeeea {formerly called "atomic weights") used in the calcv
lationa; the source of these should b* recorded by the anatyet
(for example (1!J>.
34L4 The role of reference matarlala In realizing unhs
ewtslde of the tl
Where the component! of a metaursmem system (for example
the Imperial system) can be related exactly to the correspond-
ing components of the SI, it la unnecessary to have indepen-
dent maena for realizing the non-Si measurement system.
Where the quantities cannot be related to those of the SI, then
independent realization of the non-Si units i* in prifxapts
necessary. In practice, however, ftw such system* remain in
us* and thus are mostly historical curiosities.
3.3 UM of reference m«»rt*t«
REMCO Intends to publish « separata guide cowering general
end statistical principles for the use of reference materiel*.
There am very few pubfiahed documents that address general
problem* eseoclsted wfth the use of refeiance materiel*. The
reader I* referred to the document* end reeomrnendation*
pubOehed by IUPAC CommiesJon 1.4 on Phyvrco-chemkal
Reference Matertala and Standard*, which deal primarty with
Vclwtn«, dtt* of pwbllctthin
an4 p*fl« number
40
40
40
46
4f
4t
60
W
C2
fa
ra
1974 : 390
1974 : 461
1974 : 463
1978:1
1979 : 241
1977 : 001
1971 : 1 477
1971:1486
1980:2393
1981 : 1 847
1961 : 1 863
proptrty
Enthalpy
Optical rotation
Optical refraction
Density
Relative molecular mas*
Absorbanc* and wavelength
Reflectance
Potsntiometric ion activrtie*
Vocoerry
Permittivity
Thermal conductivity
4 Measurement uncertainty
In discussing measurement uncertainties, the tsrms
"precision", "systematic error or bias", and "accuracy" are
usuafly used. The meanings of these terms are not rigidly fixed,
but depend to a large extent on the interpretation end use of
th* data I M. ««.
4.1 An Illustrative example}
If two aquaty trained operator*. A and 8, each make four
repficationt of a measurement on a uniform material each day
for 4 days on one instrument, and 4 day* again on a similar
instrument, the results. 16 sets of four measurements, may
look fika those in figure 1. What can b* seen from this plot ?
a) the spreads among each eat of four values are com*
parable, perhaps sltghtfy smeller for instrument 2 man in-
strument t;
b) there appears to be more variability betwsen dely
results than within sets of defly results, particutariy for in-
strument 1;
c) operator S gives lower results then operator A;
dl instrument 1 grve* tower results than instrument 2.
Figure 1 is constructed for the purpose of demonstration, and
actual measurements could be better or worse than shown.
However, this plot does show soms four type* of factors that
contributed to the totst variabXty of these measurements:
1) factors acting within days;
2) factors acting between days;
3) factors due to instrument systems;
4) factor* due to operators.
Appropriate technique* are avalabic for the separata esti-
mation of ih* effects of theee four factors and standard dsvi-
ationa could be computed corteapondmg to each of them.
I to waver, the Imhed number of opei sluts end instruments
prevent* the computation of standard deviation* as refiabry for
-------
ISO GUIDE 36:1969 IE)
I
i I 1 ! I
1 1
1 1
I 1
I i
•".
.«
%•
0
• .
"."•! •'
• •
"
•\
Oayll D«y2 | Day]
Instrument 1
.••
•
.••
0«y4
• "
•'
.•'•
0«yS
• *•
„••
1 1
^
• «
•
1
1
1
i
i
"" " "
•/•
•%
Operator A •
Operator 8 • |
i i
D«y* J Day 7 ( Dtyl
Instrument 2
Figure 1 — An example of results of meesurement*
by two operator* using two instrument* on eight
different day*
factor* 3) and 4) a* for factors 1) and 2). The time and work in-
volved certainly impose Smhs on any efforts to do so.
The faBurs to allow for factors relating to instruments and
operators is one of the main causes for the unreasonable dif-
ference* usually encountered in interiaboratory, or round-
robin, types of tests I "l. Because instruments vary from time to
time end operator* change, the result from a laboratory at a
given time represents only one of the many results that could
be obtained, and the variability caused by these two source*
must be considered as part of the precision of the laboratory.
The standard deviation computed without regard to these
effect* would underestimate the true variabKty.
tf. by the proper use of standards and reference method*!171.
these two source* of errors were eliminated, the standard devi-
ation computed from the 16 means of sets of four measure-
ment* would be the proper measure of precision. Presumably
the grand mean of the 16 mean value* would be reported.
The mean of many value* is more stable than individual
measurements. When extraneous source* of variation, such as
instrument and operator effects, are eliminated, the relation-
ship between the standard deviation of individual measure-
ments and the standard deviation of the mean of n such
measurements can be expressed a*
... (II
In other words, the standard deviation of the mean i* smaller
than the standard deviation of individual measurements by a
factor of 1/V7T. One important provision must hotd for this
relationship to be true. i.e. that the it measurements are in-
dependent of each other. "Independence" can be defined in a
probability seme, but for present purpose*, measurement*
may be considered independent If they show no trend or pat-
tern. This is certainly not true in figure 1, and to say that the
standard deviation of the mean of at 64 value* i* 1/t
(. 1/V«) of the standard deviation of individual measure-
ments would seriously underestimate its true variability. More-
over, the relationship in equation 111 is expressed in terms of
the true vstue of the standard deviation, a. which is ueuafty not
known. As the computed standard deviation, t. is itserf an esti-
mate of 9 from the set of measured vslues. the standard devi-
ation of the mean in equation (f I is onry approximated when * is
used in place of o.
The use of the standard deviation computed from dairy aver-
ages rtihar than individual values is preferred because the
former property reflects a component of variability between
day*, or over time, which i* usually present in precision
measurement.
4.2 Some basic statistical concept*
The basic information available on the meesurement error* is
summarized by:
al the number of independent determinations or the
number from which a mean was computed and reported;
b) an estimate of the standard deviation, t, defined by
I'/J
t «
where * measurement result* *re denoted by *,. xj
xt. and their mean i*
i.i
From •( and b) several useful derived statistic* can b* com-
puted:
c( standard deviation of the mean of n measurements
t
This is sometimes ceded the standard error of the mean to
differentiate it jrom the standard deviation of individual
determinations.
NOTE - As * becomes torga. the value of iLr,) become* wry
smM. snowing that the «verege of a large number of metsure-
mtntt »ppro*crtM a con*t»nt value n which « ututly the objective
al the mMSurwnent procedure.
dl confidence interval for the mean (normal distribution!.
Each time n measurement* are made, a value of the average
of the measurement* i* reported. These averages wfll differ
from time to time within certain limits. Assuming a normal
distribution, one interval of the type i t i can b* con-
structed "•' such that the interval from i - &\o i + t wfll
-------
ISO GUIDE 36:1969 (El
be fairly cartain to Include the value of n deeirad. The Intar.
vat is computed by:
- I
whara / ia a tabular value of the Student diatribution, and
dapanda on tha confidenca (aval and the dagraas of freedom
for*;
a) 2-aigma (or 2s). 3-sigma (or 3j) fim/ts. These Smks
describe tha distribution of maaauramant arror. If a
maaauramant is made by tha user of a CRM having tha
aama precision (i.a. aama 9) aa that obtained by tha certify-
ing laboratory, his measurements should fafl (with prob-
ability approximately 0.95 to 0.997) within thaaa limits whan
a is w*l-established. Otherwise there ia evidence of system-
atic difference.
4.3 Inatrumant and oparator arror*
Instrument and oparator typaa of arrora have not vat bean
treated. An ideal situation would be to efiminete them from tha
maaauramant process, or to uaa mora instruments and mora
operatora and then estimate standard deviations associated
with thsae aourcaa. Whan neither of tha above ia feasible or
practical, tha least that can be dona ia to use two instruments
and/or oparatora. H tha confidence intervals for tha mean
results of tha two instruments do not overlap, than there ia
good evidence of instrument drfferenca.
lifting his axparianoa and judgement, a measurement scientist
may arrive at reasonable bounda for thaaa types of errors. If tha
bound ia not computed from measurement data, than its val-
idhy cannot ba supported by statistical analysis. In such cases,
thaaa bounds ara "guesttimatea" and tha only recourse is to
treat them aa fimfts to systematic arrora.
Tha detection of differences and the separation of the total
variably into ha identifiable components can ba fadrteted
through careful pfenning and statistical design of the experi-
ment.
4.4 Olffarancaa among maaauramant mathoda
Each meaauramant method purports to measure tha deeirad
property of a material, but seldom does a method measure the
property directly. In moat cases tha method actually measures
aoma other property that ia related to tha property by theory,
practice, or tradition, and than converted to tha value of the
desired property through these relationahipa. Discrepancies
among results of different measurement mathoda are common.
even for measurements leading to tha determination of fun-
damental physical constants "•'.
In tha preparation of a CRM, usually two or mora measurement
mathuda ara employed for eech property measured. If these
mathoda ara wall established by virtue of paat experience, tha
reeutta yielded by these mathoda uauafty agree to within the
uncertainty aaaignad to each method.
In a few eases theee differences are ao large that the reeurta
cannot ba raconcead, and theee results are than reported
aaparatary for each individual method. Tha RM ia either not car-
trfied or certified on a method-dependent baaia. A hiatorica)
axampla of thia type of reporting ia N8S CRM 1091, Stalntaaa
Steal. Tha nitrogen content w»a maaaurad by vacuum fuiion
and praaaura bomb-diatiKation. and gave raaulta of Ml and
946mg/kg. with standard daviationa of 3 and ZOmg/kg.
reapactivary. Qaarty one or both mathoda have a ayttematie
error that ia large compared to tha variaWrty of materiel or tha
maaauramant uncertainty. A report of tha average of tha two
mathoda would ba highly mialaading.
Measurement accuracy in its abeotute aanaa ia never raaTaad. In
practice, certified values of aoma reference materials ara
defined by using a referee method or assigning a value by a
wen-defined procedure ao that at least tha aama benchmark wi
ba used by everyone in tha field. Tha importance of reference
methods to supplement tha uaa of thaaa maaauramant stan-
dards ia also being emphaauediW. A good example is the
reference method for blood haemoglobin and tha value
aaaignad as a benchmark to tha reference material issued by the
International Committee for Standardization in Hematology
(ICSH) <»•»'!.
4.5 Uncartalntlaa of cardflad valuaa
The uncertainty of a CRM value ia usuafty made up of several
componanta. aoma supported by data and aoma not:
a) a atatiatical tolerance interval grving bounds to material
inhomogenerty based on data and atatiatical computations;
bl a confidence interval for tha maan giving bounda to
measurement error baaed on data and atatiatical compw
tationa;
c) componanta of measurement uncertainty due to vari-
ation among laboratoriaa and/or operators and measure-
ment methods;
d) a combination (addition of abeotute valuaa or tha
square root cf tha sum of tha aquaraa) of estimated bounda
to "known" aourcaa of poaaibla ayatamatic error based on
experience and judgement (in other words, thara ara no
data, or an insufficient number of data, to make a atatiatical
calculation).
Tha word "known" ia quoted abova to contrast with sys-
tematic arrora that are "unknown" or unsuspected. Thaaa un-
suspected arrora could occur in a number of ways - a compo-
nent in tha physical system, a minor flaw in tha theoretical eon-
sideration, or tha rounding error in a computation. Aa more
homogeneous materials become evaiaMe. and mom precise
maaauramant methods ara developed, these types of arrora wi
ba detected by design or by chance and hopetufty wl be
eliminated. Improved accuracy in tha measurement of a prop-
erty ia basically an expensive iterative process and unwarranted
demand for accuracy could maan tha waste of resources.
4,t Statamanta of uncertainty on CRM
cartlflcataa
A variety of statements of uncertainty can ba found In paat and
currant cartiflcatae issued for CflMs around tha world. Some of
thaaa atatements ara wal formuiatad and eupported by data,
-------
ISO GUIDE 36:19»(E)
others are not; torn* of thee* statements eoouin • wealth of
information that i* useful to exacting user*, but overwhelming
to other*; some statement* are oversimplrfied with • resulting
IOM of information. Because the originator of • CRM h«« to
keep al dsss«s of UMK« in mind. the UM of • single, form of
statement i« not usually possible. The mention is that al that*
statement* are unambiguous, meaningful, and contain an the
infonnation that i» relevant for potential user*.
Some commonly used statements, takan from existing car-
tificates. ara listed in 4.6.1 to 4.6.4.
* f .1 Exampla 1: 96 % confidence limit* for the mean
Rubidium chloride
Absolut* abundance ratio 2.693 t 0.008
"The indicated uncertainties are overall Emits of error bated on
95 % confidence limits for the mean and allowance* for the
effect* of known sources of possible systematic error."
Because the isotoptc ratio is a constant for a given batch of
material and is not subject to error* of material inhornogeneity.
the 96 % confidence limits for the mean rater to measurement
error only. This is computed from
as described in equation (21.
The effects of known source* of possible systematic error are
discussed in detail in "Absolute isotopic abundance ratio and
atomic weight of terrestial rubidium" TO.
4.0.2 Example 2: 2-slgma or )-algma Omits
Glass Filters for Molecular Absorption Spectrometry
Absorbance 0.5000 10.002 5
"This uncertainty i* the sum of the random error of t 0.1 %
relative Oa Hmhl and of estimated biases which are t 0.4 %
relative."
Each glass filter was individually calibrated, and the standard
deviation refers to measurement error, including the cleanliness
of the surface. As these glass filters win be used time after time.
a murtipte of the standard deviation is a proper measure of
variability.
4.6.3 Example 3: Uncertainty expressed In significant
dlgfte
AISI 4340 Steel
Element Mass Fraction
Carbon 3.»2*10-J
Manganese 6.6 x 10-»
According to the explanation given in the text: "The value
listed is not expected to deviate from the true value by more
than 11 in the last significant flgurt reported; for a subscript
figure, the deviation is not expected to be more than t 8."
Thus, the mass fraction of carton, expressed as a percentage.
i* between 0.377 and 0.367; and that for manganese is between
0.66 and 0.67. These uncertainties include material inhomogerv
eity. measurement imprecision, and possible bias between
laboratories and implicit rounding, because these value* are
"... the present best estimate of the true value based on the
results of a co-operative intertaboratory analytical programme."
When 20 to 30 elements ara to be certified for one material, this
method gives a concise and convenient summary of the results.
As these limits are expressed in units of S and 10, some infor-
mation is unavoidably lost for som* of the elements. However.
when the certified value is used, it is important to use alt of the
digits given including the subscripts. The uncertainty stated on
this certificate depend* heavily on the use of chemical judge-
ment.
4.6.4 Example 4 : Standard deviation, and number of
determinations
Method
Vacuum
fusion
i
I
*
Neutron
activation
*
t
Inert oas
fusion
i
t
M
Oxygen in ferrous matala
CRM
A
(Ingot iron)
484
14
216
492
2ft
4*7
13
12
CRM
(StsMess steel :
AISI 431)
131
•
286
CRM
C
(Vacuum
melted steeO
21
2
KB
132 ! 2ft
7 i «
« 1 S
129
ft
11
2)
S
20
where
i is the mean oxygen value;
* is the standard deviation of an individual determination;
M is the number of determinations.
NOTf — The standard deviation includes error due both to the «we-
tiHen of the analytical method and to possible heterogeneity of in*
msteriel analysed.
One criticism againet this mode of presentation is that the user
wl have to compute the uncertainty based'on his own
understanding of the relationship*.
-------
ISO GUIDE 36:1869 (E)
6 Homogeneity of material*
Most RM* are *ubjected to • preparation procedure which
uttimetery indudM *ubdM*ion into uMbtt units. A subset of in*
dividual unto from the batch ie choMn for measurement
according to • stetlsticaBy vafid sampling plan. A measurement
uncertainty it derived taking into account malarial inhomogen-
aity M wefl a* othar factor* (*ee clause 4). Other type* of RM
a/a prepared M individual artifact* and the certification is based
on aaparata measurement of each unit rather than on statistical
sampling of tha comptata batch. The second approach is useful
whan tha RM can ba msasured non-destructrvefy.
S.1 Malarial*
RM* praparad a* aolutJon* or port compound* era expected to
ba homoganaou* on physical (thermodynamicl ground*. Tha
object of tha test for homoganaity i* mainly to detect any im-
purities, intarfaranc** or irragularitiaa.
Malarial* auch a* mixad powdar*. ora*. alloys, ate. ara hetero-
ganaou* in composition by nature. RMs praparod from such
malarial* must therefore ba tested to assess tha dagraa of
homoganaity.
6.2 Concapt of homogeneity
In thaory, • matarial« perfectfy homoganaou* with respect to a
given charactaristic rf thara « no diffaranca batw«an tha valua
of thi* charactaristic from ona part (unit) to anothar. However,
in practice a matarial is accepted to ba homoganaou* with
raapact to a givan characteristic M a diffaranca between tha
valua of thi* charactaristic from ona part (or unit) to anothar
cannot ba detected experimentally. The practical concept of
homogeneity tharafora embodies both a specificity to tha
characteristic and a parameter of measurement (usualy tha
standard deviation) of tha measurement method used, in-
cluding the defined sample size of tha test portion.
6.2.1 Characteristic of Interest
A material may ba sufficientfy homogeneous with respect to
the characteristic of interest to ba useful as an RM even though
it la inhomoganeou* with respect to othar characteristic*, pro-
vided that thi* inhomogeneity exert* no detectable influence on
the accuracy and precision of the commonly used methods of
determination for tha charactaristic of interest.
•JL2 Homogeneity measurement method
The degree of homogeneity that a material must have for use as
an RM is oommensurm with tha praciaion attainable by tha
beet available method* for the determination of tha charac-
teristic for which the RM i* intended. Therefore, the greater the
precision of the measurement method, the higher i* tha re-
quired degree of homogeneity of the materiel.
The precision attainable by tha homogeneity measurement
method vane* with both tha characteristic measured and it*
value for the RM. An RM intended for more than one charac-
teristic i* described by a corresponding number of *t*tements
of homogeneity, each of which ahouid ba traceable to an
experimema»y determined praciaion. Tha magnitude of tha pre-
cieion can very widely-
In many cases, the praciaion attainable by a measurement
method i* affected by the aUe of the teat portion taken from the
RM. The degree of homogeneity of an RM is therefore defined
for a given tact portion sue.
•.2.3 Practice
Weatty. an RM should be ch*r*ctaruad with respect to tha
degree of homogeneity for each characteristic of Interest. For
RM* intended for • relatively large number of characteristics,
the assessment of the degree of homogeneity for al charac-
teristics is both economically and physically burdensome, and
in some case* unfeasible. In practice therefore, the degree of
homogeneity of auch RM* i* attested only for selected charac-
taristic*. It ia recommended that these characteristic* ba ap-
propriately selected on tha basis of established chemical or
physics* relationship*; for example, an intarelemant con-
comitance in the mineral phases of an RM make* reasonable
the assumption that tha RM also has an acceptable degree of
homogeneity for tha non-selected element*.
5.3 Experimental design
6.3.1 Objectives
For reference materials that are expected to be homogeneous
on physical ground*, the main purpose of homogeneity testing
is to detect unexpected problems. Some examples are differen-
tial contamination during the fine) packaging into individual
units, or incomplete dissolution or equiibration of an enatyte «
a solvent (which could lead to staadiry changing concentration*
from the first vial filled to the last). A statistical trend anetysi*
would be helpful in the latter case. If the materiel i* produced in
more than one batch, it i* necessary to test the equality of the
batches lor to certify the batches separately!.
When the natura of a reference matarial lead* one to expect
some inhomogeneity, the goaf of tha testing programme i* not
simply detection of inhomogenerty. but rather tha animation of
it* magnitude. Thi* may require a more extensive testing pro-
gramme than is required for detection.
Inhomogeneity can manifest rtseff in at least two way* :
•) different subsamples of an RM unit may differ on the
property of interest;
b) there may be differences between units of the RM.
Differences among subsamples can usually be reduced or con-
troRed to an accepubt^owlevei by maXing tha sue of the sub-
sample sufficients/ large. Often a study to determine tha ap-
propriate aubsampie size i* conducted before the certification
experiments ara begun. Differences which exist between in-
dividual unit* of tha candidate RM must be reflected in the
uncertainty statement on the certificate.
In statistics* term*, the experimental design must satisfy the
following objectives:
1) to detect whether the wrthin-unh (short-range) varia-
tion ia statisticety significant In comparison with the known
variation of the measurement method;
-------
ISO GUIDE 36:1989 IE)
2) to detect whether the between-unhs (long-range) vari-
•lion it statistically significant in comparison with the
wrthin-unrt variation;
3) to conclude whether a detected statistical significance
for one either soM or pondered form, and finely ground oxide
m*i«nel* that «>e intandad for use •« reference materials in X-ray
•m«s
-------
ISO GUIDE »: 1969 IE)
"rtrbstical toieranos interval" can be ueed. To Muatraie thfc
concept, auppoee • eolution • prepared end packaged into
1 000 empoutee, ol which 30 ere measured for aome property.
For (hit aumpl«, the tolerance limit concept "•' tutee aaaen-
tiety »h«t baaed on the measured valuee of the 30 ampoule*
etmoet al of the 1 000 empovlea wi» not differ from tht average
of the 30 ampoule* by more than th« constructed (imit. In stat-
istical terme. it would reed: "The toleranc* interval
(mean t AI ia constructed such that rt w* cover n toast 95 %
of the Deputation with probability 0,99". "
Thie statement dOM not guarantee that th« tol*nnc« intwvtl
w» indud* an of tht unpoui**. h Myt that 99 % of tn« lima
tha tolaranca intarvat wfl includa at teatt 96 % of th« a/rv
pouctad ? First, tha maan (aquation (3)]
and atandard d«viation |*quation 441) from tha 30 ampoutea ara
oomputad :
X m
.-1
whar*
T|.
.. 13)
... 14)
. jr. ara tha ma«»ur*d valua», with
n m 30;
jt ia an a«timaia of tha mean, u, of tha 1 000 ampoulaa;
t ia ao *atimata of tha maasura of tha dispersion, o.
among theaa ampoulaa.
Tha v*K>«a i and i contain practicalty al tha information
avaiabla on the 1 000 wnpoukM and can be used to calculate
tha totoranc* intarval i t A .
Tha value of A ia computed as a multiple of s. i.a. 6 * *V».
The value of *j depends on three parameters :
al tha number, *. of samples measured (30);
bl the proportion, p, of the total population to t>a covered
(0,961;
e) tha prototMlrty (aval. 1 - a. apecrfwd (0,99).
A table of factoca for two-aided tc4eranca Nmrw for norm*
diatributiooa grvaa tha va/ua for fj aa 2.841 foi n • 30;
1 - a • 0,99; and p - 0,95. Tablaa of trvaae factors are given
in ISO 3207» and in many atanda/d sutirical tartt"*'.
Tha term "two-sided" means that wa are interested in both
over and under limits from tha average Tha term "normal d«S-
tricHjtion" rater* to the diatribution of all tha values of intaraat
and is • symmetrical, bed shaped distribution usuelh/ en-
countered in precision measurement wort.
Rgura 2 is a histogram of tha ratios of tha emission rate ol
'"Ci. in a '"Ct nwcJear fue< bum-up reference material, to a
radium reference standard. A frequency curve of a normal dis-
tribution can be fined to these data. Thar* were 98 ampoules of
'"Cs involved; each ampoule was measured in Aprt. Septem-
ber, and November. 1972. By averaging tha three measure-
ment*, the measurement error was considerably smaller than
tha difference of masses of act/va. sotutiona among these am-
poule*, and the ptet in figure 2 show* essentially the in-
homogenerty of the rross of solution in the ampoules.
30 r-
CRM
12
?nfl
%
\
I
^M
I
tour
^ 1 iM
»o»-'~z:z:*^£ircs»;S
O«5«O«k*O*°*O*O*D
a o" o' o" o o' o" o' o' o o o"
'"Cs
RRS3D
Figure 2 — Histogram of tha frequency (number of
ampoultsl versus the ratio of tha actrvrty of 1>7Ca
standards to a radium reference* standard IRRS20)
II Th« rutement m tn* only for a popuc*iion al infant tat; how»v*f, the contemn tor a population of IMt« w« • ne0lg<>«a wtw* firwte tat •
laraa
21 ISO 3207. StrbncJ nr*rpr*t>(xyi or* dttt - Dtt+rmnttnn of t mteaicaf foMrenc* •ttarva/.
-------
ISO GUIDE 36:1989 (E)
ComoutaiKXt
O< COftSOMUS
v»iue« *nd
unc«rt»mtv
M
-------
ISO GUIDE 36:1969 (E)
lni«rt«b
-------
ISO GUIDE 36:1969 (E)
1.3.1 Statistical outfters
A tingle result or an entire set of results is suspected to be •
statistical outlier rf its deviation either in accuracy or precision
from others in the set or other sets, respectively. » greater than
can be justified by statistical fluctuations penmen! to • given
frequency distribution. Therefore, the effectiveness for the
detection of outliers depends on the validity of the assumption
of the frequency distribution. Tha lt*t for oujjjam should ba tha
statistician's prerogative. For an interieboauxy programme
outlying ttitut may ba conferred on individual results, ratults
for individual unit* or tha antira cat of results from a laboratory.
8.4 Statistical analysis
8.4.1 Two-tttga o»tt«d daslgn
This modal is usad whan tha results of an intartaboratory pro-
gramma ara usad to confinn tha homoganaity as wafl as to
charactarua tha matarial. Tha experimental schema is iflus-
tratad-schematicanY in figure S a). The results can be expressed
by tha aquation
where
. «5J
is tha * th result of sample unit j reported by
laboratory i;
it « the grand mean;
a, is tha error due to laboratory i;
f¥ is tha error due to tha /th sample unit in laboratory /;
tt is the measurement error.
AN these parameters can ba estimated aimunaneously by the
analysis of veriar.ce, (ANOVA) method (see 8.4.3.1) if there ere
sufficient resorts of equal replication tihe same number of
replicate determinationa from each unit and the same number
of units par laboratory) after outliers have been eiciuded. If this
ANOVA requirement cannot be met because of the number of
outliers and/or missing results, the significance of the
between-units finhomogeneityl variance can be tested by the
simple procedure for unbalanced data given « 8.4.3.2.
Theoretical details and additional methods for balanced and un-
balanced ANOVA are given in standard textbooks. «" »
8.4.3.1 Computation of two-stage ANOVA
x,^ is the *th result of sample unit / reported by labora-
tory *;
p is the number of participating laboratories;
q is the number of units per laboratory:
unit.
is the number of replicate determmations per sample
JT m —
8.4.2 One-stage netted design
This model is used when tha material is accepted to ba
homogeneous by the organizers. The experimental scheme is il-
lustrated schematicaBy in figure S bl. Equation (5) can then be
simplified to
X* - it + a, * e*
8.4.3 Analysis of two-stage nested design
Parameters to be estimated ara
— it. the Ofo-xl mean (which is used as the consensus
value);
—
-------
ISO GUIDE 36:1969 (El
and each mMn square it given ••
MS, - SS,//,
MS, - SS,//,
MS, - SS,//}
These results should 64 tabulated (s«« table D
Table 1 - ANOVA table
Source
tabors-
tons*
orwu
mem
e«w
turn of
square*
SS,
SS,
SS,
Degree* ef
freedom
,-.
p(q - 1)
"'-"
Meen • Expectation
•o,uar* • of m««n *qu*re
MS, «w » "*2 * «"«*
MS, «w » »Wy
MS, • 0W
Each parameter is estimated by the following equations, where
the circumflex denotes the estimate :
i mi
SI - (MS, - MSj)/ - 11 y —
PI*
'i-«/a
-------
ISO GUIDE 36:1969 (E)
T«vo-ti«g«
I) AH Mmple units »r* diHeftnt Kow«v«f. m c«Ch UtXVdory Ihcy »'«
1. 2.
On* «t»8« netted d«*ign
RM
CMrKtertliC
1 '2
0«le>min«iion
Figure S •— Enp«rim«nt«l sch«m« for an intcrlaborctory ppogramm*
-------
ISO GUIDE 36:1989.(E)
T.bU 2 - ANOVA tebto
Source
B«rw»*n g«vts
M**tur«m«m •nor
Sum
of squares
$$,
SS,
O+frees
ef freedom
h
/,
MM*
square
MS,
MS,
The twt (or statistical significance of the between-units
(inhomogeneity) vsritnce «
MSj/MSj
which Should be compared with th« critical value of the
/""•distribution for degrees of freedom
{£(8,- 1)}and{13>(/- U}.
t.4.4 Analyst! of one-stage netted design
For cas«* where the material it considered to be homogeneous.
i.e. that all units are identical, all results reported by a labora-
tory are considered at replicates.
xv a the yth result reported by laboratory /;
p is the number of participating laboratories;
n, is the number of results reported by laboratory i.
I.I
The variance of the consensus value, x is simply estimated by
—J-r- V tf, -
ptp-1)£rf '
with degrees of freedom (p - II.
The confidence interval for the consensus value (mean of
means) is the interval from A to B where
- 1|(tr))
l/a
- «/2
-------
ISO GUIDE 36:1969 (E)
own eetimates of accuracy ar* correct and that no error ha*
escaped it* attention.
Inter comparisons add confidence to the uncertainty computed
by th« metrology laboratories indMdueny. Sometimes they use
safety factors which are not necessary; sometimes they under-
estvnate their own uncertainties.
Tht present practice by which each metrology laboratory evalu-
ates the uncertainty of a particular measurement on in own i*
inherently dangerous. It is not possible for a laboratory alone to
avoid ai error* in af circumeiancee. in particular for derived
uniti. Intercomparisons detect errors that were not taken into
account and situations where el parameters influencing the
measurements are not sufficiently wet controlled.
There is unfortunately no general requirement in metrology that
uncertainty statements be based on appropriate iniercompari-
son*. Certifying a reference material on the besis of results of
one single metrology laboratory may therefore imply a risk
which should not be overlooked.
When the certification of a physical property or quantity is
undertaken, it is therefore important to have an intercom-
parison between the major metrology laboratories followed by
a fui discussion of the results with all participants to resolve
any possible discrepancy. If the primary metrology laboratories
are not themselves involved in the measurement, complete
treoaebifity of the participating laboratories to the respective
national laboratories must be established before starting.
The participants must then compare their measurements and
discuss afl the possible errors responsible for discrepancies and
efiminet* them while remaining independent. This is described
in more detal for chemical measurements in 9.3.2.
If more than one method is possible, and if these methods ap-
pear equaflY valid, it is important to compare them. However, it
is useful to remember that the method with the shortest trece-
eb&ty route or. in other words, with the most direct connection
to the fundamental units, has a higher probability of being more
accurate.
At the Emit, there can of course exist situations where one
single laboratory, having compared its method with al possible
others and having efirnnated most causes of errors, is able to
refine its method to reduce the uncertainty while taking con-
siderable precautions to avoid any accidental source of errors.
Some measurement problems in the field of physical properties
can be briefly ilustrated by thermal conductivity of insulation
and refractory materials. Until some years ago. laboratories
were not able to carry out such measurements with appropriate
accuracy although the ca&bratipn of the instrumentation ap-
peared satisfactory. The guarded hot-plate used for the
measurement was constructed and operated in accordance
with existing national and international standards. The agree-
ment appeared satisfactory for simple technical applications.
However, in most laboratories there was a systematic error.
Heat loeaes occurred above room temperature because the
guard ring was not sufficient. Any reference material certified
on that basis would have a wrong traceabffity. The method and
equipment w«re therefore modified until the heat tosses
became negligible.
The accurate determination of thermal conductivity of refrac-
tory materials it very difficult by the direct method using the
guarded hot-plate apparatus mainly because of the heat losses
and experimental difficulties. Methods such as the hot-wire
method or the flash method do not present such difficullie*.
but their traceabitty is not easy to establish and therefore these
methods are not the best for certificatioft. However, th* results
of these methods are important as a verification of the results
of the guarded hot-plate.
9.3 Certification of • chemical composition
1.3.1. Traceabilftv
In the field of analytical chemistry, there is no established
measurement system organized as in the field of metrology.
with primary and calibration laboratories, and measurement
standards available for circulation. The concept of accuracy is
hence more difficult to reach and the tric«b*tv is more dif-
ficult to realize.
In chemistry, the calibrations in the usual sense are not the
major source of difficulties although the task of th* chemist is
heavier than that of the metrologist. He needs not only physical
standards of mass, volume, temperature, etc..but also stan-
dards of alt chemical species he has to determine : elements.
organic compounds, etc. Each one of these chemical standards
has an uncertainty (e.g. impurities) which is sometimes under-
estimated.
The biggest problem is however the traceabUty of the overall
analytical process : the uaceabffity chain it broken every time
the sample is physically or chemically modified in the analytical
process.
As the variety of sample processing procedures is Urge, it is not
possible to discuss the treceability in general. The following
paragraphs are to be considered only as examples.
$.3.1.1 Sample weighing
The first step of the analytical process is the weighing of the
sample. This does not pose problems of traceabity if the
balance is periodicals- calibrated. Human errors are not ex-
cluded but they are not frequent.
S.3.1.2 Sample treatment
Whenever the sample is disserved or submitted to simlar treat-
ment, the traceabilhy chain is broken and any uncertainty
•valuation should take this into account. To establish trace-
ability for that part of the measurement procedure, a laboratory
must demonstrate the relationship between th* initial sample
and the solution prepared from it. The main questions to be
answered are. was the sample totally dissolved, what were the
losses, were there contaminations? If the analysis is to deter-
mine not one element but a compound, was the compound
changed during the dissolution step? In the case of organic
compounds, the efficiency of extraction is one of the main
cause* of drrncurtie*.
-------
ISO GUIDE 36:1969 (E)
Ttbto S - Tre.ce elements In milk
Vibes *» nanogremt per grsm
Element
Cd
H«
rt>
Cu
Plret
Inter eomperiaon
Irenoa of reeultsi
0.4 to 4 500
0.6 to 42
8* to 6 500
47tf to * 2S7
Certification
campaign
(range of results!
1 to 6,6
0.73 to 1,27
12.4 to 112.5
475 to 700
1 Certified
1
i af
1.0
! 104.5
1 • «*
Table 4 - Results of analyses of olrve-tre« leave*
Element
Cd
r*
Hfl
Cu
Zn
Mr.
IffTtreeutte
V9'9
O.OM»o 6.864
17.6 to 33.3
0.006 to 0.708
O.S to 131.*
12.3 to 31.6
0.4 to 4.6
Retto
1X3
1.»
140
264
2.6
11.5
1M1 resuNe
P0'«
0.054 3 to 0.121
20J to 26.4
0.247 to 0.336
434 to 50.1
14.5 W 17.7
$1 to61.6
Ratio
2.2
1.3
S
i ;-J,
1.2
T«W« S - 0«t«rmln«tlon of poftlcld.t In powd.rod milk iplk.d with certain compound.
Compound
HCH
a-HCH
7- HCH
DOE
op'DOT
P-HCH
4-HEPO
0*k>rin
ppOOT
Results
mg/kg
0.001 to 0,22
0.00* to 0.60
0.001 14 to 0.16
0.0043 to 0.47
0.003 to 0.24
0.01 to 0.13
0.001 to 0.13
0.01 to 0.10*
0.005 to 0.38
r
Ratle
220
67
IS*
109
60
13
130
10
72
Quantities sdded
mg/kg
0.26
0.11
0.20
0.54
0.06
0.12
0.10
— solution VMtmant*.
— •mx* incluo>d in tht cafibration curv*.
- mnching the calibration to th« product to analyM
matrix affactt, intarfwanoM;
- a aocond round of analyM* with tha aama labontona*
but possibly with a malarial of slightty drNarant compo-
sition;
- discussion;
- further rounds of analysas at nacassary.
Tha procadura dascrfcad oftan laads to raiecting soms
mathod(s) or to abandoning aomt laboratoriaa which cannot
improv* thatr parformanca. At tha and of this long procadura.
ona has a sat of tachnicany consistant rasufts for which ona
calculatas tha maan valua. and hs S5 % confkfanca intarval
(adoptad as uncertainty). Examplaa of auccassrva stags* art
grvan in figura* 6 and 7. Statistics ara usad for no othar purpoaa
than for verifying that tha conditions ara futftllad to calculata a
96 % confidence intarval.
Tha statistics for tha calculation ara the same as shown in ISO
Guide 331 a i.
When the rasufts are not consistent, one must conclude that
tha technical work is not terminated and thst certification is not
possible.
h is to be noted that for trace elements or for the cartificstion of
impurity levels, the distribution of results can be tog-normal.
Tha confidence internal can be non-symmetrical.
NOTES
1 Th* rrwthodUl u*»d to certify S r*f*r«nc« m»t*ritl K» »om»tifnei
very different from tr* methods u*»d in routine praetic* U.Q. to certify
oortisol in *erum one has to use GCMS. whil* n pttctxa ttie
commonry uMd method is redio-immurtostMyl. In ttteia cs*M it is
important to verify that tha RM is suitable for KM with the roubne
In figure *. it should be noted diet only the GCMS iwuhs war* in-
tended for certification. The other methods were uMd to verify the
MjrujMrry of the KM.
-------
ISO GUIDE 36:1969 (E|
If. after sample treatment. the »oMion it subject to further
manipulations (pr econcentration. precipitation, etc.) each «tap
complicate* tha traceabitity rout* end adds n«w possibilitie* of
losses or contaminations which mutt be investigated.
h is wef known that torn* of th« parameters listed
nxxa on tha maun than on tha clamant or compound to ba
determined.
f .3.1.3 Final determination
Tha third ttap in an analytical process is tha final datarmination.
Apart from gravimatrv. trtrimetry, and coutometry. moat
methods, for axampia *pactromatry and atomic abaorption. ara
indiract. Tha rotrumantation used for these mea*ur«manta
provide* a signal which mutt ba correiatad with tha concen-
tration of tha substance of interest in tha unknown sample.
That corraiation a established by means of • calibration curve.
H«»a there ara two groups of problems to consider:
- is any error introduced in producing the calibration
curve and what a the accuracy ?
— is it correct to use that particular calibration curve ?
H we suppose that tha calibration can ba dona by means of
solutions, then the most important parameters to take into
account ara
— tha accuracy of the measurements (mass, volume!
made for the preparation of the solution;
- the purity of the elements or substances, the stoi-
chiometry of the compounds, etc.;
— the purity of the water or solvent.
Errors due to the calibration curve are not rare even in good
laboratories.
However, as pointed out in 9.3.1.4 even larger errors are due to
the fact that users sometimes produce calibration curves which
•re not appropriate to tha solutions they have to analyse; these
•re named matrix effects, interferences, etc.
In metrotogicel terms, this could ba expressed M follows : each
laboratory produces for itserf a measurement scale which is not
fully appropriate to the measurements to be made, and each
one produces a different measurement scale.
t.3.1.4 Matrix effect
The response of a particular element to a measurement process
(•.9. spectrometry. atomic absorption) may depend on the
solution (viscosity, conductivity, ionic strength) or on the ions
present in it (interferences).
Besides • large number of such cases in inorganic analyse*.
severe matrix effects ara found in clinical chemistry, where
some methods designed to analyse a serum can be wrong for
aqueous solutions. For such methods the calibration should ba
done with human serum; if this is not possible, the validity of
any other matrix should be demonstrated.
In this respect tha term "cetibrsnt" used by biochemists c*n be
misleading. Similarly, in inorganic chemmry. a calfersnon sol-
ution should simulate very cioseiy the solution to be analysed.
t.3 J Certification work
The task of any laboratory participating in an exercise to certify
a new reference material includes the study of the parameter*
mentioned in 9.3.1. A fufl study requires the comparison of dif-
ferent method* of sample treatment and different method* o«
determination. This c*n. however, ba bast done colectivery in
order to have the coftaborstion of experienced specialist* in
each method. In addition, for each method there should be
more than one laboratory in order to avoid systematic error*
due to laboratory effects or operator affect*. It can ba pointed
out that error* (e.g. those due to contaminationsl can only be
detected by comparison of results from different laboratories.
The need tor scrutinizing carefufty tha results of the different
participants can be Jiustrsted by the examples given in tables 3
and 4. which are rather typical of trace element analysis at very
low levels. The laboratories c-ften find values which are too high
because they ai produce some contamination. If one too
quickly adopted the mean value of they results, one would
have a systematic error by excess, and a reference matenal
totally unreliable from the point of view of traceabaity. This ax-
plains why the procedure proposed to approach accuracy «
composed of severs! steps in which the participants discuss si
sources of error* in si parts of the analytical procedure and
then try to reduce them. Analyses are then repeated (possibh/
not on exactly the same samples) and the results are discussed
again a* many times as necessary to reach sufficient con-
vergence.
The need for severe! laboratories also exists in the case of so-
called "definitive" methods like IOMS. For one particular deter-
mination there may be more than one "definitive" method, or
several variations of a definitive method; it is of course essential
to verify that they provide the same result and this is not
necessarily the case. H. after detailed comparison of the results
of several laboratories. H is not possible to identify errors, the
variation of results (between laboratories) represents the uncer-
tainty of the technique in the current state of the art. Working
with one single laboratory would perhaps toad to a smaller
spread of results but this would not necessarily represent the
real uncertainty.
To summarize, the certification work in accordance with the
approach proposed here would include tha following steps for a
homogeneous and stable material:
— examinationT with experienced laboratories, of the
most reliable (accurate) methodologies for the analysis of
the element or substance in the particular matrix con-
sidered;
— a first round of analyses;
- a detailed discussion of the results wrth al participants
to try to discover explanations of the differences; particular
attention is given to
— sample treatment.
— possible tosses, contaminations.
-------
ISO GUIDE 38:1969 (E)
iv» •
102'
|100-
a
8 98 .
96-
•
94-
•
L__
*
I
•
I ,
_ f
1 J T „_
. i r_J
2
Tint mi
™*~
KOXl
— «• w- «».
»P*n*on
1 2 3 4 5 6 7 8 9 10 11 12 13 U 1S 16
UbO/ttoriM
O
U
52
50
48
46
-, j 4- i 4 1_ J
Stcond int«toomp«n«on
1 2 3 4 5 67 8 9 10 tl 12 13 14 15 16 17
UboritoriM
• -
of th« flr« and ••cond lnt.rcomp.ri»on of •n.ly,,, of yrbon monoxld. In nhrog.n
-------
ISO GUIDE 38:1968(8
2 For II* preparation of I reference meter* in the Ottmedical f*W in
particular. Wood serum • tinted with tubtan? »genu or » lyopNi-
tied. H * ttien menna) to verify the appropriateness O< the reftrtnot
matcnai »tlw than treatments.
9.4 Certification of conventional properties)
In chemistry, biochemistry and other technologies, many prop-
erties ere defined only by a method, a test procedure or per-
ticular equipment. Examples are mechanical properties of
materials, activity of eruyme*. etc. The return of the*e
measurements or tests can be tub/act to great variability with
heavy economic consequences.
A* in any other measurement, the results depend on the way in
which the procedure is applied. Howtver. the procedura is not
always described in al necessary detail in the written stan-
dards and th« operator has no means of verifying rf the way ha
has interpreted and applied the procedure is correct. Hence the
need for the reference material.
The diegrtms in figure 10 show results of determination of the
activity of an enryma (x-glutamyttransferasel in an albumin
matrix with the same IFCC method. Laboratories shown on the
right-hand side had previous training with the method.
Laboretories on the left-hand side were high-level scientific
laboratories but with no previous experience in the method.
While the two upper diagrams in figure 10 relate to one
material, the bottom diagram concerns a different material.
Similarly, where a test depends on the use of a particular
machine or equipment it is possible, but extremely time-
consuming and expensive, to verify that the machine satisfies
al specifications. A simple way to by-pass this is to measure or
test a reference sample. If the results are satisfectory. h means
that the machine is in good condition end that therefore the
results can be considered traceable to the measurement scale
established by the relevant written standard.
Of course, the certification wort to establish reference
materials for such properties or measurement scales requires
the application of the same principles as explained before. The
measurements of these parameters, which may be mass,
volume, length or temperature, must themselves be accurate
and traceable and therefore may require extensive calibration.
Considerable effort is often necessary to investigate the in*
fluence of the various parameters of the procedures end of the
equipment on the measurement results. The verifications and
calibrations must be done independently in a few. if not
several, laboratories in order to avoid a uniform bias that would
appear as a good agreement and grve an illusion of accuracy.
9.9 Use of reference materials for eatabllshlnj
traceablllty
In J.3.1. a review was given of a number of parameters that a
laboratory should control and verify to ensure the traceebiKty of
the determinations. To do this in aH necessary details is very
hard work.
This can be considerably simplified by the use of a certified
reference material of established traceabilitv. The reference
materiel must be sufficiently simiar (in matrix) to the actual
sample to be analysed in order to include afl analytical problems
which might cause errors in the determinations. Of course, the
user should apply to the reference material the same anatyticat
procedure aa for his unknown sample.
When the laboratory using such a reference material finds only
a negligible difference with the certification value, this indicates
both that the result is accurate and that it is traceable to the
fundamental measurement scale. If the difference is not
acceptable, it indicates that the measurement procedura in-
cludes errors which must be identified and eliminated. It is sug-
gested that the most critical steps subject to errors are tr-«
sample treatment and the matching of the calibration.
Hence the role of the reference material is comparable to that of
the transfer standards used in metrology laboratories in in-
dustry, in that it allows working with a specified margin of
uncertainty.
The reference materials also make it possible to establish the
uncertainty of a measurement for analytical determinations or
technological testing.
The importance of a certified reference material goes therefore
beyond the definition of the reference material given in ISO
Guide 30I>I.
A reference materiel is used not only
- for calibration of an apparatus.
- for the verification of e measurement procedure.
but also
- for establishing traceabiity of the measurement results.
- for determining the uncertainty of these results.
FmeHy. one should not forget that the use of a reference
materiel does not eliminate completely the importance of
audits, the purpose of these being to verify that no mistake »
made in the use of the RM.
-------
ISO GUIDE 38:1969 IE)
1,4-
1.2-
1 -
0.8-
0,6-
«
0.4-
0.2-
|
Sn -
]
_ o •
I
1X2-
.£
a 1 -
lojB-
•5
| 0,6-
| 0.4^
0
,1
•
t
f»
I
I
I
I
I
]
[
I!
i
F»m mttf a
f
i
xrxxnwn
L«bof«lonM
'
I
•
•
I
']
I
i 1
1
I
J
I
I
bndiniffcc
I
imp«n«ort
Uborrton*t
1 -
OJ8-
0.6-
0.4-
0.2-
A —
"•j. k J— ,— .L
1
i iil
RM284 j f I
I I T ^ I
C*nific»iton o( RMt 2
I
— __.
i
1 I
M * as
Ub(X*l(Xi«l
Flgurt 7 - Evolution of r«tuh« in »ucc«M)vt lnt«rcomp«ri»oft« for th» d*t*rmlnatioA
of aftctoxln in milk powdor
-------
ISO GUIDE 36:1969 (E)
20-
o
1
S 10
A
14 t
Ij
*»VNAA
I'1,'
Photon
«Ctiv«tKX»
,'
CPA
,1
—
««OuC«ng fufon
'YI,!1'1.'
1
•
SSMS
1
*
123456
Figure I — Result* of Individual laborstoric* for oxyg«n/nick»l ratio
-------
ISO GUIDE 36:1989 (El
130 -
125-
120-
• 115-
t
i
| 110-
105-
"* 100-
95
90
At
IOMS
T1' l
1 l I
I
HPLC
I.
I
RIA
•»•
X
1
i
i
«
i
123456 ?• ABCOE 125
Ubortlontt
CKlificatiOA
f - D*t«rmination of eortlsot In r»con«Hut»d human Mrum
-------
ISO GUIDE 36:1989 IE)
250 H
200-
8 60-
100-
50
1 2 3 fc 5 6 7 8
L»bcx»too«*
Pint «u«tco">Mn*on
12 13 •; 15
2SO-
240-
| 230-
5
220-
210-
•»ftft-
I
i
!' I
1 I
4»
1 I
Ml.1
I
i
I . i
J^ S«cond Kit»feo^o*'i»ori
1 2 3 4 5 6 ? 8 9 10 11 12 13 U 15
Figur* 10 •) - Successive results lor the dstsrminstion of y-glutsmyltrsrtslsras* in slbumin
First and second int«rcomp«risons
-------
ISO GUIDE 36:1989 (E)
5 90-
"5
§
80-
4 ' U
. _.J.._ -J- —
I
T «
"1 "
I
123AS67891011 13
16
Figur* 10 b) - SUCC«MIV« results for th« d«t*rmiruition of r-0tuUmyHran«f«rat« in albumin
Final campaign
-------
Appendix B
Participating Laboratories
-------
LEGEND
(Appendix B)
OBS
METH
LAB
ANAL
EXTR
Reported Result
Method Number
1
2
3
4
5
Microwave/Atomic Absorption Spectrometry
Hotplate/Atomic Absorption Spectrometry
Microwave/Inductively Coupled Plasma Emission Spectrometry
Hotplate/Inductively Coupled Plasma Emission Spectrometry
Laboratory XRF
Code Assigned to Laboratory
Analytical Method
AA = Atomic Absorption Spectrometry
ICP = Inductively Coupled Plasma Emission Spectrometry
XRF = Laboratory XRF
Extraction Method
NIO = NIOSH Method 7082
EPA = EPA/AREAL Method
-------
List of Participating Laboratories by Method
OBS METH LAB ANAL EXTR
1 1 10 AA EPA
2 1 11 AA EPA
3 1 12 AA EPA
4 1 13 AA EPA
5 1 14 AA EPA
6 1 15 AA EPA
7 1 16 AA EPA
8 2 20 AA NIO
9 2 21 AA NIO
10 2 22 AA NIO
11 2 23 AA NIO
12 2 24 AA NIO
13 2 25 AA NIO
14 2 26 AA NIO
15 2 27 AA NIO
16 2 28 AA NIO
17 3 30 ICP EPA
18 3 31 ICP EPA
19 3 32 ICP EPA
20 3 33 ICP EPA
21 3 34 ICP EPA
22 3 35 ICP EPA
23 3 36 ICP EPA
24 3 37 ICP EPA
25 3 38 ICP EPA
26 4 40 ICP NIO
27 4 41 ICP NIO
28 4 42 ICP NIO
29 4 43 ICP NIO
30 4 44 ICP NIO
31 4 45 ICP NIO
32 4 46 ICP NIO
33 4 47 ICP NIO
34 4 48 ICP NIO
35 4 49 ICP NIO
36 5 50 XRF N/A
37 5 51 XRF N/A
38 5 52 XRF N/A
39 5 53 XRF N/A
40 5 54 XRF N/A
41 5 55 XRF N/A
42 5 56 XRF N/A
-------
LABORATORIES PARTICIPATING IN EPA/RTI ROUND-ROBIN
Alpha Analytical Labs
8 Walkup Drive
Westboro, MA 01581
Ms. Kathleen O'Brien
(508) 898-9220
American Medical Laboratories
11091 Main Street
Fairfax, VA 22030
(703) 802-6900
Azimuth, Inc.
9229 University Blvd.
Charleston, SC 29418
(803) 553-9456
Clayton Environmental Consultants
1252 Quarry Lane
Pleasanton, CA 94566
Mr. Ron Peters
(510) 426-2641
Clayton Environmental Consultants
22345 Roethel Drive
Novi,MI 48050
Ms. Ellen Coffman
(313) 344-1770
EOHSI
681 Frelinghuysen Road
P. O. Box 1179
Piscataway, NJ 08855
Dr. Clifford Weisel
(908) 932-0154
ESA Laboratories, Inc.
Industrial Hygiene Analytical Laboratory
43 Wiggins Avenue
Bedford, MA 01730
Mr. Paul Ullucci
(617) 275-0100
-------
Galson Technical Services
Industrial Hygiene Laboratory
6601 Kirkville Road
East Syracuse, NY 13057
Ms. Mary Withrow
(315) 432-0506
IT
5103 Old William Perm Hwy.
Export, PA 15632
Mr. Lyle Linsenbigler
(412) 731-8806
Keystone NBA Environmental Services
. 12242 S.W. Garden Place
Tigard,OR 97223
Mr. Thomas Nadermann
(503) 624-2773
Lawrence Livermore National Laboratory
Hazards Control Laboratory
7000 East Ave. P. O. Box 808 L-383
Livermore, CA 94550
Mr. Ray Szidom
(415) 423-7348
Liberty Mutual Insurance Company
Industrial Hygiene Laboratory
71 Frankland Road
Hopkinton, MA 07148
Mr. Ken Muzal
(503) 435-9061
Maryland Department of Health and Mental Hygiene
Division of Clinical Lab Services
Lead Lab, Room 509
201 W. Preston
Baltimore, MD 21201
Ms. Marilyn Gallagher
(410) 225-6184
Massachusetts State Laboratory Institute
Environmental Lead Laboratory/Room 311
305 South Street
Jamaica Plain, MA 02130
Ms. Phyllis Madigan
(617) 522-3700, Ext. 363
-------
Materials Analytical Services
2418 Blue Ridge Road, Suite 105
Raleigh, NC 27607
Mr. Don Porterfield
(919) 881-7708
Metro Denver Wastewater Reclamation
6450 York Street
Denver, CO 80229
Ms. Molly Lee Castleberry
(303) 289-5941
Midwest Research Institute
425 Volker Blvd.
Kansas City, MO 64110
Dr. John Stanley
(816) 753-7600, Ext. 160
National Loss Control Service Corporation
Environmental Sciences Laboratory
Rt. 22 and Kemper Ctr.
Long Grove, IL 60049
Ms. Joan A. Wronski
(800) 323-9585
NIOSH
Alice Hamilton Laboratories, R-8
4676 Columbia Parkway
Cincinnati, OH 45226
Mr. Peter Eller
(513) 841-4256
OWMC Laboratory
555 North Service Road
Burlington, Ontario L7L5H7
Mr. Joe Lesko
(416) 332-6711
Pennsylvania Department of Environmental Resources
712 Maryland Avenue
Erie, PA 16505
Mr. Gary Manczka
(814) 871-4291
-------
Research Triangle Institute
Analytical and Chemical Sciences
P. O. Box 12194
Research Triangle Park, NC 27709
Dr. Margaret Martin-Goldberg
(919) 541-7211
Roche Analytics Laboratory
P. O. Box 25249
Richmond, VA 23260
Ms. Sue Salkin
(800) 888-8061
SRI International
Physical and Analytical Chemistry Laboratory
333 Ravenswood Avenue
MenloPark, CA 94025-PS-177
Ms. Helen Parish
(415) 859-6177
Swanson Environmental
3150 Brookfield Road
Brookfield, WI 53045
Ms. Rosemary Dinen
(414) 783-6111
UEC Laboratories
4000 Tech Center Drive, MS#15
Monroeville, PA 15146
Mr. Mark Banister
(412) 825-2400
University of Cincinnati Medical Center
Department of Environmental Health
Kettering Laboratory Analytical Section
3223 Eden Ave., ML-56
Cincinnati, OH 45267-0056
Ms. Sandy Roda
(513) 558-1705
U.S. AEHA
Bldg. E, 2100
APGEA
HSHB-ML-R-M
Aberdeen Proving Ground, MD 21010
Mr. Dave Rosak
(410) 671-2619
-------
U.S. Department of Labor/Salt Lake Technical Center
P. O. Box 65200
Salt Lake City, UT 84165-0200
or
1781 S. 3rd West
Salt Lake City, UT 84115
Dr. Ray Abel
(810) 524-4270
U.S. Environmental Protection Agency, EMSL/Las Vegas
Environmental Programs Office
Lockheed ESC
1050 E. Flamingo Road
Suite 120
Las Vegas, NV 89119
Dr. Harold Vincent/Ms. Dawn Boyer
(702) 798-2129
U.S. Environmental Protection Agency, EMSL/Las Vegas
Methods Research Branch
944 E. Harmon Street
Las Vegas, NV 89119
Mr. Thomas Hinners
(702) 798-2140
U.S. Environmental Protection Agency
Region VII
25 Funston Road
Kansas City, Kansas 66115
Mr. Raymond Paus
(913) 551-5155
Wisconsin Occupational Health Laboratory
Department of Hygiene
979 Jonathon Drive
Madison, WI 53713
Mr. Terry Burke
(608) 263-6550
-------
Appendix C
Standard Operating Procedures
-------
Appendix C-l
AAS/ICP SOP - "Standard Operating
Procedures for Lead
in Paint by Hotplate- or Microwave-based
Acid Digestions and Atomic Absorption or
Inductively Coupled Plasma Emission
Spectrometry11
-------
RESEARCH TRIANGLE INSTITUTE
ZEI!
Center for Environmental Measurements and Quality Assurance
March 18,1992
Ms. Kathleen O'Brien
Alpha Analytical Labs
8 Walkup Drive
Westboro, MA 01581
Digestion Methods: NIOSH 7082 and EPA/AREAL
Analysis Method: ICP
Dear Ms. O'Brien:
Please find enclosed the RTI report, "Standard Operating Procedures for Lead in Paint
by Hotplate- or Microwave-based Acid Digestions and Atomic Absorption or Inductively
Coupled Plasma Emission Spectrometry." The report describes protocols to be followed for
digestion of paint and dust samples by the the NIOSH 7082 (Hotplate) and EPA/AREAL
(Microwave) methods for the EPA/RTI round robin. Paint and dust samples are being
shipped under separate cover.
Once again, thank you for your participation in the round robin.
Sincerely,
Emily Williams
Post Office Box 12194 Research Triangle Park, North Carolina 27709-2194
Telephone 919541-6914 Fax:919541-5929
-------
v>EPA
United States Atmospheric Research and Exposure Environmental Criteria
Environmental Protection Assessment Laboratory and Assessment Office September tMt
Agency Research Triangle Park, NC 27711 Research Triangle Park, NC 27711
Research and Development & Pesticide* and Toxic Substances
EPA 600/8-91/213
Standard Operating Procedures
for Lead in Paint by Hotplate- or
Microwave-based Acid Digestions
and Atomic Absorption or
Inductively Coupled Plasma
Emission Spectrometry
-------
PB92-114172
September 1991
Standard Operating Procedures for
Lead in Paint by Hotplate- or Microwave-based
Acid Digestions and Atomic Absorption or
inductively Coupled Plasma
Emission Spectrometry
Prepared by
D. A. Binstock
D. L. Harbison
P. M. Grohse
W. F. Gutknecht
Center fa Environmental Measurements and Quality Assurance
Research Triangle Institute
Research Triangle Park, North Carolina 27709-2194
EPA Contract No. 68-02-4550
RTI Project No. 91IM699-100
EPA Project Officers:
M. E. Beard
S. L Harper
D. J. von Lehmden
Atmospheric Research and Exposure Assessment Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
Prepared fa
Environmental Criteria and Assessment Office
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
TABLE OF CONTENTS
Sect I on Page No.
Disclaimer
1.0 Principle and Applicability 1
1.1 Scope and Application 1
1 .2 Summary of Method 1
2.0 Apparatus 5
2.1 Sampling 5
2.2 Instrumentation 5
3.0 Procedure 8
3.1 Sample Preparation 8
3.2 Sample Extraction 9
4.0 Analysis 1O
4.1 AAS-Cal Ibrat Ion 10
4.2 ICP - Calibration 11
4.3 Quality Control Prior to Sample Analysis .... 12
4.4 Quality Control During Sample Analysis 14
4.5 Sample Determination 16
5.0 Data Processing 17
5. 1 AAS 17
5.2 ICP 17
5.3 Calculation - Field Sample Concentration . . . . 18
6.0 References 18
-------
DISCLAIMER
The Information In this document has been funded wholly or In
part by the United States Environmental Protection Agency under EPA
Contract No. 68-O2-4550 to the Research Triangle Institute. It has
been subjected to the Agency's peer and administrative review, and
It has been approved for publication as an EPA document. Mention
of trade names or commercial products does not constitute endorse-
ment or recommendation for use.
-------
1.O PRINCIPLE AND APPLICABILITY
1.1 SCOPE AND APPLICATION
The adverse health effects resulting from exposure of young
children to environmental lead has received Increasing attention In
recent years. Studies have shown that chronic exposure even to low
levels of lead can result In Impairment of the central nervous
system, mental retardation and behavioral disorders. Although
young children are at the greatest risk, adults may suffer harmful
effects as welI.
The major sources of exposure to lead In housing units are
thought to be.pa Int. dust and soil. Food, water and airborne lead
are also potential sources but are considered to be minor avenues
of exposure. Though soil and dust serve as the principle vehicles
of direct exposure, lead-based paint Is receiving emphasis as the
source of lead in these two media and Is the focus of this
document.
Under Section 302 of the Lead-Based Paint Poisoning Prevention
Act, as amended, Public Housing Authorities (PHAs) are required, by
1994, to randomly Inspect all their housing projects for lead-based
paint1. Currently, the device most frequently used for testing In
housing Is the portable x-ray fluorescence (XRF) spectrometer,
which gives rapid results and Is non-destructive. However,
uncertainty In accuracy and precision of XRF measurements Is a
major problem, especially at and below the abatement level for
paint. I.e., 5000 pg/g or 1 mg/cm2.2 Inconclusive XRF measurements
currently must be confirmed In the laboratory using a more accurate
method such as atomic absorption spectrometry (AAS) or Inductively
coupled argon plasma emission spectrometry (ICP). This standard
operating procedure describes use of these two methods for
determination of lead In paint.
1.2 SUMMARY OF METHOD
1.2.1 Sampling and Measurement
Paint chips will be collected In the field according to HUD
guI delInes.2 The collectIon of blank paint film samples will also
be performed wherein these blanks consist of non-lead-based paint
(as determined by XRF or some other screening technique) collected
In the vicinity of the lead-based paint.
Lead In the paint Is solubllized by extraction with nitric
acid (HNO3) and hydrogen peroxide (H202) facilitated by heat
(modification of NIOSH 7O82)3, or by a mixture of HNO3 and hydro-
1
-------
chloric acid (HCI) facilitated by microwave energy.4
The lead content of the sample Is measured by atomic absorp-
tion spectrometry (AAS) using an air-acetylene flame, the 283.3 or
217.0 nm lead absorption line and the optimum Instrumental
conditions recommended by the manufacturer. Alternatively the lead
Is measured by Inductively coupled argon plasma emission spectrome-
try (tCP). the 220.35 nm emission line, and the optimum Instrumen-
tal conditions recommended by the manufacturer.
1.2.2 Range. Sensitivity and Method Discrimination Limit
The values given below are typical of the method's capabili-
ties. Absolute values wtll vary for individual situations
depending on the complexity of the paint sample, the type of
Instrument used, the lead line and operating conditions.
1.2.2.1 Range—
Using the NIOSH method (without additional dilutions), a
typical sample analysis range for AAS Is 1OOO - 20,000 |ig Pb/g
(0.1O - 2%) assuming the Instrument Is linear up to 20 pg/mL, while
for ICP, the typical range Is 10O - 200,OOO ng Pb/g (0.010 - 20%)
assuming the Instrument Is linear up to 2OO ng/mL. A paint sample
mass of 0.1 g and a solution volume of 10O mL Is assumed for
determination of both of these ranges.
Using the microwave method (without additional dilutions), a
typical range for AAS Is 20O - 4,000 tig Pb/g (0.020 - 0.4%) while
for ICP, the typical range Is 20 - 40,OOO |*g Pb/g (O.002 - 4.0%).
The upper linear ranges and sample mass are assumed to be the same
as presented In the previous paragraph; the solution volume Is
assumed to be 20 mL. In order to analyze high levels of lead by
AAS In samples prepared using the microwave method, the samples
will need to be dI Iuted. A 1 to 5 dI IutI on wI I I extend the I I near
range to 20,OOO ng Pb/g (2.O%).
1.2.2.2 Sensitivity—
Typical AAS sensitivities for 1 percent change In absorption
(O.OO44) absorbance units) are 0.2 and 0.5 ng Pb/mL for the 217.0
and 283.3 nm I Ines, respectIvely. ICPsensltlvlty Is a functIon of
the photocurrent Integration time as well as other Instrumental
parameters. However, an Indication of ICP sensitivity at a given
wavelength Is the ratio of net analyte intensity to background
analyte Intensity, 'n/lb- For the 22O.35 nm line, a reasonable
-------
value for this ratio Is 50 - 1OO, which would result In a detection
limit of approximately 0.05O ng/mL (50 ppb).5
1.2.2.3 Method Discrimination Limit (MDL)—
A typical MDL for AAS Is 500 ug Pb/g and for ICP Is 50 ng Pb/g
using the HNC^/H^ hotplate method and for AAS Is 100 |ig Pb/g and
for ICP Is 10 pg Pb/g using the HNO3/HCI microwave method. The
smallest mass of lead that can be detected by flame AAS (assuming
a solution volume of 100 mL) is 100 t*g while the smallest mass of
lead that can be detected by ICP (assuming a volume of 100 mL) Is
1O ng. These values were calculated as equivalent to twice the
wIthIn-laboratory standard deviation obtained for the lowest
measurable lead concentration In a test of the method.6*7 A paint
sample weight of 0.1 gm Is assumed.
1.2.3 Interferences
Interferences for AAS and ICP can be manufacturer and model
specific. The following are general guidelines.
1.2.3.1 AAS—
1.2.3.1.1 Chemical Interferences—Chemical Interferences,
that Is Interactions between molecular and/or Ionic species during
the absorption process, are not expected and therefore no correc-
tion for chemical Interference Is given here. If the analyst
suspects that the sample matrix Is causing chemical Interference,
the Interference must be verified and corrected by carrying out the
analysis with and without the method of standard additions.7
1.2.3.1.2 Light Scattering—Nonatomlc absorption or light
scattering, produced by high concentrations of dissolved solids In
the sampte, can produce a significant Interference, especially at
low lead concentrations. The Interference Is generally greater at
the 217.O nm line than at the 283.3 nm line. Light scattering
Interferences can be corrected Instrumentally. Since the dissolved
solids can vary depending on the origin of the sample, the
correction may be necessary, especially when using the 217.0 nm
line. Dual beam Instruments with a continuum source give the most
accurate correction. A less accurate correction can be obtained by
using a nonabsorblng lead line that Is near the lead analytical
line. Information on use of these correction techniques can be
obtained from Instrument manufacturers' manuals.
-------
If the instrumental correction Is not feasible, the effects of
the Interference can be el Immated through a prelImtnary separation
of the lead from the sample extract. The lead Is complexed by
ammonium pyrroJId InecarbodIthlonate and the complex then extracted
Into methyl Isobutyl ketone.8 The complex-ketone solution Is then
analyzed directly by atomic absorption spectrometry.
1.2.3.2 ICP—
1.2.3.2.1 Spectral Interference—The efficient excitation of
sample constituents at high temperature results In the possibility
of spectral overlap Interferences. A mathematical correction can
be applied for the Interference If the interfering element and the
magnitude of the Interference are determined. As an alternative,
an interference-free line may be chosen If the line exhibits an
adequate detection limit. Background shifts due to stray light,
line broadening and recombination continuum and other less well-
defined sources, require correction by background measurement near
the analysis line. This correction normally Is done dynamically
within the Instrument.
1.2.3.2.2 Physical Interferences—Paint digest samples may
contain species that affect the efficiency of nebullzatlon with
respect to standards when matrix matching Is not possible. The
existence of physical Interferences may be checked for by using the
method of standard additions. It has been observed that the high
concentrations of dissolved materials In paints may depress the
lead values. This effect can be tested by analyzing a set of
serial dilutions of the original digest. An Increase in the value
(properly corrected for the dilution) Indicates a matrix effect.
1.2.3.3.3 Chemical Interferences—Chemical Interferences,
that Is, Interactions between molecular and/or Ionic species during
the emission process, are Insignificant for ICP because of the
completeness of destruction of the sample by the high energy of the
pIasma.
1.2.4 Precision and Bias
Precision of samp I Ing of paint chips Is principally dependent
upon the number of layers of paint In the chip and the variability
In the thickness of these layers, some of which may contain more
lead than others. No typical value for sampling precision has been
establI shed.
-------
The combined extraction-analysis relative standard deviations
are as fol lows :
Hotplate Extraction
ICP 6 - 10% (at >3OO pg Pb/g)
AA 4-8% (at >100O jig Pb/g)
HNO^/HC I Microwave Extraction
ICP 2-6% (at >3OO pg Pb/g)
AA 2 - 4% (at >1OOO |ig Pb/g)
Single laboratory experiments Indicate that there Is no
significant difference In lead recovery between the hotplate and
microwave extraction procedures, and recovery of lead from
synthetic paint samples and NBS SRM 1579 (lead In paint) was found
to be greater than 90 percent In an I nter laboratory study.
2.0 APPARATUS
2.1 SAMPLING
The paint sample collection apparatus is described tn Section
A. 5. 3.1. of the HUD Guidelines.2
2.2 INSTRUMENTATION
2.2.1 Atomic Absorption Spectrophotometer
Flame a torn I zat I on spectrophotometer equipped with lead hollow
cathode or electrode I ess discharge lamp. Perk In Elmer Model 603 or
equivalent may be used.
2.2.1.1 Acetylene —
The grade recommended by the Instrument manufacturer should be
used. Change cylinder when pressure drops below 50 - 100 pslg.
2.2.1.2 Air —
Filtered to remove partlculate, oil and water.
-------
2.2.2 Alternatively, Inductively Coupled Argon Plasma Emission
Spectrometer
Computer-control led plasma emission spectrometer with
background correction and radio-frequency generator. Leeman Labs
Plasma Spec ICP 2.5 or equivalent may be used.
2.2.2.1 Argon Gas Supply—
Ensure that adequate argon, water and electrical power are
available. Liquid argon Is the most desirable source of argon,
especially for dally use from a cost and labor perspective, if gas
Is used, ensure adequate purity.
2.2.2.2 Cooling Water—
ReclrculatIng or fresh water that meets flow rate and
temperature specifications.
2.2.3 Hotplate
Surface temperature, 140*C,
2.2.4 Alternatively, Microwave Digestion System
Nominal 6OO watts power. Includes turntable, 120 mL Teflon
vessels and Capping Station. CEM Corporation MDS-81D or equivalent
may be used. The power available for heating Is to be evaluated
weekly. This quality control function is performed to determine
that the microwave has not started to degrade and that absolute
power settings (watts) may be compared from one microwave unit to
another.
This power evaluation Is accomplished by measuring the
temperature rise In 1 kg (1.0 liter) of water exposed to microwave
radiation for a fixed period of time.9
The water Is placed In a Teflon" beaker and stirred before
measuring the temperature. The beaker Is circulated continuously
through the field for 2 minutes with the unit at full power. The
beaker Is removed, the water vigorously stirred, and the final
temperature recorded. The final reading Is the maximum temperature
reading after the energy exposure. These measurements should be
accurate to +_ 0.1*C and made within 30 sec of the end of heating.
The absorbed power Is determined by the following relationship
p - (K) (Cp) (m) (AT)
-------
P - the apparent .power absorbed by the sample In watts (W).
(W-Joule* sec"1)
K « the conversion factor for thermochemlcal ca lor les'sec"'
to W (-4.184)
Cp « the beat capacity, thermal capacity, or specific heat
(cal«o" •C"r), of water
m - the mass of the water sample In grams (g).
AT - Tf, th« final temperature minus Tl, the Initial tempera-
ture CC). and
t - the time In seconds (s).
Using 2 minutes and 1 Kg of distilled water, the calibration
equation simplifies to: P - (AT) (34.87).
The microwave user can now relate power In watts to the
percent power setting of the unit.
2.2.5 Apparatus - HMO^/H^? Hotptate Digestion
Beakers: Phillips, 125 mL or Griffin, 50 ml_ with watchglass
covers.
Volumetric Flasks: 200 and 100 mL.
Assorted Volumetric PI pets: As needed.
Bottles with caps: Linear Polyethylene, 10O mL.
NOTE; Only boros I I Icate, Class A glassware Is to be used.
Also, before use, all labware should be scrupulously cleaned. The
recommended procedure Is:
1. Wash with hot, laboratory detergent solution or ultrason-
Icate with laboratory detergent solution.
2. Rinse and then soak a minimum of 4 hours In 50% V/V
nltrIc acid.
3. Rinse 3 times with doubly delonlzed water.
2.2.6 Apparatus - HNOa/HCI Microwave Method
Centrifuge: International Equipment Company Model UV or
equivalent.
Centrifuge Tubes: Oak Ridge 3O mL polysulfone tube, polypro-
pylene screw closure, Nalgene 3115-0030 or equivalent.
Pipette, Automatic Dispensing Class A: SMI Incorporated
Unlpump 200 or equivalent.
Shaker, Mechanical: Eberback Corporation 6460 or equivalent.
-------
2.2.7 Reagents - HNOg/H^ Hotplate Digestion
Nitric Acid; Concentrated, spectrographlc grade
Nitric Acid, 10% (W/V): Add 100 mL concentrated nitric acid
to 5OO mL delonlzed water; dilute to 1L.
Hydrogen Peroxide: 30% HjO^, W/W, ACS reagent grade.
Doubly Delonlzed Water: Building water passed through a
PolymetrIcs, 3 cartridge system or equivalent, then through a
MM I Ipore Corporation Mlltl-Q deIonizer or equivalent, and having
a minimum of 15 megohm-cm resistivity.
2.2.8 Reagents - HNOa/HCl MIcrowave OIgest ton
Doubly Delonlzed Water: Building water passed through a
Polymetrlcs, 3 cartridge system or equivalent, then through a
MlfiI-Q deIonizer or equivalent, and having a mlnImum of 15 Megohm-
cm resistivity.
Hydrochloric Acid: Concentrated, ACS reagent grade.
Nitric Acid: Concentrated, spectrographlc grade.
Extraction Solution: In a 1 liter volumetric flask, combine
In order and mix well: 500 mL doubly delonlzed water, 55.5 mL of.
concentrated spectrographlc grade nitric acid (16.0 N) and 167.5 mL
of concentrated hydrochloric acid (12.3 M). Cool and dilute to l
liter with doubly deion I zed water.
CAUTION; Nitric Acid and hydrochloric acid fumes are toxic.
Prepare in a well ventilated fume hood.
2.2.9 Reagents - Measurement
Master Stock Solution: 1OOO |ig Pb/mL. Commercial standard;
alternatively, weigh out 1.5985 g ACS reagent grade Pb(NO3) that
has been dried for two hours at 110'C and dissolve In 200 mL water
In 1 L volumetric flask. Add 10 mL concentrated HNO3 and dilute to
volume with water. Store In a linear polyethylene or Teflon
bottle. Stable - one year.
3.O PROCEDURE
3.1 SAMPLE PREPARATION
Final results may be reported In area concentration (mg/cm2)
or mass concentration ((ig/gm). If area concentration Is desired,
be sure that areas are provided for each paint chip. Then proceed
to weigh each total chip sample; only a fraction will be taken for
analysis and final concentration will be determined by relating
fractional mass to total mass.
8
-------
Cut the paint chips Into small pieces using a sharp blade2.
or alternatively, crush them In a beaker using a glass rod. The
sample may be further ground to a fine powder using a mortar and
pestle. Alternatively, a small motorized hammermlI I or other
grinding device may be used. Reducing the sample to a fine powder
further assures that the extraction methods will be acceptably
effIclent.
3.2 SAMPLE EXTRACTION
3.2.1 HNO?/H?O? Hotplate Extraction
Weigh out 0.1 g (nearest milligram) of sample Into a 50 mL
beaker or 125 mL PhI I IIps beaker. Add 3 mL concentrated HNO3 and
1 mL 30X H;j02 and cover with a watchglass. Start a reagent blank
at this step. Heat on a hotplate <140*C) until most of the actd
has evaporated. Remove the sample from the hotplate and allow It
to cool. Repeat this process two more times using 2 mL concentrat-
ed HNO3 and 1 mL 30% H^ each time. finally, heat on a 140'C
hotplate until the solution Is near dryness.
Rinse the watchglass and walls of the beaker with 3 to 5 mL
1O% HNO3- Allow the solution to evaporate gently to dryness. Cool
each beaker and add 1 mL concentrated HNO3 to the residue. Swirl
to dIssolve soluble spec Ies. Next perform fI 11rat I on, wh ten shouId
take pi ace under the hood. Use a wash bottle fllled with deIon I zed
water for rinsing. Set up the glass funnels over 1OO mL pre-
labeled volumetric flasks. In each funnel, place a folded Whatman
•54 filter paper. Before filtering, wet filter paper and rinse
glassware with about 20 - 30 mL of water. Discard waste rinse. To
filter, decant the liquid from the sample first, then pour the
solids onto the filter. Once this has drained, wash the beaker
with 3 small (3 mL) portions of water, adding each wash to the
fI Iter paper. Rinse the fllter paper wlth 3 smaI I (3 mL) port Ions
of water. After the filter paper Is thoroughly drained. It Is
discarded. Rinse the glass funnel with one small portion of water.
Ollute to volume with del on I zed water. The sample Is 1% In nitric
acid. Caution: Nitric acid fumes are toxic.
3.2.2 HNOq/HCI Microwave Extraction
Weigh out 0.1 gram (nearest ml II I gram) of sample Into a 30 mL
polysulfone Oak Ridge centrifuge tube. Add 10 mL of extraction
solution (Section 2.2.8) using Class A automatic dispensing pipette
-------
(SMI Incorporated Unlpump 2OO or equivalent). Cap the tube
tightly.
Pipette 31 mL of double del on I zed water Into a 120 mL Teflon
microwave digestion vessel. Place an Oak Ridge centrifuge tube
containing the sample In the 120 mL Teflon microwave digestion
vessel. Place a safety valve and cap on the vessel and tighten the
cap using the capping station. Fill the microwave turntable with
12 vessels containing the centrifuge tubes. Put the filled
turntable In the microwave oven; activate the "on" switch and the
"turntable" switch. Set the exhaust fan to maximum speed. Program
the microwave oven for a time of 23 minutes and a power of 81% (522
watts) and press the "start" button.
At the end of the program, remove the turntable containing the
microwave vessels and cool It In tap water for 1O minutes. Open
the microwave vessels and discard the water they contain. Open the
Oak Ridge centrifuge tubes and add 10 mL of doubly delonlzed water
using a Class A automatic dispensing pipette (SMI Incorporated
Unlpump 200 or equIvalent). Cap the tubes tightly and mechanically
shake 5 minutes. Centrifuge 25 minutes at 200O RPM (International
Equipment Company Model UV or equivalent). Open the centrifuge
tubes and decant or pipette off the clear solution Into an acid
cIeaned 20 mL scIntI I I at I on vial for analysis. Use a samp Ie voIume
of 20 mL to calculate analytical results. The sample Is 1.03 M In
hydrochloric acid and 0.45 M In nitric acid.
NOTE: The sample solutions may need to be further diluted to stay
within the linear calibration range.
4.0 ANALYSIS
4.1 AAS-CALI BRAT I ON
4.1.1 Working. Standard, 2O t*g Pb/mL
Prepare by diluting 2.O mL of the 1OOO tig/mL master stock
solution (Section 2.2.9) to 10O mL In 1% HNO3 If the HhK^/H^
hotplate extraction was used, or 0.45 M HNO3/1.03 M HCI If the
HNOg/HCI microwave method was used. The working standard should be
prepared at least weekly; dally preparation Is preferred.
4.1.2 Calibration Standards
Prepare dally by diluting the working standard, as Indicated
below with acid solution to match the sample matrix (1% In HNO3 or
O.45 M HNO3/1.03 M HCI). Other lead concentrations may be used.
10
-------
Volume of 20 |tg Pb/mL working Final Concentra-
standard. mL volume, mL tlon, |*g Pb/mL
0 1OO 0
5.O tOO 1.0
25.0 100 5.0
50.0 100 1O.O
100.0 100 2O.0
4.1.3 Calibration Curve
The calIbratlon curve may be manually plotted, determined with
a hand calculator using linear regression analysis or calculated
automatically. Some automatic systems will simply display the
analysis results calculated by the Internal electronics and/or
computer. Other, more complex systems will allow selection of the
curve fitting function (e.g., linear, polynomial, segments I) and
provide values for the function constants (e.g., slope and
Intercept for the linear function y - mx + b). When first
calibrating the system or after any significant change to or work
on the Instrument, a manually plotted standard curve should be
compared to the standard curve calculated from the mathematical
function. Any difference In the curves of more than 1O% needs to
be Investigated and corrective action taken. Such action may
Include selection of a different curve fitting function.
4.2 ICP - CALIBRATION
4.2.1 Working Standard, 100 pg/mL
Prepare by diluting of 10.0 mL of the 10OO ng/mL master stock
solution to 1OO mL In 1% HNO3 If the HhK^/HjO? hotplate extraction
was used or O.45 M HNO3/1.03 M HCI If the HNO3/HCI microwave method
was used. The working standard should be prepared at least weekly;
dally preparation Is preferred.
4.2.2 Calibration Standards
Normally 2 to 5 standards are used for ICP calibration.
Typical concentrations are shown below. Prepare dally by diluting
the working standard, as Indicated below.
11
-------
Volume of 100 |*g Pb/mL working Final Concentra-
standard, mL volume, mL tlon, |ig Pb/mL
0 1OO 0
1.O 200 0.5
3.O 100 3.0
10.0 100 10.0
30.0 100 30.0
100.0 100 100
Higher lead concentrations may be used as long as I I near I ty of
response Is maintained.
4.2.3 CaII brat Ion Curve
The calIbratlon curve (Integrated photocurrent [or equivalent]
ys concentration) will be calculated automatically. When first
calibrating the system or after any significant change to or work
on the Instrument, a manually plotted standard curve should be
prepared and then compared to the standard curve calculated by the
system. Any difference In the curves of more than 10% needs to be
Investigated and corrective action taken.
4.3 QUALITY CONTROL PRIOR TO SAMPLE ANALYSIS
Qua I Ity control Is necessary to assure that resulting data are
of adequate quality. Several tests are to be performed prior to
sample analysis. These are as follows:
4.3.1 Blank Check
Laboratory or reagent blanks are analyzed to determine the
background or contamination levels. Contamination levels above
detection limit must be accounted for and eliminated, If possible,
before proceeding with sample analysis. Field blanks (that is,
paint samples testing very low In the field) that show lead levels
well above levels for "lead-free" paint, that Is, above 500 - 1000
pg Pb/g, Indicate possible cross contamination of samples. As with
laboratory blanks, high lead values for field blanks must be
accounted for and corrective action taken. If necessary.
4.3.2 Matrix Interference Check
Chemical and/or physical Interferences may cause error. These
are checked by the methods of standard additions and sample
dllutIon.
12
-------
4.3.2.1 Method of Addition Check—
Allquots of digests representing each source of paint samples
are spiked with lead solution after Initial analysis to approxi-
mately double the concentration. The recovery must be within 80%
to 12O% of the known value. The spike addition should produce a
minimal level of 10 times and a maximum of 100 times the Instrumen-
tal detection limit. If the spike Is not recovered within the
specified limits, a matrix effect should be suspected. The use of
a standard-addition analysis (MSA) procedure can usual ly compensate
for this effect. If an MSA procedure does not produce acceptable
recovery, then the digestion procedure must be regarded as suspect.
CAUTION: The standard-addition technique does not detect
coincident spectral overlap. If suspected, use of
computerized compensation, an alternate wavelength,
or comparison with an alternate method Is recom-
mended .
4.3.2.2 Dllutlon Check—
It has been observed that the high concentrations of dissolved
materials In paints depress the values measured by ICP. The effect
must be tested for by analyzing a set of serial dilutions of the
original digest, e.g., 1:10, 1:25, 1:100. An Increase In the value
(properly corrected for the dilution) Indicates a matrix effect.
Such a dilution test should be performed for each new matrix type.
The final dilution ratio used will be limited by the lead concen-
tration, which should be between 1 and 10 ppm for optimum measure-
ment .
4.3.3 ICP Interfering Element Check
When lead In paints Is being measured by ICP, It Is Important
to be aware of the potential for spectra I Interferences due to the
ex I stence of potent I ally high levels of Interferences (e.g. Tl, Al,
Cr, etc). It Is Important to periodically analyze InterferIng
Element Check Samp Ies that contain known high levels (200 - 1OOO
ppm) of each suspected Interfering element. Such solutions are
available from a variety of vendors. Once the solutions are
analyzed, the data must be evaluated to determine the existence of
a false lead value attributed to the Interferences that are more
than 2 x the solution detection limit. If the false values do
exceed this criteria, an Interfering element correction factor
(FjEC) must be determined as follows:
13
-------
F|tc- Falsa analyte si gnat
concentration of .inter fie rant
For example - 1000 ppm of aluminum causes an approximately
false lead signal of 0.250 ppm (7 x DL^)
Therefore, F(K - (0.25/1000)-0.00025
This value Is used to correct lead data In the presence of
high aluminum. The Interfering element Identified In the above
manner Is therefore added to the analytical program. This
procedure must be applied to all potential Interfering elements.
4.3.A CaI Ib r a 11on Check Samp Ies
A check sample prepared from an Independent master stock
solution must be run after standardization to determine the
accuracy of the simple aqueous standards. The concentration of the
check sample should be approximately 75% of the highest catIbratlon
standard. Agreement must be within *5% of expected or a recaIIbra-
tlon must be performed, possibly with fresh standards.
4.4 QUALITY CONTROL DURING ANALYSIS
During the course of.ana lysis, the following quality control
activities are to be performed.
4.4.1 Reagent Blanks
A reagent blank (extraction reagent carried through entire
analytical process) Is to be run after every 20 samples. A sudden
Increase would Indicate a contamination problem.
4.4.2 CaIIbrat Ion Checks
High and low. Independently prepared check samples are to be
run alternately after every 10 samples to determine that calibra-
tion has not drifted. If a change of more than 10% is measured,
the system must be recalibrated and all samples run since the last
calibration check rerun.
The results should be plotted on a control chart at the end of
each sample analysis session, although real-time checking Is
preferred.10 The analysis Is concluded to be out of control If any
one or more of the following Is met.
1. One or more points outside of the control limits.
14
-------
2. A run of at least eight points, where the type of run
could be either a run up or down, a run above or below
the center line, or a run above or below the median.
3. Two of three consecutive points outside the 2-slgma
warning limits but still Inside the control limits.
4. Four of five consecutive points beyond the 1-slgma
Iimlts.
5. An unusual or nonrandom pattern In the data.
6. One or more points near a warning or control limit.
4.4.3 Pup IIcates
Analyze one duplicate samples for every 20 samples. A
duplicate sample Is a sample brought through the whole sample
preparation and analytical process. The acceptance criteria for
precision of the duplicate analyses varies with proximity of the
analytical result to the detection limit and Is as follows:
Average Analyte Concentration
Concentration (Multiples of
Detection Limit
O - 2
2-10
Maximum Acceptable,
Average Relative Percent
Dlfference
200%
17.3%
8.6%
Where Average Relative Percent Difference -
((X, - X2)/((X, - X2)/2)) X 100
These values result In estimates of the 95% confidence Intervals
for the method of (1) _+ 30% for concentrations 2 - 10 x the method
discrimination limit, and (2) £ 15% for concentrations > 10 x the
method discrimination limit.11 If unacceptable precision Is
obtained, corrective action Is to be taken Including review of all
original data and calculations and possible analysis of a second
dupI Icate samp Ie.
4.4.4 Standard Reference Materials (SRMs)
Depending on the matrix, a standard reference material should
be analyzed once per sample batch or, at a minimum, once per day to
check the entire extraction/analysis procedure. Lead recovery
should be within 90 to 110% of the known value. An appropriate
reference material for lead at the present time Is NIST 1579
Powdered Lead-Based Paint at 11.87%. Additional paint standards
having lower lead concentrations will be available from NIST
15
-------
sometime In 1992. Plot results on a control chart as outlined In
Section 4.4.2. If the sample Is out of control, sources of error
must be Identified and appropriate corrective action taken.
4.5 SAMPLE DETERMINATION
4.5.1 AAS
Most pertinent startup procedures may be found In the
manufacturer's operation manual. The operator should be reasonably
familiar with the operation manual regarding basic operation and
safety. However, these procedures are outlined below.
1. Turn on the power and Install the appropriate lamp and
burner head.
2. Set the source lamp current to proper value.
3. Set the silt to the proper value. Set the wavelength to
proper value and peak the wavelength setting. Align the
lamp.
4. Set the control switch to the desired measurement mode
(absorption).
5. Turn on and adjust background correction. If available.
6. Select the proper, flame and flow rates and Ignite the
gases according to the manufacturer's procedure manual.
The proper flame Is listed I n the manufacturer's ana IytI -
caI methods manual. Follow manufacturer's recommenda-
tions regarding warm up times.
7. Select the desired Integration time.
8. Aspirate a blank solution and auto zero the Instrument.
9. Aspirate the calibration standards and establish a
calibration curve either manually or automatically such
that the standards bracket the samples.
10. Run a calibration check sample as described In Section
4.3.4.
It. Aspirate a sample solution and measure the absorbance
and/or the concentration.
4.5.2 ICP
1. Ensure that adequate argon, water and electrical power
are aval I able. LI quid argon Is the most desIrable source
of argon, especially for dally use from a cost and labor
perspective. If gas Is used, ensure adequate purity.
2. Adjustment of NebuI Izer Spray - See operator's manual for
procedure.
3. Ignition of Torch - Check argon supply Is on.
16
-------
4. After startup - Be sure plasma does not flicker or
present an orange corona around torch. if the plasma
fIIckers, be sure the spray chamber Is draining properly.
If the orange corona Is observed, make sure that the
nebulizer argon Is on. Otherwise some residual salt may
be present In the nebulizer spray that must be flushed
out or the entire spray chamber assembly must be cleaned.
5. Warmup - Allow the Instrument to warm up at least 30
minutes before serious analyses are Initiated and the
standard readings have stabilized.
6. Optical Calibration/Torch Alignment Procedures - Before
analytical calibration procedures are performed, It Is
Important to perform the optical calibration procedures
and the torch alignment operation. Each of these Is
described In the operator's manual.
7. Select program that Includes wavelength, Integration
time, number of replicate readings, sample uptake time
and rinse time.
8. Aspirate the calibration standards and establish a
calibration curve.
9. Hun a calibration check sample as described In Section
4.3.4.
10. Aspirate a sample solution and measure the emission
signal.
5.O DATA PROCESSING
5. 1 AAS
The absorbance of each sample result Is recorded. If the
readout Is In absorbance, this value Is entered Into the linear
regression equation and the concentration Is calculated. Alter-
nately the Instrument will provide a direct readout In concentra-
tion.
For direct determination, read the element value (ng/mL) from
the calibration curve or readout. If dilution of the sample has
been performed, then
ng/mL element In the sample - pg/mL In the dilution X D
Where D • (mL of aliquot) + (mL of diluent)
Of
5.2 ICP
The ICP will provide direct readout In concentration.
Correction for dilution Is made as described In Section 5.1.
17
-------
5.3 CALCULATION - FIELD SAMPLE CONCENTRATION
5,3.1 Area Concentration
The area concentration of lead In a paint chip Is calculated
as fol lows:
mg Pb/cm2 - (CTS x VTS x MQJ /M^ ) / ( 1 000 x A^)
where CTC - lead concentration In test solution, corrected
for dilution, \ig Pb/mL
VTS « volume of sample digest solution, mL
mass of original sample, g
* mass of sample aliquot digested, g
« area of or I g Inal sample, cm2
5.3.2 Mass Concentration
The mass concentration of lead In a paint chip Is calculated
as fol lows:
ng Pb/g - (CTS x v^/M^
where Cjc «• lead concentration In test solution, corrected
for dilution, |*g Pb/mL
VTS m volume of sample digest solution, mL
* mass of sample aliquot digested, g
6.0 REFERENCES
1. Lead-Based Paint Poisoning Prevention Act, 42 U.S.C. 4822
(d)(2)(A), 1971.
2. Lead-Based Paint: Interim Guidelines for Hazard I dent I-
TTcatlon and Abatement In Public and Indian Housing,
Department ofhousing and Urban Development,September
1990.
3. National Institute for Occupational Safety and Health,
NIOSH Manual of Analytical Methods, Third Edition, 1964.
NIOSH Method 7082, Issued 2/15/84.
4. Pranger, Louis, J., Standard OperatIng Procedure for
Microwave Extraction of Giass-FJber_ FI Iters, u7s.
Environmental Protection Agency, AREAL/RTP-SOP-MRDD-037,
January, 1990.
18
-------
5. Wlnge, R. K., V. A. Fasse I , V. j. Peterson and M. A.
FIoyd, Inductively Coupled Plasma-Atomic Emission
Spectroscopy. Elsevler. New York, p. 276,1985.
6. "Proposed Recommended Practices for Atomic Absorption
Spectrometry. " ASTM Book of Standards, part 30, pp. 1596
- 1608 (July 1973~!T
7. Blnstock, D. A., D. L. Hard I son, J. White, P. M. Grohse
and W. F. Gutknecht, Eva IuatI on of Atom Ic SpectroscopIc
Methods for Determination of Lead in P<*' nt, soil and
Dust,U.S.E.P.A.contractNo.68-O2-4560, September
1991 .
8. Kolrtyohann, S. R. and J. W. Wen, "Critical Study of the
APCD-MIBK Extraction System for Atomic Absorption."
Anal. Chem.. 45, 1986-1989 (1973).
9. Blnstock, D. B., W. M. Yeager, P. M. Grohse and A.
Gask III, Validation of a Method for Determining Elements
In Solid' waste by Microwave Digestion, u.5.6. .P . A.
Contract No. 68-O1-7266, November1989.
10. Montgomery., D.C., I ntroduct I on to Stat | st I ca I Qua I I ty
Contro t, 2 ed, John wiiey & Sons, 1991.
11. Personal communication, John Moore, U.S. Fish and
Wildlife Service, Patuxent, Maryland, 1991.
19
-------
Appendix C-2
Laboratory XRF SOP - "Standard Operating
Procedures for Energy Dispersive
X-ray Fluorescence Analysis of
Lead in Urban Soil and Dust
Audit Samples"
-------
RESEARCH TRIANGLE INSTITUTE
Center for Environmental Measurements and Quality Assurance
May 29,1992
Mr. Thomas Nadermann
Keystone NBA Environmental Services
12242 S.W. Garden Place
Tigard, Oregon 97223
Dear Mr. Nadermann:
Please find enclosed the document, "Standard Operating Procedures for Energy
Dispersive X-Ray Fluorescence Analysis of Lead in Urban Soil and Dust Audit Samples,"
referenced in the letter sent to you with the round robin samples. If your laboratory has
established protocols for the analysis of dust, please follow these established protocols. We
are including the SOP only as a reference for laboratories that do not have standard
procedures for these analyses.
Once again, thank you for your participation in the EPA/RTI round robin for lead-
based paint and dust.
Sincerely,
Emily Williams
Post Office Box 12194 Research Triangle Park, North Carolina 27709-2194
Telephone 919 541-6914 Fax: 919 541-5929
-------
STANDARD OPERATING PROCEDURES
FOR ENERGY DISPERSIVE X-RAY FLUORESCENCE ANALYSIS
OF LEAD IN URBAN SOIL AND DUST AUDIT SAMPLES
by
Dawn M. Boyer & Daniel C. Hillman
Lockheed Engineering & Sciences Company
Las Vegas, Nevada 89119-3705
Contract No. 68-CO-OQ49
Project Officer
Harold A. Vincent, Quality Assurance Division
Environmental Monitoring Systems Laboratory
Las Vegas, Nevada 89193-3478
ENVIRONMENTAL MONITORING SYSTEMS LABORATORY
OFFICE OF RESEARCH & DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
LAS VEGAS, NEVADA 89193-3478
-------
Notice
This document is a preliminary draft. It has not been formally
released by the u. S. Environmental Protection Agency policy. It
is being circulated for comments on its technical merit and policy
implications.
Mention of corporation names, trade names, or commercial products
does not constitute endorsement or recommendation for use.
-------
Table of Contents
List of Abbreviations i
INTRODUCTION ii
1.0 SAMPLE PREPARATION 1
1.1 Soil Samples 1
1.2 Dust Samples 1
1.3 Loading XRF Sample Cups for Analysis 1
2.0 ENERGY DISPERSIVE X-RAY FLUORESCENCE ANALYSIS 1
2.1 Summary . . l
2.2 Instrument Parameters 2
2.3 Peak Processing Procedure 2
2.4 Calibration 3
2.5 Determination of Unknown Sample Concentration ... 3
3.0 QUALITY CONTROL 3
4.0 LABORATORY SAFETY ..... 4
REFERENCES 5
-------
Page: i of ii
List of Abbreviations
DL detection limit
GFAA Graphite Furnace Atomic Adsorption
ICPAES Inductively Coupled Plasma Atomic Emissions Spectroscopy
HCV high calibration verification sample
LCV low calibration verification sample
LCS laboratory control sample
MCA multichannel analyzer
MDL minimum detection limit
QA quality assurance
RM reference monitor
RSD relative standard deviation
SOP standard operating procedure
ULADP Urban Lead Abatement Demonstration Program
XRF X-ray fluorescence
-------
Page: ii of ii
INTRODUCTION
Lead in the human body, whether at high or low concentration,
temporary or long lasting, may result in a broad spectrum of
adverse health effects. These effects, sometimes called "lead
poisoning1* when severe, range from dizziness, hearing impairment,
destruction of red blood cells, and delayed cognitive behavior, to
convulsions, coma, and death. While lead poisoning can be treated,
many of its developmental effects are irreversible.
Young children are the population most at risk from excessive lead
exposure due to their physiological development and their frequent
contact with lead-contaminated parts of their environment (dust,
leaded paint chips, soil, etc.). Lead exposure may result from
normal outdoor play activities as well as from indoor contact with
paint and contaminated dust which pay collect on carpets, floors,
and furniture. The human fetus is also part of this high-risk
population; lead in the maternal bloodstream may produce toxic
fetal effects including reduction in gestational age, birth weight,
and mental development'.
Energy dispersive x-ray fluorescence (XRF) has been identified as
an effective analytical tool for measuring lead in solid materials
including dust, soil, and paint. XRF advantages are that it quick,
precise, cost effective, nondestructive and requires minimal sample
preparation. This standard operating procedure (SOP) was designed
to provide a method suitable for measuring lead in urban soil and
dust audit samples for the Urban Lead Abatement Demonstration
Project (ULADP).2
-------
Date: 04-09-92
Revision No.: 1
Page: 1 of 5
1.0 SAMPLE PREPARATION
1.1 Soil Samples- It is assumed that soil samples have
previously been reduced to < 60 mesh. This procedure is
written assuming an initial sample size of about 20 g.
1.1.1 Homogenization and Subsampling to 5-g Aliquots
Initial Homogenization- Position the two receiving pans
under the small riffle splitter. Pour the entire 20-g
aliquot from the distribution pan evenly across the
baffles of the riffle splitter. Transfer the soil from
each receiving pan into the distribution pan and replace
the receiving pans under the riffle splitter. Repeat
this step five times in succession.
Splitting into 5-g Aliquots- Pour a 20-g aliquot evenly
across the baffles of the small riffle splitter. Place
the soil from one receiving pan into a plastic bag.
Transfer the soil from other receiving pan to the
distribution pan and continue splitting as necessary
until approximately 5 g of soil occupies each receiving
pan. Place the entire contents of the pan into pre-
labeled sample container. Repeat the procedure until the
entire 20-g sample is split into an even number of 5-g
aliquots.
1.2 Dust Samples- It is assumed that soil samples have
previously been reduced to < 60 mesh and that the sample
size of about 2 g.
1.2.1 Homogenization- Position the two receiving pans under the
small riffle splitter. Pour the entire 2-g aliquot from
the distribution pan evenly across the baffles of the
riffle splitter. Transfer the dust from each receiving
pan into the distribution pan and replace the receiving
pans under the riffle splitter. Repeat this step five
times in succession.
1.3 Loading XRF Sample Cups for Analysis- Pour a 5-g soil
aliquot or 2-g dust aliquot into an XRF sample cup and
seal with 3.6 MB mylar film.
2.0 ENERGY DISPERSIVE X-RAY FLUORESCENCE ANALYSIS
2.1 Summary - Samples are loaded into the spectrometer and
the sample is with irradiated x-rays. The characteristic
line spectrum consists of a series of discrete
wavelengths, x-ray spectral lines, characteristic of the
emitting element and having various relative intensities.
-------
Date: 04-09-92
Revision Ho.: l
Page: 2 of 5
The line spectrum of an element originates when electrons
are expelled from inner levels of its atoms, and
electrons from levels farther out fall into the
vacancies. Each transition constitutes an energy loss
which appears as an x-ray photon. The minimum photon
energy that can expel an electron from a given level in
an atom of a given element is known as the absorption
edge of that level of that element. Each element has as
many absorption edges as it has excitation potentials3
X-ray spectral lines of all elements in the sample are
excited and detected simultaneously, then the resulting
detector output pulses are separated electronically on
the basis of their pulse height.4 Loose powder samples
are analyzed by XRF. The Pb L-beta peak/ Ag Compton peak
ratio is calculated. The lead concentration is
determined from the ratio and the calibration curve
(Ratio vs. Concentration). Quality control is described
in Section 1.4.
2.2 Instrument Parameters
Instrument: Kevex Delta Analyst 770
Sample Form: Oust (< 60 mesh)
Cup Diameter: 31 mm
Counting Time: 200 sec
X-ray Tube Voltage: 35 KeV
X-ray Tube Current: 3.0 Ma
Secondary Target: Silver
Analysis Atmosphere: Air
2.3 Peak Processing Procedure
A.) Acquire the spectrum: This routine begins the
acquisition of data into the currently enabled
multichannel analyzer (MCA) memory group.
B.) Save the spectrum: This routine save the spectra in a
spectrum file.
C.) Process the escape peaks: This routine corrects spectral
data for losses due to fluorescence and subsequent escape
of silicon K-a x-rays in the detector crystal.
D.) Smooth the spectrum: This routine smooths the spectrum
using a pseudo-Gaussian 1:2:1 3-point smoothing
correlator.
E.) Deconvolute the Scatter peaks: This routine fits
Gaussians to the Compton and Rayleigh peaks, and computes
the Corapton-to-Rayleigh intensity ratio for the current
spectrum.
-------
Date: 04-09-92
Revision No.: l
Page: 3 of 5
F.) Save the Compton intensity: This routine save the
Compton intensity in a specified file.
G.) Recall the old spectrum: This routine recalls the last
spectra in memory prior to any spectral processing.
H.) Process the escape peaks: This routine corrects
spectral data for losses due to fluorescence and
subsequent escape of silicon K-ot x-rays in the detector
crystal.
I.) Process the summation peaks: This routine removes
undesired sum peaks from spectra, due to trailing-edge
pulse pileup during high deadtime acquisition.
J.) subtract the background: This routine subtracts the
background stored in the processing group P2 from the
spectrum stored in group PI.
K.) Identify the Pb peak: This routine adds specified
elements to the current element list of the current
spectrum.
L.) Deconvolute Pb L0 intensity by integration: This
routine extracts intensities by integration.
M.) Clear the background: This routine erases any
background presently stored in group P§, whether or not
it is being used.5
2.4 Calibration and Quantification- The XRF is calibrated by
acquiring spectra from a series of urban soil standards with
known lead concentrations. Currently we use a series
containing 443, 849, 995, 1069, 2455, 3772, and 17993 mg/kg
Pb. Acquisition conditions are given in Section 2.2. The Pb
L/S peak and Ag Compton peak are measured from the spectra and
the Pb Lj3 peak/Ag Compton peak ratios are calculated. A
calibration line is calculated using linear regression of the
ratio vs. the lead concentration.
2.5 Determination of Unknown Sample concentration - The Pb L0 peak
and Ag Compton peak are measured from the spectra and the Pb
L0 peak/Ag Compton peak ratios are calculated. Unknown
concentrations are determined from the calibration line
discussed in Section 2.4.
3.0 QUALITY CONTROL
Laboratory control sample (LCS) - One LCS sample will be
prepared and analyzed per group of 20 samples. A LCS is a
real sample with a matrix similar to the samples being
analyzed which contains a known concentration of lead.
Reference Monitor (RMJ - Prior to analysis, a reference
monitor sample is measured. It is an in-house synthetic
sample containing 1.273% Fe, 1.505% Sb, 1.507% Y, 9.65% Br,
-------
Date; 04-09-92
Revision No.: 1
Page: 4 of 5
17.69% Na, and 19.89% Cl. The reference monitor intensity
provides a standard measure of the x-ray flux that irradiates
the samples being analyzed. The reference monitor provides a
method of standardizing and/or compensating for changes in the
x-ray tube flux.
High Calibration Verification Sample (HCV) - A HCV sample is
a real sample containing lead at a concentration near the
upper end of the calibration line. It is analyzed after the
RM and after the last sample in a run. The concentration of
Pb (17993 mg/kg) is at the high end of the range of interest.
Low Calibration Verification Sample (LCV) - A LCV sample is a
real sample containing lead at a concentration near the lower
end of the calibration line. It is analyzed after HCV sample
in a run. The concentration of Pb (443 mg/kg) is at the low
end of the range of interest.
Detection limit (DL) Determination. - the smallest
concentration/amount of a the analyte of interest that can be
measured by a single measurement with a stated level of
confidence. This must be determined for each new sample
matrix.
Minimum Detection Limit (MDL) - the concentration/amount of
analyte that gives a net line intensity equal to three times
the square root of the background intensity. This must be
determined for each new sample matrix.
4.0 LABORATORY SAFETY
Environmental samples often contains hazardous materials and
must be handled with respect. Special equipment and
facilities are must be used to prevent safety hazards and
eliminate cross contamination of space and other samples.
Sample preparation must be performed in a fume hood and
personnel must wear a dust mask, PVC gloves, and a lab coat.
Personnel engaged in handling hazardous samples undergo
initial and periodic medical examinations to insure that they
have not contracted medical problems related to the materials
with which they are involved.
-------
Date: 04-09-92
Revision No.: l
Page: 5 of 5
REFERENCES
Aschengrau, Ann et al. (1991) Three City Urban Soil-Lead
Demonstration Project. Midterm Project Update. Unpublished
report.p.2.
Papp,M. ,Hillman,D. ,Boyerf D. ,Kohorst,K.,Vincent,H. (1990)
Standard Operating Procedures for the Preparation and
Characterization of Soil, Dust, and Handwipe Audit Samples for
the EPA Lead Abatement Demonstration Project.
Bertin, E. (1975) Principles and Practices of X-ray Spectro-
metric Analysis, p 38-40.
Bertin, E. (1975) Principles and Practices of X-ray Spectro-
metric Analysis, p 21.
Kevex Instruments (1985) Kevex XRF Toolbox™ II Reference
Manual, 3-1 - 3-218.
-------
Appendix D
Instructions to Laboratories
-------
Appendix D-l
Letter of Instruction to
AAS/ICP Laboratories
-------
RESEARCH TRIANGLE INSTITUTE
/RTI
Center for Environmental Measurements and Quality Assurance
March 31, 1992
Ha. Joan W. Etheridge
OWMC Laboratory
845 Harrington Court
Burlington, Ontario L7N3P3
Laboratory I.D. No.t
Digestion Methods: NIOSH 7082 and EPA/AREAL
Analytical Method: ICP
Dear Ma. Etheridge:
Thank you for your willingness to participate in a round robin
analysis for lead in paint and dust supportive to the U.S. Environmental
Protection Agency.
The round is designed to evaluate the level of lead in, and the
homogeneity of, a group of performance evaluation samples for lead in paint
and dust. A total of 35 laboratories will be participating in the round.
Seven laboratories will be analyzing by laboratory XRF, and the remaining
labs will be analyzing by AAS or ICP. Two of the participating labs will
analyze the samples using both laboratory XRF and AAS/ICP. Your laboratory
identification number and method of digestion {NIOSH 7082 (hotplate) or
EPA/AREAL (microwave) ) and analysis (AAS or ICP) selected by your
laboratory is shown at the top of this letter and on the enclosed data
reporting form.
Please find enclosed five (5) bottles of paint (P-l through P-5), and
five (5) bottles of dust samples (D-l through D-5) for analysis. Upon
receipt of the samples, please rotate the bottles gently through all axes
for a couple of minutes in order to compensate for any.separation that may
have occurred during shipment.
At the time of sampling, please remove two aliquots from each sample
and digest and analyze each aliquot separately. The enclosed data
reporting form provides a blank for reporting the concentration of
Aliquot 1 and Aliquot 2 for each sample, for a total of twenty (20) results
for the analysis of the paint and dust materials. It is recommended that
samples analyzed by ICP be diluted to a final solution concentration of
less than 10 ppm.
Post Ofiice Box 12194 Research Triangle Park. North Carolina 27709-2194
Telephone 919 541-6914 Fax: 919 541-5929
-------
Protocols for preparation and analysi* of samples are given in the
report, "Standard Operating Procedures for Lead in Paint by Hotplate- or
Microwave-baaed Acid Digestiona and Atomic Absorption or Inductively
Coupled Plasma Emission Spectrometry," already mailed to you under separate
cover. Centrifuge tubes are required for the EPA/AJUSAZ. digestion method,
and are enclosed. These tubes are not clean, and will need to be cleaned
per the method described in the SOP report. Pleaae follow the protocol
given to clean the centrifuge tubes (EPA/AREAL digestion method), to carry
out the digestion and to analyze samples.
An ICP Instrument Parameter sheet is enclosed. Please complete it,
along with the data reporting form, and send results to RTI no later than
Thursday, April 30, 1992. The forms should be submitted tot
EPA/RTI Round Robin for Lead in Paint and Dust
Center for Environmental Measurements and Quality Assurance
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, NC 27709
Attns Emily Williams
Building 7
A statistical analysis and report of the round will be sent to
participating laboratories by the end of June.
Again, thank you for your participation. If you have questions,
please call either David BinstocJc or Emily Williams at (919) 541-6896 or
(919) 541-6217, respectively.
Sincerely,
David Binstock
Emily Williams
-------
ICP PARAMETER SHEET
Instrument
(Manufacturer/Model)
Nebulizer
Wavelength_
Grating
Resolution
Focal Length_
Background Correction
Interference Correction
Forward Power
Reflected Power
Plasma Frequency_
Auxilliary Gas Flow Rate_
Sample Introduction Rate_
Calibration Standards and Calibration Check Samples
-------
EPA/RTI Round Robin for Lead In Paint and Dust
Round Robin No. 002
Lab ID No.
Digestion Method NIOSH 7082 Laboratory _
Experience with this Method years Laboratory
Analysis Method ICP
Experience with this Method
Approval Signature:
years
Sample ID No.
P-1
P-2
P-3
P-4
P-5
D-1
D-2
D-3
D-4
D-5
Reagent Blank
Gross Concentration of Lead (ppm)
Aliquot 1 Aliquot 2
-------
Appendix D-2
Letter of Instructions to
Laboratory XRF Laboratories
-------
RESEARCH TRIANGLE INSTITUTE
Center for Environmental Measurements and Quality Assurance
March 31, 1992
Ms. Phyllis Madigan
Massachusetts State Laboratory Institute
Environmental Lead Laboratory
Room 311
3305 South Street
Jamaica Plain, MA 02130
Laboratory I.D. No.i
Analytical Method: Laboratory XRF
Dear Ms. Madigan:
Thank you for your willingness to participate in a round robin
analysis for lead in paint and dust supportive to the U.S. Environmental
Protection Agency.
The round is designed to evaluate the level of lead in, and
homogeneity of, a group of performance evaluation samples of paint and
dust. A total of 35 laboratories will be participating in the round.
Seven laboratories will be analyzing by laboratory XRF, and the remaining
labs will be analyzing by AAS or ICP. Two of the participating labs will
analyze the samples using both laboratory XRF and AAS/ICP. Your laboratory
identification number is shown at the top of this letter and on the
enclosed data reporting form.
Please find enclosed five (5) bottles of paint (P-l through P-5), five
(5) bottles of dust samples (D-l through D-5), and two bottles of Dust
Reference Materials, CIN 1 (2275 ppm), and BAL 1 (58 ppm). Upon receipt of
the samples, and before sampling, please rotate the bottles gently through
all axes for a couple of minutes in order to compensate for any separation
that may have occurred during shipment.
At the time of analysis, please remove two aliquots from each bottle,
prepare the aliquots as individual samples and analyze each. The enclosed
data reporting form provides a place for reporting the concentration of
Aliquot 1 and Aliquot 2 for each sample, for a total of twenty (20) results
if your lab is participating in the analysis of both paint and dust.
Post Office Box 12194 Research Triangle Park, North Carolina 27709-2194
Telephone 919 541-6914 Fax: 919 541-5929
-------
We are requesting that laboratories follow their own protocol for the
XRF analysis. Please use an amount of material that corresponds to an
infinitely thick sample relative to the excitation beam, and run the sample
in a cup that is approximately 31 mm in diameter, otherwise, please select
parameters that optimize your laboratory operations, and enter these
parameters on the enclosed XRF parameter form. • Laboratories using a
wavelength-dispersive instrument, rather than an energy-dispersive
instrument, are asked to contact RTI before the analyses are begun. As a
reference, a protocol from the EPA 3-City Study will be mailed to you under
separate cover at a later date.
When analyzing the paint samples, please calibrate the Instrument with
the standards routinely used in your operations. For the dust samples, we
are requesting that you calibrate with the two reference materials enclosed
(BAL 1 and CIN 1). If you have your own dust standards, please run yoyr
standards as samples relative to the calibration curve generated with CIN 1
and BAL 1; and report the values for your standards on the enclosed XRF
Parameter Sheet for Dust.
Please use the enclosed data reporting form to submit results to RTI
no later than Thursday, April 30, 1992. The XRF parameter form and data
reporting form should be submitted to:
EPA/RTI Round Robin for Lead in Paint and Dust
Center for Environmental Measurements and Quality Assurance
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, NC 27709
Attn: Emily Williams
Building 7.
A statistical analysis and report of the round will be sent to
participating laboratories by the end of July.
Again, thank you for your participation. If you have questions,
please call me at (919) 541-6217.
Sincerely,
0
Emily Williams
-------
LABORATORY XRF PARAMETERS - PAINT
Sample Quantity_
Sample Preparation
Instrument
Description of X-ray Source_
Description of Secondary Target_
Description of Detector_
Reference
-------
LABORATORY XRF PARAMETERS - PAINT
Counting Time_
Counting Rate_
Total Counts
Calibration Standards
Results of Calibration Check Samples_
-------
LABORATORY XRF PARAMETERS - DUST
Sample Quantity_
Sample Preparation
Instrument
Description of X-ray Source
Description of Secondary Target_
Description of Detector
Reference
-------
LABORATORY XRF PARAMETERS - DUST
.Counting Time
Counting Rate_
Total Counts
Calibration Standards — CIN 1 and BAL 1
Results of Calibration Check Samples
-------
EPA/RTI Round Robin for Lead In Paint and Dust
Round Robin No. 002
Lab ID No.
Digestion Method N/A
Experience with this Method
Analysis Method Lab XRF
Experience with this Method
Laboratory HA state
years Laboratory Institute
Approval Signature:
years
Sample ID No.
P-1
P-2
P-3
P-4
P-5
D-1
D-2
D-3
D-4
D-5
Reagent Blank
Gross Concentration of Lead (ppmj
Aliquot 1 Aliquot 2
N/A
N/A
N/A
N/A
-------
Appendix D-3
RTI Copy of Data Reporting Form
with Sequence Tracking
-------
EPA/RTI Round Robin for Lead In Paint and Dust
Round Robin No. 002
Digestion Method N/A
Lab ID No.
Experience with this Method
Analysis Method Lab XRF
Experience with this Method
years
Laboratory MA state
Laboratory Institute
Approval Signature:
Sample ID No.
P-1 -
P-2 -
P-3 -
p-4 -
P-5 -
D-1 -
D-2 -
D-3 -
D-4 -
D-5 -
Reagent Blank
Gross Concentration of Lead (ppm)
Aliquot 1 Aliquot 2
N/A
N/A
C/A/ - 1
N/A
N/A
-------
EPA/RTI Round Robin for Lead In Paint and Dust
Round Robin No. 002
Lab ID No.
Digestion Method EPA/AREAL . Laboratory wr occupational
Experience with this Method years Haith laboratory
Analysis Method ICP
Experience with this Method
Approval Signature:
years
Sample ID No.
P-1-4
p.2-q
P-3 'M-
p.4-4-
P-5-M-
Gross Concentration of Lead fppm)
Aliquot 1 Aliquot 2
D-2- V
D-3-4-
D-4- ^
D-5- 4-
Reagent Blank
-------
EPA/RT1 Round Robin for Lead In Paint and Dust
Round Robin No. 002
Lab ID No.
Laboratory CTs>1c
Digestion Method NIOSH 7082
Experience with this Method years Laboratory
Analysis Method Icp
Experience with this Method
Approval Signature:
years
Sample ID No.
P-1 -M
P-2 -
-------
Appendix E
Reported Results
-------
Appendix E-l
MW/AAS Laboratories
-------
EPA/RT1 Round Robin for Load In Paint and Dust
Round Robin No. 002 Lab ID No. J£.
Digestion Method EPVAKEAL Laboratory
Experience with this Method < / years
Analysis Method ** Aoweval Stanature/
Experience with this Method _££. years _ .
Gross Concentration of Lead (pom)
Sample ID No. AltouolJ_ AHouot 2
P-1 ISDO
P-2 ItH'OOP M^OQO
P-3
P-4 23*0 ^6&b
P-5
0-1
D-2* 40 <\\
D-3 HOP tldO
D-4^
D-5 5100
', /Iff
Reaoent Blank
-------
XAS INSTRUMENT PARAMETER SHKET
Kidth
Background Corr«utton_
lnt»rt«r«nc« corr«otion_
He i
Light source
Typ«_
/
calibration St«nd«rdi «nd calibration Ch«CR» r^**^ 5/5/
-------
EPA/RTI Round Robin for Lead In Paint and Dust
Round Robin No. 002
Lab ID No. 11
Digestion Method EPA/AREAL
Experience with this Method
Analysis Method AA
Experience with this Method
years
years
Laboratory
Signature:
Sample ID No.
P-1
P-2
P-3
P-4
P-5
D-1
D-2
D-3
D-4
D-5
Reagent Blank
Gross Concentration of Lead (pom)
Aliquot 1 Aliquot 2
I5HO
20*53
IOU)
4^30
3.Q
?.OtO
14.40
JliL
((40
103
45oO
-------
AAS INSTRUMENT PARAMETER SHEET
Instrument rnttQ^ EslttAgg^ 3oBO "B
(Manufacturer/Model)
Have length/Si it Width '2-\>.Q V\M P."? Arts. 3>b
Background Correction^
Interference Correction t>UUXA gflUftA Olj»U ilUWV.
Light Source
Flame Type
7 ^X^i^lfJto
Calibration Standards and Calibration Checks
1.0 . ^.0 , 6~.Q , 10-0
I '
(-0
-------
EPA/RTI Round Robin for Lead In Paint and Dust
Round Robin No. 002
Lab ID No. 12
Digestion Method EPVAREAL Laboratory
Experience with this Method / years
Analysis Method AA
Experience with this Method
Approval Signature:
years
Sample ID No.
P-1
P-2
P-3
P-4
P-5
D-1
D-2
D-3
D-4
D-5
Reagent Blank
Gross Concentration of Lead fppm}
Aliquot 1 Aliquot 2
2.IC5
451
O. 39
22.Q0
-------
AAS INSTRUMENT PARAMETER SHEET
Instrument
PC
(Manufacturer/Model )
/Slit Width **&3.3 nm
(P,"? S,l.'f ^hifll,^
Correction J^i
ce Correction nOrtC,
ce Pi> Koljowj Co^-^iode
AiA/Ke-me/^
Calibration Standards and Calibration Checks 0^1 5 I (£ 2-@ WnL- ' *
.Q u*
-------
EPA/RT1 Round Robin for Ltad In Paint and Dutt
Round Robin No, 002
Ub ID No. 13
Dlgtftlon Mtthod EPVMEM, Laboratory
Experience with this Method O years
Analyeli Method **
Experience with thlt Method
Sample ID No.
P-1
P-2
P-3
P-4
P-5
D-1
D-2
D-3
D-4
D-5
Reaocnl Blank
ae
'
Approval Signature:
ytan
Gross Concentration of Lead
Aliquot 1
1110
11*4000
two
Allouot 2
HJ200
'2JOIO
f/30
-------
AAB INSTRUMENT PARAMETER SHUT
Instrument
(Manufacturer/Modal>
Wav»Ungth/«lit Width
Background Correction
Int*r£«r«nc« Correction
Light Souroa-
", **
calibration Standard* and Calibration chaeka
iJ . I . 3, /U,
-------
EPA/RTl Round Robin for Lead In Patnt end Dust
Round Robin No. 002
Analysis Method **.
Experience with this Method
years
Lab ID No. _J1
Digestion Method EPA/AREAL Laboratory
Experience with this Method years .
AoDrowal Slanatura:
Gross Concentration of Lead toomi
gampie ID No.
P-1
P-2
P-3
P-4
P-5
0*1
D-2
D-3
D-4
D-5
Aliauot 1
1773
10941*
38312
1576
35498
5022
196
1292
97
4797
Ailauot 2 ,
1669
127416
36048
1522
36621
4210
177
1277
87
4686
Reaaerrt Blank
-------
AAS INSTRUMENT PARAMETER SKBBT
HI08B 7082
inatrumont
(Manufacturer/Mod*!)
Wavelangth/Slit Width 283.3/4
Background Correction .
Interference Corr«otion
Light Bouroa Bo11n« rarVin/f^ (iqiqp f
Type lean Air- C2H2
Calibration Standards and calibration cb«ok«_
Btandardat 0.5, 1, 2, 5. 10, 20, 40 ppa
Calibration checks> One of.the above after every 6th tample
plotted on linear squar* calibration curve.
-------
EPA/RTI Round Robin for Lead In Paint and Dust
Round Robin No. 002 Lab ID No.
Digestion Method EPA/AREAL Laboratory
Experience with this Method O years
.}
Analysis Method *A _ Approval Sianature:
Experience with this Method 4*5 years __, T _ .
Gross Concentration of Lead (ppm>
Sample ID No. Aliquot 1 Aliquot 2
P-1 Q130-
P-2 /so, OOP . /a. 7, OOP.
P-3 _ J" 3700 - 42 300.
P-4
P-5 4/600- 40 a. 00.
D-1 4?20- 54-50.
D-2 99- 105.
/3oo.
D-4 /6^. 97.
D-5 &/$Q. 47 TO
Reagent Blank OO*0
OO.O
-------
AAS INSTRUMENT PARAMETER SHEET
instrument
SpeCttAA
(Manufacturer/Vtodel)
Wavelength/Slit Width *£%.) 7*^h>rv J /«O/im
Background Correction & *U.-fcrU^ PJ>
Flame Type A\ 1 - A <&j-U l
Calibration Standards and Calibration Checks C A& blfAI^OTl SfQ n QT^ Wvg
AA/y/fcA . Pfe>
Ml s T srt> c/C s-fa n 0(5) r o'
-------
EPA/RTI Round Robin for Lead In Paint and Dust
Round Robin No. 002 Lab ID No. *&
Digestion Method EPA/AREAL Laboratory
Experience with this Method O years
Analysis Method M Approval Signature:
Experience with this Method 2.Z years
Gross Concentration of Lead fopm)
Sample ID No. Aliquot 1 Aliquot 2
p'1 J/ȣQ //2Q
P-2 __
P-3
P'4 jgjD5O 1,740
p'5 4^.7CO 43.£ft£>
D-t
°-2 130.
D-3 i
D'4 /4C\ /4A
D'5 4. feftft. ^".04^
Reagent Blank
-------
AAS INSTRUMENT PARAMETER SHEET
inatrument Per km - Elmer Mode.) 5"QOO AA5
(Manufacturer/Model)
Wavelength/Slit Width £63 . 3 n "1 / Q.T H Ol
Background Correction_
Interference Correction
Light source Pe^k-m ^El^gr 3In'|-eAs.r'\"ron holloa C£rfko^i>.
Flame Type_
Calibration Standards and Calibration Checks
-------
Appendix E-2
HP/AAS Laboratories
-------
EPA/RT1 Round Robin for Lead in Paint and Dust
Round Robin No. 002
Digestion Method N/O5H
Experience with this Method.
Analysis Method frA
Experience with this Method
Lab ID No. 20
Laboratory
years
Approval Signature:
years
Sample ID No.
P-1
P-2
P-3
P-4
P-5
D-1
D-2
D-3
D-4
D-5
Reagent Blank
Gross Concentration of Lead fppm)
Atiquot 1 Aliquot 2
?> I
/<*/
/ 0 0 0 O
-------
AAS INSTRUMENT PARAMETER SHEET
Instrument tKA "3 W ££.£// /}$ /f l/JS/tn
( Manufacturer/Model)
Wavelength/Slit Width_
Background Correction_
3x3 1 ?) 0*
Interference•Correction
Light Source /£y?2> fa {/***>
},'& /$
Flame Type fr/ & / // d L
Calibration Standards and Calibration Checks ^f/P.f f r*i $/)
( o*j ox, /0.< *<>t
-------
EPA/RTI Round Robin for Lead in Paint and Dust
Round Robin No. 002
Lab ID No. 21
Digestion Method NIOSH 7082 Laboratory,
Experience with this Method 0^0 years
Analysis Method M
Experience with this Method
Aooroval Signature:
years
Sample ID No.
P-1
P-2
P-3
P-4
P-5
D-1
D-2
D-3
D-4
D-5
Reagent Blank
Gross Concentration of Lead fopm}
Aliquot 1 Aliquot 2
noto
bOO
103O
III?
130
1,1
Oo
tiio
loo
-------
AA5 INSTRUMENT PARAMETER SHEET
Instrument
(Manufacturer/Modal)
Wavelength/Slit Width Q~ I / &
Background Correction_
Interference Correction
Light Source
Flanie Type
Calibration standards and Calibration Checke
fjf>k Qc, tia^b Vks I?W film
-------
EPA/RTI Round Robin for Lead In Paint and Duet
Round Robin No, 002
Ub ID No, 22
Digestion Method NIOSH 7082 Laboratory,
Experience with this Method ^/°._yean
Analysis Method AA
Experience with this Method .*/? years
Approval Signature:
Sample ID No.
P-1
P-2
P-3
P-4
P-5
D-1
D-2
D-3
D-4
D-5
Reaoent Blank
Groat Concentration of Lead (ppmj
.Aliquot 1 AllQUQt 2
3SO/0
o.oo
o.oo
o.oo
3355*0
3V/VO
100
Hszo
-------
AA8 INSTRUMENT PARAM3ST8R SH1ET
TJl «.<» JW)I Asfc \A
-------
EPA/RT1 Round Robin for Lead In PaJnt and Dust
Round Robin No. 002
Lab ID No. 23
Digestion Method NTOSH 7082 Laboratory.
Experience with this Method I _ years
Analysts Method AA
Aooroval Sionature:
Experience with this Method i years
Sample ID No.
P-1
P-2
P-3
P-4
P-5
D-1
D-2
D-3
D-4
D-5
Gross Concentration of Lead fppm)
Aliquot 1 Aliquot 2
Kt*
tun
Ktrf
IM4
kut
vt
ion
IMC 5 I, "3
»irs- riB-w-
+G&-UO
Reagent Blank
-------
AAS INSTRUMENT PARAMETER SHEET
inatrument
(Manufacturer/Model)
Wavelength/Slit Width
Background Correction_
Interference Correction
Light Source
Flame Type
Calibration Standards and Calibration Checks
-------
EPA/RTI Round Robin for Lead In Paint and Dust
Round Robin No, 002 LablDNo.Jil
Digestion Method NIOSH 7082 Laboratory,
Experience with this Method 5 years
Analysis Method M AoorovaJ Signature:
Experience with this Method i3 years ....-_
Gross Concentration of Lead fooml
Sample ID No: Allouot 1 Aliquot 2
' P-1 1510 1.790
/ p-2 102.000 JlltOOO
' p"3 aa.sno . 39.SQQ
1.940 1,790
' P"5 36.900 41.600
s D-1 3,990 4,390
/ D-2
/ D-3
/ 0-4
^ 0-5
Reagent Blank *TQQ
^-100
<100 140
1.130 1.240
108
4603
171
5.710
-------
AAf INSTRUMENT PARAX1TI* 8HKXT
Xn*trum«nt
(Manufaoturcr/Xod*!)
Wav«l«ngth/«lit Midth J/t^.O
Background Correct ion Afrfrjk. 6sur6-&jOM
Int«rf*r«nc« Correction
Source
Typ« i/i /
Calibration fttandarda «nd Calibration Ch«ckt
0 /V>»M , / pp/n &a*m
am
-------
EPA/RTI Round Robin for Lead In Paint and Dust
Round Robin No. 002
Lab ID No. 25
Digestion Method NIOSH 7082 Laboratory,
Experience with this Method b years
Analysis Method AA,
Approval -Signature:
Experience with this Method \O years
Sample ID No.
P-1
P-2
P-3
P-4
P-5
D-1
D-2
D-3
D-4
D-5
Reagent Blank
Gross Concentration of Lead fopm)
Aliquot 1 Aliquot 2
1-3 \o
1C
C>
51
o
-------
AAS INSTRUMENT PARAMETER SHEET
Inatrument
Interference Correction n.OV\
Light Source f-jc|(cuJ Co.^-K.6fj-P /•
Flame Type f\\r
(Manufacturer/Model)
Wavelength/Slit Width "ft •- ^Hnrx 0.1
Background Correction i Jg'Lv'i 6 <* i
Calibration Standards and Calibration Checks CdUb. S.4Js, I -^. ^ (t .
T f f / f
-------
EPA/RTI Round Robin for Lead In Paint and Dust
Round Robin No. 002
Lab ID No. 26
Digestion Method NIOSH 7082 Laboratory
Experience with this Method / years
Analysis Method M
Experience with this Method
Approval Signature:
years
Sample ID No.
P-1
P-2
P-3
P-4
P-5
D-1
D-2
D-3
D-4
D-5
Gross Concentration of Lead topm>
Aliquot 1 Aliquot 2
1100 /<
/7-6'Q
36, C c'C'
oo
lie
. C'OO
oc
Reagent Blank
, C-
-------
Instrument
AAS INSTRUMENT PARAMETER SHEET
I OO ~
(Manufacturer/Model)
Wavelength/Slit Width --? / ?• / ^>-» / 0.
Background Correction A i'
Interference Correction
Light Source L&&£- f~k- 11 <•**-> (txlW^cCc. / c<-4?^ LA.f*-/O
Flame Type .^/^ - /? GL
Calibration Standards and Calibration Checks ^.~h^..c(Af£(^ Ci ~? *»-,\gf
/cc
tvt pi' t'u.kL. \ if^i Ct-f: ijt-L. "hi*. u-'6$ c^j.e.Lc^ gc'g.-c/ S'
— • —••- -...-- •- ----• ..-- -- — ..... -..
-------
EPA/RTI Round Robin for Lead In Paint and Oust
Round Robin No. 002
Lab ID No. 27
Digestion Method NIQSH
Experience with this Method,
years
Laboratory
Analysis Method
Approval Signature}
Experience with this Method
(13 ye«r* experience with
Samole ID No.
p-1
P-2
P-3
P-4
P-5
D-1
D-2
D-3
D-4
D-5
o years
AA method)
Gross Concentration
Aliquot 1
1542
93,532
37,699
1,805
37160
4567
109
1199
109
5096
of Leatf (pom)
ARgupt 2
2096
99,463
35.974
1.879
37002
5014
HI
1207
140
4071
Reagent Blank
-------
XA8 INSTRUMENT PARAMBT*R SHEIT
NIOSH 7082
Inatrument F>rVin
(Kanufaoturer/Kodal)
Wavelength/Slit Width 283.3/4
Background Corr«ation_
Interference Correction
Light Source HolTm*
Typ^ lean Air- C2H2
calibration Standard* and Calibration chaok»_
StandardBi 0.5, 1,_2. 5, 10, 20, 40 ppa
Calibration cbeckt; Qg« of.th« above after every 6tb
Standards/Sanplea plotted on linear square calibration curve.
-------
EPA/RT1 Round Robin for Lead in Paint and Dust
Round Robin No. 002 LablDNo._28_
Digestion Method NIOSH 7082 Laboratory
Experience with this Method I years
Analysis Method AA Approval Signature:
Experience with this Method Z2/ years
Gross Concentration of Lead fppm>
Sample ID No. Aliquot /L Aliquot 2
P-1
Tf^o&p ti3>c&&
P-2 "iKJP^ "'*'
P-3
P-4
p-5 3ftj 7QQ
D-1
D'2 < 3OO < v
°-3 UftQ (32Q
D-4 ^j
D-5
Reagent Blank
-------
AAS INSTRUMENT PARAMETER SHEET
instrument Pfcrfcm Elmer Modc.1 *S"OOO A AS
(Manufacturer/Model)
Wavelength/Silt Width £&3.3 f)l*\ I O.7 Hm _
Background Correction
Interference Correction
Light Source rCr^Cm-
Plame Type
Calibration Standards and Calibration Checke 2 j S} fC> ; Qnn JZft LL
-------
Appendix E-3
MW/ICP Laboratories
-------
EPA/RTI Round Robin for Lead In Faint and Dust
Round Robin No. 002 Lab ID No. 30_
Digestion Method t?A jfitfAL- Laboratory
Experience with this Method o years
Analysis Method TCP Approval Signature:
Experience with this Method ST* „ years .
Gross Concentration of Lead fppm)
Sample ID No. Aliquot 1 Aliquot 2
P-1 _
P-2
P-3 j(>/0g. }\'t,oo.
P-4
P-S
I>1
D-2
D-3
°'4 tit '<>*'
D-5
Reagent Bfank
-------
EPA/RT1 Round Robin for Lead In Paint and Dust
Round Robin No. 002
Digestion Method EPA/AREAL
Experience with this Method __;
Analysis Method ICP
Experience with this Method
Lab ID No. 31
Laboratory ,
years
Approval Sianakira:
years
Gross Concentration of Lead
Sample ID No.
P-1
P-2
P-3
P-4
P-5
D-2
D-3
D-4
D-5
Reagent Blank
Aliquot 1
Afiauot 2
\cw i.tzc
l^fr.c-C
ii i *\ if" L
/j Si ?» P
/
/a"2> t> o o
>>3 00 0
j ^
* % -\ - '~f s\ ^
.rvr
u
1
1060
L£/ rr(?j/€
O /-. ^J I /
O.C")
-------
Instrument
ICP PARAMETER SHEET
f S .3. tT'PiP
(Manufacturer/Model)
x^—
Nebulizer Ut I Jt> t fi-A \.'t) (* t-\ 0
ff\
Wavelength ,> // - .
Cc l\ e.
Resolution 6* UP ~] ->
Focal Length
'• ix' .v*.M?.l£_
Background Correction Y *~
Interference Correction
/ t!
/
1C _^_
Forward Power . - -v- : y
Reflected Power.
Plasma Frequency
/ ^ . 'J V
'_AL
t "3
Sample Introduction Rate
Auxilliary Gaa Flow Rate
/. ^ -Y*J_
|
W-
Aig Otf/S *?$*! *- ft/A c
-------
EPA/RTI Round Robin for Lead In Paint and Dust
Round Robin No. 002 Lab ID No. 32
Digestion Method EPA/ARFAL Laboratory
Experience with this Method O years
Analysis Method ICP Approval Signature:
Experience with this Method 1 years . .
Gross Concentration of Lead fppm)
Sample ID No. Aliquot 1 Aliquot 2
P-1 _
P-2
p-3 3^400
P"4 /37O /6>OO
p'5 3SSOO 35400
D-1 __
D-2 /c 7
D-3 __
D-4
D-5
Reagent Blank
-------
Instrument
ICP PARAMETER SHEET
(Hanuf acturer/Model )
Nebuliser
Wavelength 1.2.0 » b C\V\
Reaolution H. K^Ci^ I K H
Focal Length
Background Correction \— Q <\^ KV^gV ' \\
Interference Correction A\ . LC v V(L . ' >
----- ---- ^ "~"~" ~"" " "
Forward Power
\ J> O y
Reflected Power
Plasma Frequency c. JU | v. I \ i>
Auxilliary Gao Flow Rate
,1 I/,
I r- , /
Sample Introduction Rate \ . S fn\ I / f*\
Calibration Standards and Calibration Check Samples
Jl
6" f ^-t Tvr Z H Co ^rcft-p e /
H ^
-------
EPA/RTI Round Robin for Lead In Paint and Dust
Round Robin No. 002
Analysis Method Icp
Experience with this Method a years
Lab ID No.
Digestion Method EPA/AREAL Laboratory
Experience with this Method 4 years ^
Approval Signature:
Sample ID No.
P-1
P-2
P-3
P-4
P-5
D-1
D-2
D-3
D-4
D-5
Gross Concentration of Lead fppm)
Aliquot 1
1432.
109400.
34000.
1518.
32400.
4160.
87.
1142.
145.
3960.
Aliquot 2
1408.
109600.
34100.
1502.
32600.
4170.
89,
1104.
98.
3960.
Reagent Blank
< 10.
-------
ICP PARAMETER SHEET
instrument Leeman Labs Inc.; [CP/PS 1000
(Manufacturer/Modal)
Nebu 11 zer 42 PSI
Wavelength 220.353
Grating Fixed Echelle Grating
Resolution
Focal Length_
Background Correction 220.330
interference correction Inter Element Correction for Aluminum
Forward Power 1.1 kW
Reflected Power
Plasma Frequency 40.68 mHz
Auxilliary Gas Flow Rate .00 LPH
Sample Introduction Rate 1.7 ml/min
Calibration Standards and Calibration Check Samples,
Calibration Standards: 0.0. 0.5. 3.0. 10.0. 30.0. and 100.0 PPM
Calibration Check Standards: 0.0. ~2.Q and 100.0 PPH +/- 10 t
run after every 10 samples.
-------
EPA/RTI Round Robin for Lead In Paint and Dust
Round Robin No. 002
Lab ID No. 34
Digestion Method EPA/AREAL Laboratory
Experience with this Method years
Analysis Method TCP
Experience with this Method 6 years
Approval Signature:
Gross Concentration of Lead toom)
Sample ID No.
P-1
P-2
P-3
P-4
P-5
D-1
D-2
D-3
D-4
D-5
Aliquot 1
1,500 ppm
117,000 ppm
35.200 ppm
33.700 ppm
4,070 ppm
80 nnm
1.170
17ft
a 110 ppm
Aliquot 2
1.880 ppm
120.000 ppm
3 6,7 0 0 ppm
1,830 ppm
35.200 ppm
4,960 ppm
140
-PPB-
1,\fifi ppm
110 ppm
3.9QQ pnm
Reagent Blank
ug/sample
-------
ICP PARAMETER SHEET
instrument Jarrell-Ash 9000 Air Spectrometer
(Hanufacturer/Hodel)
Nebuiixer Fixed Cross Flow
Wavelength 2203.00
1516 groves/mm ruled grating at 500 nm
Resolution *045 nm' First Order, .023 nm, Second Order, .015 nm Third
Focal Length Focal curve is 580 ntn in length .
Background Correction No t
Interference Correction Yes Fe, Mg, Al
Forward Power 1.2 Kilometer
Reflected Power
Plasma Frequency_
Auxilliary Gas Flow Rate 22 LPM
Sample Introduction Rate 2.7 ml per min
Calibration Standards and Calibration Check Sarople«__
Fisher Lead Reference Solution 1,000 ppm + 1%
NIST Re,erence std OC 3,6,
Lead Reported 40 ug/f
NIST Reference Std QC 34370 Actual 37.8 ug/f
-------
EPA/RT1 Round Robin for Lead In Paint and Dust
Round Robin No. 002
Lab ID No. 35
Digestion Method EPA/AREAL Laboratory
Experience with this Method years
Analysis Method ICP
Experience with this Method
Approval Signature:
years
Sample ID No.
P-1
P-2
P-3
P-4
P-5
D-1
D-2
D-3
D-4
D-5
Reagent Blank
Gross Concentration of Lead fopm)
Aliquot 1 Aliquot 2
|5VO
I l
-------
ICP PARAMETER SHEET
Instrument
(Manufacturer/Model)
Nebulizer Ay^d
Wavelength
Grating ho I 0 f
Resolution
Focal Length
Background Correction_
Interference Correction
Forward Power
Reflected Power
Plasma Frequency
(X
Auxilliary Gas Flow Rate j__
Sample Introduction Rate ' * I
Calibration Standards and Calibration Check Samples
Tie*
-------
EPA/RTI Round Robin for Lead In Paint and Dust
Round Robin No. 002
Digestion Method EPVAREAL
Experience with this Method i MON
Analysis Method *CP
Experience with this Method
years
Lab ID No. 36
Laboratory.
Approval
Gross Concentration of Lead fppm)
Sample ID No.
P-1
P-2
P-3
P-4
P-5
D-1
D-2
D-3
D-4
D-5
Reagent Blank
Aliauot 1
1600
116,000
35,800
2120
39,400
4260
126
1220
88
4720
0
0
Aliquot 2
1400
115,000
35,000
1590
37,600
3940
98
1150
98
5360
-------
ICP PARAMETER SHEET
Instalment Perkin Elmer 6000 ICP
(Manufacturer/Model)
Nebulizer Cross Row
Wavelength 220.353 nm
Grating UV grating - holographic
Resolution 0.001 nm
Focal Length 408mm
Background Correction Yes
Interference Correction Yes
Forward Power 1.20 Kilowatts
Reflected Power Less than 5 watts
Plasma Frequency 27.12 MHz ISM Band
Auxiliary Gas Row Rate 0.6 L/mln
Sample Introduction Rate 1.1 mL/mln
Calibration Standards and Calibration Check Samples
1. Calibration standards - 0.00 and 10,00 ug/mL
2. Check samples - 0.00 and 10.00
ug/mL
3. The 0.00 ug/mL check sample could not drift beyond ±0.05 ug/ml and the 10.00 ug/mL sample
beyond 5% (9.50 and 10.50 ug/mL).
' -™-""™1™"1"--- - - f • T..-J ......".. iir .: ......_...-•••_..-.. T/....V.-T ----......
4. A manually plotted line using 0.00.0.50. 3.00, and 10.00 ug/mL standards resulted in an r square
value of 1.000.
-------
EPA/RTt Round Robin for Lead In Paint and Dust
Round Robin No, 002 Lab ID No. 37.
Digestion Method EPA/ARE&L Laboratory,
Experience with this Method o years
Analysis Method *CP Approval Signature:
Experience with this Method 8 years ...
Concentration of Lead fapml
Sample ID No. Aliquot 1_ Aliquot 2
P-1
P-2
P-3
P-4
P-5
D-1
D"2 *•* 1 l & I
D-3 J/«f2.
D-4
D-5
Reaoent Blank
10
-------
ICP PKRAMBTKft SHEET
instrument
flKL.
(Manufacturer/Mod*!)
Plr
"" "
Orating ___/
FOC»I Lttnoth I /ngrgo ffr^ fit r D. ob&X Pff*.
forward fowar (o 5*0 {J/tT7% (#i*/lTb6c.rt (9
R«fl*ct«d pewar
Plmem* Frequency ^. > . )^ A\ H a
Auxilliaxy cm* Flow R«t»
sample Introduction Hata I -O
Calibration Standardi and calibration Check s*mple» F^A^r <0. £ 3
Qf
-------
EPA/RT1 Round Robin for Uad In Paint and Dust
Round Robin No. 002
Lab ID No, 38
Digestion Method EPA/AREAL Laboratory,
Experience with this Method O years fc
Analysis Method ICP
Experience with this Method
Approval Signature:
years
Sample ID No.
P-1
P-2
P-3
P-4
P-5
0-1
D-2
D-3
D-4
D-5
Reaoent Blank
Gross Concentration of Lead foom)
J_ Aliquot 2
vxoooo
HI- Oft
-A
y>^ AJkj^. _
\ l>cooo
3*000
\%00
000
vaoo
3V
O.
-------
ICP PARAKETXK SHEET
Instrument
(Manufacturer/Model>
Nebuliser.
Wftvelength_
Grating
-v^A
Raaolutlon
Focal Length
Background
Int«rfwr*nc» COrraction
Forward Power N^LQ 0
Reflected Powar
\
Plasma Frequency
Auxilliary Oa« Flow Rate \\. ^> t\ 3i«N •cAe^rv^ *++
Sample Introduction Rate
Calibration Standards and calibration Check Samplee_
0. \Q
-------
Appendix E-4
HP/ICP Laboratories
-------
EPA/RTJ Round Robin (or Lead In Paint and Dust
Round Robin No. 002
Lab 10 No.
Digestion Method
Experience with this Method 0 yearn
Laboratory
Analysis Method |pp
Experience with this Method
Approval Signature:
years
Sample ID No.
P-1
P-2
P-3
P-4
P-5
D-1
D-2
D-3
D-4
D-5
Reaoent Blank
Gross Concentration of lead fppmi
Allouot 1 Atlouot 2
\D\0-
flJ'6.
-------
EPA/RTI Round Robin for Lead In Paint and Dust
Round Robin No. 002
Lab ID No. 1L
Digestion Method NIOSH 7082 Laboratory;
Experience with this Method 2. years
Analysis Method ICP
Approval Signature:
Experience with this Method 5 years
Sample ID No.
p-1
P-2
P-3
P-4
P-5
Gross Concentration of Lead fopml
Aliquot 1
1.151
41.30
I.S12
Aliquot 2
l.fcES
1(1.62
it.SH
0-2 o k (* n
D-3 0 C? k> / 8
D-4 6G (o H
D-5 OC
Reagent Blank
3.
. loo
-------
ICP PARAMETER SHEET
Instrument
(Manufacturer/Model)
1 A
Nebulizer
Wavelength 120,353
Grat ing KuJfc j
Resolution
Focal Length
Background Correction A/ct ntfCgSSo.ru.
Interference Correction
Forward Power 471
Reflected Power 001 V\jL|mtn.
Sample Introduction Rate 2.b
Calibration Standards and Calibration Check Samples^
SOL m
plc
c*vi
BlANK,
Ik II w.nfu^ tHe*i
-------
EPA/RTI Round Robin for Lead In Paint and Dust
Round Robin No. 002
Lab ID No. 42
Digestion Method NIOSH 7082 Laboratory,
Experience with this Method 4 years
ICP
Analysis Method
Experience with this Method 3 years
Approval Signature:
Sample JD No.
P-1
P-2
P-3
P-4
P-5
D-1
D-2
D-3
D-4
D-5
Gross Concentration of Lead (ppm)
Aliquot 1
1790.
119000.
34500.
U5Q.
34500.
4060,
93.
1120.
74.
4220.
Aliquot 2
1615.
115000.
34700.
1630.
34100.
4460.
108.
1100.
90.
4110.
Reaoent Blank
< 50.
-------
ICP PARAMETER SHEET
Iceman Labs Inc.; ICP/PS 1000
(Ha nu f actu rer/Mode1)
Nebuliier 42 PSI
Wavelength 220.353
orating Fixed Echelle Grating
Resolution
Focal Length_
Background Correction 220.330
interference correction Inter Element Correction for Aluminum
Forward Power 1.1 kW
Reflected Power
Plasma Frequency 40.68 fflHz
Auxilliary Gas Flow Rate .00 LPH
Sample Introduction Rate 1.7 tnl/min
Calibration Standards and Calibration Check Sample«_
Calibration Standards: 0.0; 0.5; 3.0; 10.0: 30.0: and 100.0 PPM
Calibration Check Standards: 0.0; ?.Q; and 100.Q PPH 47- 10 t
run after every 10 samples.
-------
EPA/RTI Round Robin for Lead In Patnt and Oust
Round Robin No. 002
Analysis Method ICP
Experience with this Method II years
Lab ID No. 43
Digestion Method NIOSH 7082 Laboratory
Experience with this Method 3 years
Approval Sianalure:
Sample ID No.
P-1
P-2
P-3
P-4
P-5
D-1
D-2
D-3
D-4
D-5
Reagent Blank
Gross Concentration of Lead fppm)
Aliquot 1 Aliquot 2
I/O.
37f
. Of 3
- o.ol
0-02-
, 8?o
/.
loo
/
4-,
-------
ICP PARAMETER SHEET
Instrument
(Manufacturer/Model)
Nebulizer C rg&S • plotO /O-g-kyKPt'jl
Focal Length
Background Correction
Interference Correction
Forward Power
Reflected Power
< IO.O
Plasma Frequency
Auxilliary Gaa Flow Rate Qt. <3.
Sample Introduction Rate
/*
jl H
Havelength
Crat ing U^ Q/la/fc-U* ^ "~ £2 %Q jLswLS /7t/X< / l/JSik/l>
Resolution 0 i
Calibration Standards and Calibration Check Samples / O <. &&
*- Co.//
, r>r.
H ft
-------
EPA/RTI Round Robin for Lead In Paint and Dust
Round Robin No. 002 Lab ID No. 44
Digestion Method NIOSH 7082 Laboratory,
Experience with this Method 10 years
Analysis Method Icp Aearoval Signature:
Experience with this Method /% years ^
P-4 /.TO 6
p'5 3VC.OO 33^00
D-1 s~& Qt (*>
D-3 /,
-------
ICP PARAMETER SHEET
instrument ^TA^K^LC -#5't AT Q/*i<~.o sn £>
(Manufacturer/Model)
Nebu 1 izer/^y/rp f ROSS Fi.
Wavelength 2
Grat ing j/^/"/?^/.^ •y?^'/y / v// £> t. /} f
Resolution
Focal Length /
Background Correction_
Interference Correction
Forward Power /, / 5" /C (A/
Reflected Power
Plasma Frequency % 7» /'3
Auxilliary Gas Flow Rate
Sample Introduction Rate /* £> "7 w /, /sy>,
Calibration Standards and Calibration Check Samples
-------
EPA/RTI Round Robin for Lead In Paint and Dust
Round Robin No. 002
Lab ID No. 45
Digestion Method NIOSH 7082 Laboratory
Experience with this Method years
Analysis Method _ ICP
Approval Signature:
Experience with this Method 6 years
Sample ID No.
P-1
P-2
P-3
P-4
P-5
D-1
D-2
D-3
D-4
D-5
Gross Concentration of Lead fopm)
Aliquot 2
1,810 ppm
Aliquot 1
1,720 ppm
115.000 ppm
36.900
1 Q & fl nnm
37,000
4,170 ppm
270 com
-ppniL.
A f\
94,700 ppm
37.400 ppm
1tQQO ppm
36i400 ppm
4 , 750 ppm
150 com
1 1 A fl
•PP"
5 o ppm
-pp«
Reagent Blank
10.1 ug/sample
-------
ICP PARAMETER SHEET
instrument Jarrell-Ash 9000 Air Spectrometer
(Manufacturer/Model)
Nebulizer Fixed Cross Flow
Wavelength 2203.00
1516 groves/mm -ruled grating at 500 nm
Resolution <045 nm> Flrst Order, .023 nm. Second Order, .015 nm Thir<
Focal Length Focal curve is 580 nfrp in length
Background Correction_
No
Interference Correction Yes Fe. Mg, M
Forward Power 1.2 Kilometer
Reflected Power
Plasma Frequency_
Auxilliary Gae Flow Rate 22 LPM
Sample Introduction Rate 2.7 ml per min
Calibration Standardo and Calibration Check Samplea
Fisher Lead Reference Solution 1,000 ppm + 1%
Lead Reported 58.2 uq/f
NIST Reference Std QC 3469 Actual £6.9 ug/f 9/
Lead Reported 40 ug/f
NIST Reference Std QC 34370 Actual 37.8 ug/f
-------
EPA/RTt Round Robin for Lead In Paint and Dust
Round Robin No. 002
Lab ID No. 46
NIOSH 7082
Digestion Method
Experience with this Method IMON. years
Laboratory,
Analysis Method Icp
Approval Si
Experience with this Method 7 years
Gross Concentration of Lead fppm)
Sample ID No.
P-1
P-2
P-3
P-4
P-5
D-1
D-2
D-3
D-4
D-5
Reaoent Blank
Aliauot 1
1160
84.000
32,000
1280
28,600
3160
160
840
70
3580
Aliauot 2
1280
Not available
30,800
1330
30,200
4110
80
840
70
2670
0
0
-------
ICP PARAMETER SHEET
Instalment Perkln Elmer 6000 ICP
(Manufacturer/Model)
Nebulizer Cross Flow
Wavelength 220.353 nm
Grating UV grating - holographic
Resolution 0.001 nm
Focal Length 408mm
Background Correction Yes
Interference Correction Yes
Forward Power 1.20 Kilowatts
Reflected Power Less than 5 watts
Plasma Frequency 27.12 MHz ISM Band
Auxiliary Gas Flow Rate 0.6 L/mln
Sample Introduction Rate 1.1 mL/mln
Calibration Standards and Calibration Check Samples
t. Calibration standards - 0.00 and 10.00 ug/mL
2. Check samples - 0.00 and 10.00 ug/mL
3, The 0.00 ug/mL check sample could not drift beyond ±0.05 ug/mL and the 10.00 ug/mL sample
beyond 5% (9.50 and 10.50 ug/mL).
4. A manually plotted line using 0.00.0.50. 3.00, and 10.00 ug/mL standards resulted in an r square
value of 1.000.
-------
EPA/RT1 Round Robin for Lead In Paint and Dust
Round Robin No. 002 Lab ID No. 47
Digestion Method NIOSH 7082 Laboratory,
Experience with this Method 5 years
Analysis Method ICP Approval Signature:
Experience with this Method 7 years
Gross Concentration of Lead fppml
Sample ID No. Afiouot 1 Aliquot 2
P"1 1.600 1.500
P~2 110.000 110.000
P-3
r ° -tA.nnn 36.000
°~* 1.700 1.90f>
P'5 36.000 37.000
^"1 4.400 4.500
D"2 110 100
D"3 1.200 1.200
D'4 82 130
D"5 4. SOQ 5.300
Reagent Blank <^Q
"•40
-------
ICP PARAMETER SHUT
inatrument Thermo Jarrell Ash ICAP 9QOO
(Kanufacturer/Model)
Fixe Flow Rate ^ 2 LPM
Sample Introduction Rate 1.75 ml/mtn ;
Calibration Standard* and Calibration Check Sample*
Standardization 9 10 mg/Lj Check Standard 92je/L & 1QQ mg/r
Interference Check Standard g 1 mg/L with interferpnts & ?nn m£/r
-------
EPA/RTI Round Robin for Lead In Paint and Dust
Round Robin No. 002 Lab ID No. _JL
Digestion Method NIOSH 7082 Laboratory
Experience with this Method years
Analysis Method ICP Approval Signature:
Experience with this Method 3*. years , a_
Gross Concentration of Lead fopm)
Sample ID No. Aliquot 1 Aliquot 2
P-t _
P-2 __
P-3 __
P-4
p*5 3f?oo 3C.9/Q
D-1
D-2
D-3
D-4
D-5
Reagent Blank
-------
EPA/RT1 Round Robin for Lead in Paint and Dust
Round Robin No. 002
Lab ID No. 49
Digestion Method NIOSH 7082 Laboratory,
Experience with this Method O years ..
Analysis Method _
Experience with this Method
Approval Signature:
years
Sample ID No.
P-1
P-2
P-3
P-4
D-1
D-2
D-3
D-4
D-5
Reagent Blank
Gross Concentration of Lead fopm)
AHouot 1 ABouot 2
> \HOO
WOO
31-000
3&00
\\OQ
•5SOOO
\>00
\000
-------
ICP PARAMETER SHEKT
Instrument
(Kanuf acturer/Model )
Nebui iter
wav«length_
crating
v\
Resolution
/a^W^. V*A
(
% —••
Focal Length
Background Correct ion
Interference Correction
Forward Power \1LO 0
Reflected Power
Plasm* Frequency_
f *
Auxi.lli.ary Oa« Flow Rate
Sample Introduction Rate
« jiv
V rv»»V. / rtv±~
Calibration standards and calibration check sampl«>_
\0
-------
Appendix E-5
Laboratory XRF Laboratories
-------
EPA/RTI Round Robin for Lead In Paint and Dust
Round Robin No. 002
Digestion Method
Experience with this Method flfoo
Analysis Method Lab XRF
Experience with this Method
Lab ID No. 50
Laboratory
Sample ID No.
P-1
P-2
P-3
P-4
P-5
D-1
D-2
D-3
D-4
D-5
Reagent Blank
Approval Signature:
Gross Concentration of Lead fppnrQ
Aliquot 1 Aliquot 2
5~d 000
/
< 7-5"
N/A
N/A
N/A
N/A
-------
LABORATORY XRF PARAMETERS - PAINT
Sample Quantity /
Sample Preparation
instrumeat
fl
0
~
Description of X-ray Source
C/ tf( /U ^r
Description of Secondary Target_
Description of Detector
Reference
-------
LABORATORY XRF PARAMETERS - PAINT
Counting Tin*
Counting Rat*
Total Count.
Calibration Standard* /'.LA^f^' l*v\.
' /) " A* f J^S '
\« £'f4{'*^' ( //fQp
f frT_V- - \^ / /
Results of Calibration Check Samples_
-------
LABORATORY XRP PARAMETERS - DUST
Sample Quantity
ol
Sample Preparation
Instrument
Description of X-ray Source
ts OL
Description of Secondary Target_
Description of Detector_
Reference
-------
LABORATORY XRP PARAMETERS - DUST
Counting Tima
Counting Rate_
Total Count*
Calibration Standards — CIN 1 and BAL 1
Results of Calibration Check Samples
^-f.
V/
-------
EPA/RTI Round Robin for Lead In Paint and Dust
Round Robin No. 002
Lab ID No. 51
Digestion Method N/A
Experience with this Method,
Analysis Method Lab
Experience with this Method,
Laboratory
years
Approval Siqnature:
years
Sample ID No.
P-t
P-2
P-3
P-4
P-5
D-1
D-2
D-3
D-4
D-5
Reagent Blank
Gross Concentration of Lead fopm)
Aliquot 1 Aliquot 2
i
lCC
23 m.
C.
N/A
N/A
N/A
N/A
/ S
-TL-
-------
LABORATORY XRF PARAMETERS - PAINT
Sample Quantity_
Sample Preparation
- 4JL.
/
instrument
.
* 610O 'iM-t ^ ~l()0d
Description of X-ray Source A?dJLA
Description of Detector //^ fl ^ 6- < r^-\yw- Q iT I pfe f~ $ |
WW OK
Reference
-------
LABORATORY XRF PARAMETERS - PAINT
Counting Time
Counting Rate /MO ft AW > ^ OdO C f
Si
1
Total Counta flKgK- l\ QQ
Calibration Standards VjgS irj 1 ^ tJuJc "k> .€>£% ljj< J4~
Results of Calibration Check Samples
-------
LABORATORY XRF PARAMETERS - DUST
Sample Quantity
Sample Preparation
it 41 i
••—M
(X. S \
instrument
Description of X-ray Source \f\6OlUYP*-
Xtf-F O^CtO li^fctk k^U-PX 16OC)
f\f\6OlUYP*- niVnfe, _ ty$
Description of Secondary Target \2 / / j/' '
\2 / / j
Description of Detector
k?&A tittt*
f\flJ-r't<2.
Reference
< / f^JUuf ^/ L
-------
LABORATORY XRF PARAMETERS - DUST
Counting Time /OQ
Counting Rate /'A(f^ ^lOOO t
."k
Total Counts _ (^iW6^ <-IOC CCQ
Calibration Standards — CZN 1 and BAL 1
f\ft fl/\iS/(
Results of Calibration Check Samples
•JUu*AA T^lov 0 A
-------
EPA/RTI Round Robin for Lead In Paint and Dust
Round Robin No. 002
Lab ID No.
Digestion Method
Experience with this Method
Laboratory
years
Analysis Method Lab
Experience with this Method H years
Approval Signature:
Sample tD No.
P-1
P-2
P-3
P-4
P-5
D-1
D-2
0-3
D-4
D-5
Reagent Blank
Gross Concentration of Lead fppm)
Aliquot 1 Aliquot 2
jj/A
Jj/A
N/A
N/A
, 3>QQ
90k
11 73>
/ 1 00
(100
7?- ~7_T
^. ( C \o ^ 0 0
-------
LABORATORY XRF PARAMETERS - PAINT
Sample Quantity
Sample Preparation
T\O
Instrument_
- ^-
Description of X-ray Source/^vV ~Avvivr*k *Y(Vi
Description of Secondary Target
*V\ «Aa
Description of Detector_
\e<- ^T <3 \^pM<-f^v\S\ \\ fc...ft.r^Ns«^V>. -a^Vrer.
Reference
-------
LABORATORY XRF PARAMETERS - PAINT
Counting Tima / 6 0 5€.C-
Counting Rat«
Total Counts
Calibration Standarda
Resulta of Calibration Check Samples
-------
LABORATORY XRF PARAMETERS - OUST
Sample Quantity
Sample Preparation
Instrument
170
j
Description of X-ray Source rKc4Nv^y^ 3TQ-~.c*> v v ^ w> > \
o - kc k\/ ;
_
Description of Secondary Target
Description of Detector L fV t> 5 rcvf 7 £.^Cr«y^ f>-g£o)c
-------
LABORATORY XRF PARAMETERS - DUST
Counting Time I 0 0
Counting Rate
Total Counts
Calibration Standards — CIN 1 and BAL 1
Results of Calibration Check Samples_
-------
EPA/RTI Round Robin for Lead In Paint and Dust
Round Robin No. 002
Digestion Method N/A
Experience with this Method
years
Lab ID No.
Laboratory,
Analysis Method tab XRF Approval Sianatuie.
Experience with this Method 2 years
Sample ID No.
P-1
P-2
P-3
P-4
P-5
D-1
D-2
D-3
D-4
D-5
Reaoent Blank
Gross Concentration of Lead (ppm>
Aliquot 1
5434
104,510
29 .573
6,003
26.403
2.000
126
113
1.400
N/A
N/A
N/A
N/A
AIJQUOt 2
5148
101,852
27.368
5,823
26,178
2,000
137
01 fi
126
1.9QQ
-------
LABORATORY XRF PARAMETERS - PAINT
Sample Quantity Q.20g (± Q.Qlo)
Sample Preparation
Sample bottle was rotated to insure mixture of sample material,
The 0.20 gram Of Sample was weighed as r;r>mhi r>P»ri uH t-h 1 grat^ nf
cellulose and 1 gram of zinc oxide in a plastic mixing vial
with mixing balls (exact weights were recorded) . The samples
were then mixed for 10 minutes in a shaker mill. After mixing the
samples were pressed into pellets using a Carver press. Each
pellet was pressed to 10,000 Ibs. for a minimum of 5 minutes.
(NISTIR 89-4209)
Instrument Computerized JSnergy Dispersive X-ray FliinrpRr-p-T^r-P System ,
Kevex 7 70 /Delta
Description of X-ray Source Rhodium _
Description of Secondary Target Zirconium
Description of Detector Lithium-dri f frf>ri fi i 1 i CT>T>
Reference
-------
LABORATORY XRF PARAMETERS - PAINT
Counting Time 300 seconds
Counting Rat« See attached Table 1
Total counts See attached Table 1
Calibration Standards NIST SRM1589 11.87 Pb in paint
Result* of Calibration Check Sample»_
-------
LABORATORY XRF PARAMETERS - DUST
Sample Quantity 100%
Sample Preparation
Sample bottles were rotated to insure proper mixing of sample
material. A portion of undiluted sample was placed in a XRF
sample cup with mylar film covering the bottom and microporous
film over the top.
Instrument Computerized Energy Dispersive X—ray Fluorescence
Kevex 770/Delta
Description of X-ray Source Rhodium
Description of Secondary Target Zirconium
Description of Detector Lithium-drifted Silicon
Reference
-------
LABORATORY XRP PARAMETERS - DUST
Counting Time 200 seconds
Counting Rat« See attached Table 1
Total Count* See attached Table 1
Calibration Standards — CIN 1 and BAL 1
Results of Calibration Check Samples
-------
EPA/RTI Round Robin for Lead In Paint and Dust
Round Robin No, 002 Lab ID No. 54
Digestion Method KA . Laboratory,
Experience with this Method. years .
Analysis Method Ltb xs? Approval Signature:
Experience with this Method _3 years
"^ ~r~'
Gross Concentration of Lead fppm)
Sample.ID No. Aliquot 1 Aliquot 2
P-1 _..fl» 762.
P-2
P-3
P-4
p-5 Zl84g
°'1 24i7 2444
0-2 13 IH
D-3
°-4 «e 11 r
o-s 2A\ 6 2424
Beaqent Blank N/A
N/A
N/A
-------
LABORATORY XRF PARAMETERS - PAINT
Sample Quantity,
Sample Preparation
Inetcument
Description of X-ray source
Description of Secondary Target^
Description of Detectot_
iili
Reference
-------
LABORATORY" XRF PARAK2TBRS - PMNT
Counting
Counting Bat*
Total Count, t*M> tlHC
Calibration gtandarda \O
II >- 21*1*7 FfH
Result* of Calibration Check Samples 14O PAtMT
A
-------
LABORATORY XRF PARAMETERS - OUST
Sample Quantity
Sample Preparation
Inatrument
Description o£ X-r&y Source X-WV
Deecriptlon of Secondary Targ«t_
Deocription of Detector_
Reference
-------
tABORATOR* JiRF PARAMETERS - DUST
Counting Timq &C/W
Counting Rate.
r
total Count* TXJAfr 11HC *
Calibration Standards — CZK 1 and BAL 1
Hcsulta of Calibration check Samples
-------
EPA/RTl Round Robin for Lead in Paint and Dust
Round Robin No. 002
Lab ID No. 55
Digestion Method N/A
Experience with this Method
years
Laboratory,
Analysis Method Lab XRF Approval Signature:
Experience with this Method 3.5 year _.
Gross Concentration of Lead
Sample ID No.
P-1
P-2
P-3
P-4
P-5
D-1
D-2
D-3
D-4
D-5
Reagent Biank
Aliquot 1
1006
10.5 Z
31905
973
33982
2489
107
976
81
2441
N/A
N/A
N/A
N/A
Aliquot 2
910
10.4 Z
31228
1021
32388
2458
ai
962
87
2S14
-------
LABORATORY XRF PARAMETERS - PAINT
Sample Quantity 2-2.5 grams
Sample Preparation
(See Attached)
instrument Kevex Delta-770 Analyst (EDXRF)
Description of X-ray Source Rh x-ray tube; Maximum voltage; 60 KeV
Maximum amperagp: ^.^ mA
Description of Secondary Target Silver secondary target with 0.051 ran
silver secondary target filter. KeV = 35, mA = l.S
Description of Detector Silicon lithium drifted detector
ReferenceN/A
-------
LABORATORY XRF PARAMETERS - PAINT
Counting Time Llvetime: 200 seconds (351 deadtime)
Counting Rate Time constant: 1.5 microseconds
Total Counts N/A
Calibration Standards Matrix; soil and dust; Units (mg/kg Pb)
Soils: 17993, 3772, 2455, 1069, 995, 849, 443
Dust: 58
Results of Calibration Check Samples
TRUE
ID mg/kg Pb
QC1 (17993)
QC2
QC3
(443)
(6550)
AVERAGE
rag/kg Pb 2 RSD RPD
18106 0.5 0.6
458 4.0 3.3
6737 2.6 1.3
N
6
3
3
QCl: High calibration soil standard
QC2: Low range calibration soil standard
QC3: KBS 1648 (Urban Particulate)
N: Number of measurements
-------
LABORATORY XRF PARAMETERS - DUST
Sample Quantity 2 -2.5 grams
Sampla Preparation
(See Attached)
Instrument Kevex Delta-770 Analyst (EDXRF)
Description of X-ray Source Rh X-rgy tube; Maximunt voltage; 60 KeV
Maximum amperage; 3. 3 tnA _
Description of Secondary Target Silver secondary target with 0.051 mm
silver secondary target filter. KeV =35. nA = 1.5 ....
Description of Detector Silicon lithium drifted detectpr
Reference N/A
-------
LABORATORY XRf PARAMETERS - DUST
counting Tima Ltvetlme:200 seconds (351 deadtime)
counting Rate Time constant; 1.5 microseconds
Total Counts N/A
Calibration Standards — CIK 1 and BAX. 1
Results of Calibration Check Samples,
ID
QC1
OC2
QC3
TRUE
nig /kg Pb
17993
443
6550
AVERAGE 1 RSD RPD
mg/kg Pb
15131 0.4 17.3
396 4.5 11.2
6192 2.5 5.6
N
6
3
3
QC1: Soil
QC2: Soil
QC3: NBS 1648 Urban Particulate
N: Number of measurements
-------
EPA/RTI Round Robin for Lead In Paint and Dust
Round Robin No. 002 Lab 10 No. 56
Digestion Method N/A «._ Laboratory
Experience with this Method years
Analysis Method Lab XRF Approval Signature:
Experience with this Method 5" years . l—.
Gross Concentration of Lead fppm}
Sample ID No. Aliquot 1 Aliquot 2
P-1 je/o /£>£
-------
Sample Quantity
LABORATORY XRF PARAMETERS - PAINT
c><' O <3 r<3"< S ^^"1^ &- 1 [ i
Sample Preparation
re
cc>
Instrument
Description of X-ray Source o r &C&tL S /
.
/&&
Description of Secondary Target //'fo
Description of Detector Sj Jj t /^aidl fv 1 j<~T> <$&«, £jw-$jtjs /5~&i> V
Reference
-------
LABORATORY XRF PARAMETERS - PAINT
Counting Time c3 Qt) _5et£~*ls /I
Counting Rate
Total Counts
Calibration Standards C//U1 £ BfiJ- 1
Results of Calibration Check Samples
Pb
-------
LABORATORY XRF PARAMETERS - DUST
Sample Quantity_
Sample Preparation
See
Instrument
Description of X-ray Source
/cZ//i C
Description of Secondary Target
•->&£.
Description of Detector
Reference
-------
LABORATORY XRF PARAMETERS - DUST
Counting Time c*?OC)
Counting Rate
Total Counts
Calibration Standards — CIN 1 and BAL 1
Results of Calibration Check Samples C /AJ J. •'
-------
Appendix F
Letter sent to Laboratories
Reporting Preliminary
Results of Round-Robin
-------
RESEARCH TRIANGLE INSTITUTE
/RTI
Center for Environmental Measurements and Quality Assurance
October 13,1992
Mr. Terry Burke
Wisconsin Occupational Health Laboratory
Department of Hygiene
979 Jonathon Drive
Madison, WI 53713
Dear Mr. Burke:
A statistical analysis of the results of the recent RTI/EPA round robin for
lead in paint and dust is being finalized, and consensus values for concentrations
of the samples have been determined. These values are presented in the
enclosed tables that will be included in, "Preparation and Evaluation of Lead-
based Paint Contaminated Method Evaluation Materials," as presented at the
Lead Symposium of the American Chemical Society meeting in August, 1992.
The paper, to be a part of the proceedings of the symposium, is currently being
reviewed by EPA and, upon clearance, will be sent to all laboratories that
participated in the round robin. It will include the consensus values for the
concentration of the method evaluation samples, a comparision of statistically
significant differences in the analytical methods, and inter- and intra-laboratory
precision for these methods.
All laboratories received 10 samples for analysis, 5 paint and 5 dust
samples. The samples from each matrix included duplicate bottles of one high
level and one low level method evaluation material prepared by RTI, and one
SRM. For example, the paint samples were comprised of one high paint
material (P-3 and P-5), one low paint material (P-l and P-4), and a paint SRM
(P-2, NIST SRM 1579). The dust samples included one high, post-abatement
dust (D-l and D-5), one low household dust material (D-2 and D-4), and one
sediment SRM (D-3, NIST SRM 2711). In order to provide information that will
enable your laboratory to compare the results of its analysis with the consensus
values, enclosed are two tables from the draft paper that provide the consensus
values for the paint and dust samples, as determined from a "grand mean" of the
Post Office Box 12194 Research Triangle Park, North Carolina 27709-2194
Telephone 919 541-6914 Fax: 919 541-5929
-------
digestive methods (hotplate and microwave digestion, followed by AAS or ICP
analysis). Results from analysis by laboratory X-ray fluorescence were not
included in the "grand mean" consensus values because this method exhibited a
negative bias across the matrices.
A description of the preparation of the samples, and methodology used for
the verification of the method evaluation materials will be included in an RTI
report which is currently being prepared. The report will include a complete
statistical analysis of the data, as well as a summary of any problems
encountered by the laboratories in the analysis of the samples. We expect that
the report will be distributed to the participating laboratories by the end of the
year.
Also enclosed is a brochure describing the Environmental Lead Proficiency
Analytical Testing (ELPAT) Program sponsored by the American Industrial
Hygiene Association (AIHA). A number of the laboratories that participated in
the round have been interested in this program, which offers either proficiency
testing or proficiency testing and accreditation. The first round is scheduled for
November 1992.
Once again, we appreciate your participation in the round robin, and we
will be forwarding to you soon a copy of the proceedings paper. In the meantime,
we will be happy to provide assistance if you have questions.
Sincerely,
c/n^jLus
Emily Williams
-------
Emily Williams • 18
Table 4. Mean and Consensus Values for Round Robin Paint Samples
Matrix Sample No. Method
High paint P-3, P-5 MW/AAS
HP/MS
MW/ICP
HP/ICP
LabXRF
Low paint P-1.P-4 MW/AAS
HP/AAS
MW/ICP
HP/ICP
LabXRF
Paint SRM P-2 MW/AAS
HP/AAS
MW/ICP
HP/ICP
LabXRF
Mean ± SO (%RSD), ppm
41,281 ±1,274 (3.1)
36,921 ±713 (1.9)
36,654 ±672 (1.8)
35,670 ± 796 (2.2)
27,404 ±1,567 (5.7)
1,896 ±63 (3.3)
1,661 ±74(4.5)
1,603 ±45 (2.8)
1,600 ±66 (4.1)
1,034 ±76 (7.4)
122,432 ±6,507 (5.3)
104,34018,681 (8.3)
118,281 ±2,476 (2.1)
94,382 ±7,021 (7.4)
112,721 ±13,259 (11. 8)
Consensus Mean* ± SD
(%RSD), ppm
37,632 ±449 (1.2)
37,632 ±449 (1.2)
37,632 ±449(1.2)
37,632 ±449 (1.2)
37,632 ±449 (12)
1,690 ±32 (1.9)
1,690 ±32 (1.9)
1,690 ±32 (1.9)
1,690 ±32 (1.9)
1,690 ±32 (1.9)
109,859 ±3,289 (3.0)
109,859 ±3,289 (3.0)
109,859 ±3,289 (3.0)
109 ,859 ±3,289 (3.0)
109,859 ±3,289 (3.0)
"Lab XRF not included In consensus value determination.
-------
EmiJy Wliams • 19
Table 5. Mean and Consensus Values for Round Robin Dust Samples
Matrix Sample No. Method
High dust D-1, D-5 MW/AAS
HP/MS
MW/ICP
HP/ICP
LabXRF
Low dust D-2, D-4 MW/AAS
HP/AAS
MW/ICP
HP/ICP
LabXRF
Dust SRM D-2 MW/AAS
HP/AAS
MW/ICP
HP/ICP
LabXRF
Mean ± SO
(% RSD), ppm
4,847 ± 127 (2.6)
4,677 ± 103 (2.2)
4,281 ±113(2.6)
4,397 ± 133 (3.0)
2,485 ±117 (4.7)
114 ±6 (5.3)
108 ±7 (6.5)
98 ±3 (3.1)
98 ± 9 (9.2)
93 ± 8 (8.6)
1,327 ±72 (5.4)
1,1 73 ±32 (2.7)
1,133 ±24 (2.1)
1,1 12 ±42 (3.8)
1,029 ±33 (3.2)
Consensus Mean* ±
SD (% RSD), pprn
4,550 ±60 (1.3)
4,550 ±60 (1.3)
4,550 ±60 (1.3)
4,550 ±60 (1.3)
4,550 ±60 (1.3)
104 ±3 (2.9)
104 ± 3 (2.9)
104 ±3 (2.9)
104 ±3 (2.9)
105 ±3 (2.9)
1,1 86 ±23 (1.9)
1,1 86 ±23 (1.9)
1,1 86 ±23 (1.9)
1,1 86 ±23 (1.9)
1,1 86 ±23 (1.9)
*Lab XRF not Included in consensus value determination.
-------
Appendix G
Statistical Analysis of Results
-------
Appendix G-l
Report of Statistical Analysis
by Larry Myers
-------
Statistical Analyses
Brief Summaries of the statistical methods and results are provided below. All
statistical concepts, models and methods, including analysis of variance and interaction,
are treated in Kleinbaum and Kupper (1978, Applied Regression Analysis and other
Multivariable Methods, Duxbury Press, North Scituate, Massachusetts).
1. Censored, Missing and Outlying Values
42 labs were to analyze the panel of 10 samples in duplicate, which would yield
840 results. 848 results were received because two individual results were missing, and
one lab did triplicate analyses on each sample. 28 results were reported as less than a
specified level. These censored values, most of which occurred in the low dust samples,
were removed prior to statistical analysis. This reduced the dataset to 820 results. An
additional 28 observations were removed as outliers. All analyses reported below were
based on the remaining 792 observations.
Determination of Outliers
The following approach was used to determine outliers among the 820
nonmissing, noncensored observations. For each of the six combinations of matrix (dust,
paint) and level (high, low, SRM), a nominal concentration X was obtained as the median
of all reported results from methods 1 through 4. (Method 5 was dearly producing
lower values than the others.) The recovery was then calculated for each individual
result as the ratio Y/X of the reported concentration divided by the nominal
concentration. Using recoveries between 0.35 and 2, the average and standard deviation
of recovery was calculated separately for each of the thirty method(5)-by-matrix(2)-by-
level(3) combinations. The restriction to recoveries between .35 and 2 is a prescreen
intended to remove grosser outliers which can distort the mean and standard deviation.
These statistics were merged back onto the original raw data and a score was calculated
for the recovery of each reported result, by subtracting the average recovery and
-------
dividing by the standard deviation of recovery for the given condition. Any
measurement whose absolute score exceeded 2.576 was excluded as an outlier. This
corresponds to. the upper and lower one-half of one percent of a normal distribution.
This resulted in the exclusion of an additional 28 observations.
2. Consensus values (nominal concentrations)
Consensus values or nominal concentrations for each of the six samples were
calculated as the simple average of the method-specific averages, using nonmissing,
noncensored, nonoutlying values from the four wet chemical (extraction) methods. The
XRF method was excluded from the calculation of nominal values because of a
pronounced negative bias relative to the other methods.
3. Tests for sample homogeneity.
The non-SRM samples were supplied as blinded duplicates. For these samples
it is possible to test for homogeneity of the parent stocks using two-way analysis of
variance, treating sampling, analysis, and their interaction as random effects. {That is,
laboratories within a method, and replicate samples selected form the same parent stock,
such as D-2 and D-4, were both viewed as random selection from a (normally
distributed) population of same. The assumption of random effects is appropriate in
order to generalize results to a larger population of laboratories.} This model was fit
separately to all 20 combinations of method(5)-by-matrix(2)-by-level(2) which involved
non-SRM samples.
A preliminary test for the absence of interaction between sample and laboratory
indicated that this assumption was reasonable. (Only one of twenty interaction tests was
significant at the 5% level (low dust, method 1, .025
-------
Only one of twenty tests for sample main effects was significant (low dust/
method 4, .025
-------
repeatable than each of the other methods (p<.05 for each comparison). None of the
other repeatability comparisons approaches significance.
The reproducibility estimates of the two MW methods are similar and lower than
those of the HP and XRF methods. Formal comparisons are difficult because of the
complex probability distribution of the reproducibility estimate, exacerbated by the
imbalance resulting from censoring and deletion of outliers. Using Satterwaite's
approximation to the degrees of freedom, MW/ICP is significantly more reproducible
at the 1% level than XRF and both of the HP methods.
-------
Appendix G-2
Review of Statistical Analysis
by Jack Suggs
-------
Review of Statistical Analysis by Jack Suggs
The results shown in Table 7 were taken from Larry Myers' original report. The
concentration averages, X, are expressed in original units (ppm). The standard
deviations: sample-to-sample, within-lab, and between-lab are expressed as a percentage
of level (based on analysis of logarithms).
1. For non-SRM samples, the sample-to-sample variation was based on a two-
way analysis of variance of logs with no interaction applied separately to all 20
combinations of methods (5)-by-matrix(2)-by-level(2). The standard deviation for
samples (in percent) is equivalent to a percent-difference between samples. Only
one case (low dust, method 4) was observed to have a significant percent
difference between samples. In all other cases, the sample-to-sample differences
were zero (16 out of 20 cases) or nowhere near significant. The conclusion is that
bulk sample material prepared by RTI does not significantly contribute to the
overall method variation in analysis.
2. The order (or ranking) of the methods with respect to averages is consistent
and highly significant in this regard. Method 1 has the highest average on each
of the six samples. The chance of this happening is 0.000064 if all the methods
were equal. Also method 2 has the second highest average of 5 of the six
samples. Method 5 also has the lowest average on 5 of 6 samples which is also
significant.
The repeatability (within-lab) and reproducibility (between-lab) standard
deviations are based on a one-way analysis of variance of log-recoveries ignoring
sample-to-sample differences. (These differences are absorbed into the estimates of
repeatability and reproducibility, which were shown above to be non-significant.) There
were no sampling effects with regards to SRMs. These results came from Larry Myers
original report.
-------
1. Method 5 has the best repeatability in log units on all six samples. By the
same logic applied to the ranking of the averages, this result is also highly
significant. This may be due to the possibility that the log transformation did not
sufficiently stabilize the variances and that method 5 is actually operating at a
different apparent level than the other methods on some of the samples. At the
same time method 5 was fairly consistent in repeatability across all levels. No
other consistencies could be recognized.
2. The most important single measure of method performance is
reproducibility because it reflects interlaboratory as well all within laboratory
variability. Method 5 has the worst (highest) reproducibility for all three paint
samples. Method 3 has the lowest (best) reproducibility on five of the six
samples.
3. It is desirable to have a constant percent repeatability and reproducibility
apply across ail levels of measurement at least for a given method. Table 7 does
not support this. However, regressions of repeatability and reproducibility versus
level for each method may provide a useful way of estimating method variability
given a specific level of measurement. Prediction intervals could be calculated
at the 95% probability level to predict the occurrence of future values of
repeatability and reproducibility for a given method and a given level of
measurement.
If the intercepts are forced through zero, the slope represents a percent
change in repeatability or reproducibility for each unit change in measurement
level.
4. Another estimation procedure along these lines would be to pool all
information for each method separately (this includes paint, dust, SRMs) into an
analysis of variance (one-way disregarding measurement level). As I stated
-------
above, this represents an alternative to the regression approach which provides
a "single" estimate of repeatability or reproducibility as a function (or percentage
of level).
-------
Appendix G-3
Raw Data File
-------
LEGEND
(Appendix G-3)
OBS = Reported Result
BA = Laboratory Code
LEVEL = Concentration Level
L = Low
H = High
S = SRM
SAM = Sample Number
P = Paint
D = Dust
REP = Replicate Number
CEN = Censored Data - Data reported as less than a specified level
CONG = Concentration (pg/g)
ANAL = Analytical Method
ICP = Inductively Coupled Plasma Emission Spectrometry
AA = Atomic Absorption Spectrometry
EXTR = Extraction Method
NIO = NIOSH Method 7082
EPA = EPA/AREAL Method
CONCAT = Concentration Category
+ = Reported
m = Missing
-------
raw data file
OBS BA LEVEL SAM REP CEN CONC ANAL EXTR CONCAT
1 42 L P-l 1 1600 ICP NIO +
2 42 L P-l 2 1500 ICP NIO +
3 42 L P-4 1 1700 ICP NIO +
4 42 L P-4 2 1900 ICP NIO +
5 42 H P-3 1 36000 ICP NIO +
6 42 H P-3 2 36000 ICP NIO +
7 42 H P-5 1 36000 ICP NIO +
8 42 H P-5 2 37000 ICP NIO +
9 42 S P-2 1 110000 ICP NIO +
10 42 S P-2 2 110000 ICP NIO +
11 42 L D-2 1 110 ICP NIO +
12 42 L D-2 2 100 ICP NIO +
13 42 L 0-4 1 82 ICP NIO +
14 42 L D-4 2 130 ICP NIO +
15 42 H D-l 1 4400 ICP NIO +
16 42 H D-l 2 4500 ICP NIO +
17 42 H D-5 1 4500 ICP NIO 't-
is 42 H D-5 2 5300 ICP NIO +
19 42 S 0-3 1 1200 ICP NIO +
20 42 S D-3 2 1200 ICP NIO +
21 38 L P-l 1 1160 ICP NIO +
22 38 L P-l 2 1280 ICP NIO +
23 38 L P-4 1 1280 ICP NIO +
24 38 L P-4 2 1330 ICP NIO +
25 38 H P-3 1 32000 ICP NIO +
26 38 H P-3 2 30800 ICP NIO +
27 38 H P-5 1 28600 ICP NIO +
28 38 H P-5 2 30200 ICP NIO +
29 38 S P-2 1 84000 ICP NIO +
30 38 S P-2 2 ICP NIO m
31 38 L D-2 1 160 ICP NIO +
32 38 L 0-2 2 80 ICP NIO +
33 38 L D-4 1 70 ICP NIO +
34 38 L D-4 2 70 ICP NIO +
35 38 H D-l 1 3160 ICP NIO +
36 38 H D-l 2 4110 ICP NIO +
37 38 H D-5 1 3580 ICP NIO +
38 38 H D-5 2 2670 ICP NIO +
39 38 S D-3 1 840 ICP NIO +
40 38 S D-3 2 840 ICP NIO +
41 33 L P-l 1 1070 IQP NIO +
42 33 L P-l 2 1400 ICP NIO +
43 33 L P-l 3 1340 ICP NIO +
44 33 L P-4 1 1500 ICP NIO +
45 33 L P-4 2 1180 ICP NIO +
46 33 L P-4 3 1250 ICP NIO +
47 33 H P-3 1 35200 ICP NIO +
48 33 H P-3 2 32900 ICP NIO +
49 33 H P-3 3 34300 ICP NIO +
50 33 H P-5 1 34600 ICP NIO +
51 33 H P-5 2 33400 ICP NIO +
52 33 H P-5 3 34100 ICP NIO +
53 33 S P-2 1 54500 ICP NIO +
54 33 S P-2 2 46900 ICP NIO +
55 33 S P-2 3 57800 ICP NIO +
56 33 L D-2 1 < 50 ICP NIO +
57 33 L D-2 2 < 50 ICP NIO +
58 33 L D-2 3 < 50 ICP NIO +
-------
raw data file
08S BA LEVEL SAM
59 33 L D-4 1 < 50 ICP NIO +
60 33 L D-4 2 66 ICP NIO +
61 33 L
62 33 H D-l 1 5040 ICP NIO +
63 33 H D-l 2 5010 ICP NIO +
64 33 H D-l 3 4350 ICP NIO +
65 33 H D-5 1 5560 ICP NIO +
66 33 H
67 33 H D-5 3 4360 ICP NIO +
68 33 S D-3 1 1050 ICP NIO +
69 33 S D-3 2 1000 ICP NIO +
70 33 S
71 28 L
72 28 L
73 28 L
74 28 L P-4 2 1700 ICP NIO +
75 28 H P-3 1 35000 ICP NJO +
76 28 H
77 28 H
78 28 H
79 28 S
80 28 S
81 28 L
82 28 L
83 28 L D-4 1 < 40 ICP NIO +
84 28 L D-4 2 < 34 ICP NIO +
85 28 H D-l 1 3800 ICP NIO +
86 28 H
87 28 H D-5 1 3700 ICP NIO +
88 28 H D-5 2 4700 ICP NIO +
89 28 S
90 28 S
91 23 L
92 23 L P-l 2 1615 ICP NIO +
93 23 L P-4 1 1450 ICP NIO +
94 23 L P-4 2 1630 ICP NIO +
95 23 H P-3 1 34500 ICP NIO +
96 23 H P-3 2 34700 ICP NIO +
97 23 H
98 23 H
99 23 S
100 23 S
101 23 L D-2 1 93 ICP NIO +
102 23 L D-2 2 108 ICP NIO +
103 23 L D-4 1 74 ICP NIO +
104 23 L D-4 2 90 ICP NIO +
105 23 H
106 23 H D-l 2 4460 ICP NIO +
107 23 H 0-5 1 4220 ICP NIO +
108 23 H D-5 2 4110 ICP NIO +
109 23 S
110 23 S D-3 2 1100 ICP NIO +
111 44 L P-l 1 1556 ICP NIO +
112 44 L
113 44 L
114 44 L P-4 2 1744 ICP NIO +
115 44 H P-3 1 37140 ICP NIO +
116 44 H
SAM
D-4
D-4
D-4
D-l
D-l
D-l
D-5
D-5
D-5
D-3
D-3
D-3
P-l
P-l
P-4
P-4
P-3
P-3
P-5
P-5
P-2
P-2
D-2
D-2
D-4
D-4
D-l
D-l
D-5
D-5
D-3
D-3
P-l
P-l
P-4
P-4
P-3
P-3
P-5
P-5
P-2
P-2
D-2
D-2
D-4
D-4
D-l
0-1
0-5
D-5
D-3
D-3
P-l
P-l
P-4
P-4
P-3
P-3
REP
1
2
3
1
2
3
1
2
3
1
2
3
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
CEN CONC
< 50
66
50
5040
5010
4350
5560
4540
4360
1050
1000
1030
2000
1400
1600
1700
35000
35000
37000
35000
70000
88000
61
< 35
< 40
< 34
3800
4400
3700
4700
1100
1000
1790'
1615
1450
1630
34500
34700
34500
34100
119000
115000
93
108
74
90
4060
4460
4220
4110
1120
1100
1556
1537
1882
1744
37140
35870
ANAL
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
EXTI
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NJO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
-------
raw data file
OBS BA LEVEL SAM REP CEN CONC ANAL EXTR CONCAT
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
44
44
44
44
44
44
44
44
44
44
44
44
44
44
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
46
46
46
46
H.
H
S
S
L
L
I
I
H
H
H
H
S
S
I
I
I
I
H
H
H
H
S
S
L
L
L
L
H
H
H
H
S
S
L
L
L
L
H
H
H
H
S
S
L
L
L
L
H
H
H
H
S
S
L
L
L
L
P-5
P-5
P-2
P-2
D-2
D-2
D-4
D-4
D-l
D-l
D-5
D-5
D-3
D-3
P-l
P-l
P-4
P-4
P-3
P-3
P-5
P-5
P-2
P-2
D-2
D-2
D-4
D-4
D-l
D-l
D-5
D-5
D-3
D-3
P-l
P-l
P-4
P-4
P-3
P-3
P-5
P-5
P-2
P-2
D-2
D-2
D-4
D-4
D-l
D-l
D-5
D-5
D-3
D-3
P-l
P-l
P-4
P-4
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
36510
36980
110500
113400
102
100
104
97
4013
4414
4535
4532
1249
1220
1757
1685
1872
1607
47300
36540
41260
44340
114760
111620
150
142
< 100
< 100
3365
5033
5538
5112
1317
1241
1720
1810
1940
1990
36900
37400
37200
36400
115000
94700
270
150
160
< 50
4170
4750
4540
4590
1200
1140
1628
1693
1555
1451
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
NIO H
NIO H
NIO H
NIO H
NIO H
NIO H
NIO H
NIO H
NIO ^
NIO H
NIO H
NIO H
NIO H
NIO 4
NIO 4
NIO 4
NIO 4
NIO 4
NIO 4
NIO 4
NIO 4
NIO 4
NIO 4
NIO 4
NIO 4
NIO 4
NIO 4
NIO 4
NIO 4
NIO 4
NIO 4
NIO 4
NIO 4
NIO 4
NIO 4
NIO 4
NIO 4
NIO 4
NIO 4
NIO 4
NIO +
NIO +
NIO +
NIO +
NIO 4
NIO 4-
NIO +
NIO +
NIO 4-
NIO 4-
NIO 4-
NIO +
NIO +
NIO +
NIO 4-
NIO +
NIO +
NIO 4-
-------
raw data file
OBS BA LEVEL SAM REP CEN CONC ANAL EXTR CONCAT
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
46
46
46
46
46
46
46
46
46
46
46
46
46
46
46
46
45
45
45
45
45
45
45
45
45
45
45
45
45
45
45
45
45
45
45
45
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
25
25
H
H
H
H
S
S
L
L
L
L
H
H
H
H
S
S
L
L
L
L
H
H
H
H
S
S
L
L
L
L
H
H
H
H
S
S
L
L
L
L
H
H
H
H
S
S
L
L
L
L
H
H
H
H
S
S
L
L
P-3
P-3
P-5
P-5
P-2
P-2
D-2
D-2
D-4
D-4
D-l
D-l
D-5
D-5
D-3
D-3
P-l
P-l
P-4
P-4
P-3
P-3
P-5
P-5
P-2
P-2
D-2
D-2
D-4
D-4
0-1
D-l
D-5
D-5
D-3
D-3
P-l
P-l
P-4
P-4
P-3
P-3
P-5
P-5
P-2
P-2
D-2
D-2
D-4
D-4
D-l
D-l
D-5
D-5
D-3
D-3
P-l
P-l
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
36420
36410
35800
36910
83220
92530
< 200
< 200
< 200
< 200
5010
4057
4047
4352
1168
1224
1650
2330
1840
2010
34500
42200
34100
38700
78700
118000
98
54
90
48
3860
5950
3860
7150
1010
1830
5434
5148
6003
5823
29573
27368
26403
26178
104510
101852
126
137
113
126
2000
2000
1400
1900
863
916
1300
1300
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
XRF
XRF
XRF
XRF
XRF
XRF
XRF
XRF
XRF
XRF
XRF
XRF
XRF
XRF
XRF
XRF
XRF
XRF
XRF
XRF
XRF
XRF
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
-------
raw data fi1e
OBS BA LEVEL SAM REP CEN CONC ANAL EXTR CONCAT
233 25 L P-4 1 1300 XRF N/A +
234 25 L P-4 2 1300 XRF N/A +
235 25 H P-3 I 32510 XRF N/A +
236 25 H P-3 2 29650 XRF N/A +
237 25 H P-5 1 29450 XRF N/A +
238 25 H P-5 2 31980 XRF N/A +
239 25 S P-2 1 > 50000 XRF N/A +
240 25 S P-2 2 > 50000 XRF N/A +
241 25 L D-2 1 < 75 XRF N/A +
242 25 L D-2 2 < 75 XRF N/A +
243 25 L 0-4 1 < 75 XRF N/A +
244 25 L D-4 2 < 75 XRF N/A +
245 25 H D-l 1 2951 XRF N/A +
246 25 H 0-1 2 2751 XRF N/A +
247 25 H 0-5 1 2948 XRF N/A +
248 25 H 0-5 2 2921 XRF N/A +
249 25 S 0-3 1 981 XRF N/A +
250 25 S 0-3 2 1007 XRF N/A +
251 30 L P-l 1 934 XRF N/A +
252 30 L P-l 2 879 XRF N/A +
253 30 L P-4 1 881 XRF N/A +
254 30 L P-4 2 906 XRF N/A +
255 30 H P-3 1 25440 XRF N/A +
256 30 H P-3 2 24780 XRF N/A +
257 30 H P-5 1 24340 XRF N/A +
258 30 H P-5 2 24420 XRF N/A +
259 30 S P-2 1 129600 XRF N/A +
260 30 S P-2 2 133300 XRF N/A +
261 30 L D-2 1 71 XRF N/A +
262 30 L D-2 2 73 XRF N/A +
263 30 L D-4 I 78 XRF N/A +
264 30 L D-4 2 75 XRF N/A +
265 30 H D-l 1 2167 XRF N/A +
266 30 H D-l 2 2133 XRF N/A +
267 30 H D-5 1 2166 XRF N/A +
268 30 H D-5 2 2200 XRF N/A +
269 30 S D-3 1 1100 XRF N/A +
270 30 S D-3 2 1100 XRF N/A +
271 10 L P-l 1 1200 XRF N/A +
272 10 L P-l 2 1183 XRF N/A +
273 10 L P-4 1 1112 XRF N/A +
274 10 L P-4 2 1210 XRF N/A +
275 10 H P-3 1 23112 XRF N/A +
276 10 H P-3 2 23992 XRF N/A +
277 10 H P-5 1 23816 XRF N/A +
278 10 H P-5 2 23992 XRF N/A +
279 10 S P-2 1 118327 XRF N/A +
280 10 S P-2 2 118327 XRF N/A +
281 10 L D-2 1 72 XRF N/A +
282 10 L D-2 2 82 XRF N/A +
283 10 L D-4 1 76 XRF N/A +
284 10 L D-4 2 72 XRF N/A +
285 10 H D-l 1 2775 XRF N/A +
286 10 H D-l 2 2415 XRF N/A +
287 10 H D-5 1 2435 XRF N/A +
288 10 H D-5 2 2775 XRF N/A +
289 10 S D-3 1 1074 XRF N/A +
290 10 S D-3 2 1014 XRF N/A +
-------
raw data file
OBS BA LEVEL SAM REP CEN CONC ANAL EXTR CONCAT
291 5 L P-l 1 1006 XRF N/A +
292 5 L P-l 2 910 XRF N/A +
293 5 L P-4 1 973 XRF N/A +
294 5 L P-4 2 1021 XRF N/A +
295 5 H P-3 1 31905 XRF N/A +
296 5 H P-3 2 31228 XRF N/A +
297 5 H P-5 1 33982 XRF N/A +
298 5 H P-5 2 32388 XRF N/A +
299 5 S P-2 1 105000 XRF N/A +
300 5 S P-2 2 104000 XRF N/A +
301 5 L D-2 1 107 XRF N/A +
302 5 L D-2 2 81 XRF N/A +
303 5 L D-4 1 81 XRF N/A +
304 5 L D-4 2 87 XRF N/A +
305 5 H D-l 1 2489 XRF N/A +
306 5 H D-l 2 2458 XRF N/A +
307 5 H D-5 1 2441 XRF N/A +
308 5 H D-5 2 2514 XRF N/A +
309 5 S D-3 1 976 XRF N/A +
310 5 S D-3 2 962 XRF N/A +
311 49 L P-l 1 1010 XRF N/A +
312 49 L P-l 2 1089 XRF N/A +
313 49 L P-4 1 1059 XRF N/A +
314 49 L P-4 2 1076 XRF N/A +
315 49 H P-3 1 31370 XRF N/A +
316 49 H P-3 2 30760 XRF N/A +
317 49 H P-5 1 30780 XRF N/A +
318 49 H P-5 2 31140 XRF N/A +
319 49 S P-2 1 156550 XRF N/A +
320 49 S P-2 2 159390 XRF N/A +
321 49 L D-2 1 83 XRF N/A +
322 49 L D-2 2 79 XRF N/A •*•
323 49 L D-4 1 92 XRF N/A +
324 49 L D-4 2 82 XRF N/A +
325 49 H D-l 1 2703 XRF N/A +
326 49 H D-l 2 2883 XRF N/A +
327 49 H D-5 1 2716 XRF N/A +
328 49 H D-5 2 2666 XRF N/A +
329 49 S D-3 1 1134 XRF N/A +
330 49 S D-3 2 1161 XRF N/A +
331 15 L P-l 1 819 XRF N/A +
332 15 L P-l 2 782 XRF N/A +
333 15 L P-4 1 800 XRF N/A +
334 15 L P-4 2 761 XRF N/A +
335 15 H P-3 1 21591 XRF N/A +
336 15 H P-3 2 21766 XRF N/A +
337 15 H P-5 1 21845 XRF N/A +
338 15 H P-5 2 21556 XRF N/A +
339 15 S P-2 1 61123 XRF N/A +
340 15 S P-2 2 60677 XRF N/A +
341 15 L D-2 1 93 XRF N/A +
342 15 L D-2 2 114 XRF N/A +
343 15 L D-4 1 118 XRF N/A +
344 15 L D-4 2 111 XRF N/A +
345 15 H D-l 1 2417 XRF N/A +
346 15 H D-l 2 2444 XRF N/A +
347 15 H D-5 1 2415 XRF N/A +
348 15 H D-5 2 2424 XRF N/A +
-------
raw data file
OBS BA LEVEL SAM REP CEN CONC ANAL EXTR CONCAT
349 15 S D-3 1 1052 XRF N/A +
350 15 S D-3 2 1067 XRF N/A +
351 40 L P-l 1 1544 AA NIO +
352 40 L P-l 2 1438 AA NIO +
353 40 L P-4 1 1446 AA NIO +
354 40 L P-4 2 1458 AA NIO +
355 40 H P-3 1 36790 AA NIO +
356 40 H P-3 2 42605 AA NIO +
357 40 H P-5 1 37144 AA NIO +
358 40 H P-5 2 35990 AA NIO +
359 40 S P-2 1 116025 AA NIO +
360 40 S P-2 2 99577 AA NIO +
361 40 L D-2 1 96 AA NIO +
362 40 L D-2 2 100 AA NIO -f-
363 40 L D-4 1 110 AA NIO +
364 40 L D-4 2 110 AA NIO +
365 40 H D-l 1 4464 AA NIO +
366 40 H D-l 2 4504 AA NIO +
367 40 H D-5 1 4333 AA NIO +
368 40 H D-5 2 4669 AA NIO +
369 40 S D-3 1 1067 AA NIO +
370 40 S D-3 2 1113 AA NIO +
371 36 L P-l 1 1510 AA NIO +
372 36 L P-l 2 1790 AA NIO +
373 36 L P-4 1 1940 AA NIO +
374 36 L P-4 2 1790 AA NIO +
375 36 H P-3 1 33500 AA NIO +
376 36 H P-3 2 39500 AA NIO +
377 36 H P-5 1 36900 AA NIO +
378 36 H P-5 2 41600 AA NIO +
379 36 S P-2 1 102000 AA NIO +
380 36 S P-2 2 111000 AA NIO +
381 36 L D-2 1 < 100 AA NIO +
382 36 L D-2 2 140 AA NIO +
383 36 L D-4 1 108 AA NIO +
384 36 L D-4 2 171 AA NIO +
385 36 H D-l 1 3990 AA NIO +
386 36 H D-l 2 4390 AA NIO +
387 36 H D-5 1 4603 AA NIO +
388 36 H D-5 2 5710 AA NIO +
389 36 S D-3 1 1130 AA NIO +
390 36 S D-3 2 1240 AA NIO +
391 31 L P-l 1 1790 AA NIO +
392 31 L P-l 2 1700 AA NIO +
393 31 L P-4 1 2030 AA NIO +
394 31 L P-4 2 1990 AA NIO +
395 31 H P-3 1 41000 AA NIO +
396 31 H P-3 2 39500 AA NIO +
397 31 H P-5 1 43600 AA NIO +
398 31 H P-5 2 46300 AA NIO +
399 31 S P-2 1 140000 AA NIO +
400 31 S P-2 2 132000 AA NIO +
401 31 L D-2 1 116 AA NIO +
402 31 L D-2 2 98 AA NIO +
403 31 L D-4 1 130 AA NIO +
404 31 L D-4 2 100 AA NIO +
405 31 H D-l 1 5300 AA NIO +
406 31 H D-l 2 5740 AA NIO +
-------
raw data fi le
8
OBS
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
BA LEVEL SAM REP CEN
CONC ANAL EXTR CONCAT
31
31
31
31
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
21
21
21
21
21
21
21
21
21
21
21
21
21
21
21
21
21
21
21
21
16
16
16
16
16
16
16
16
16
16
16
16
16
16
H
H
S
S
L
L
L
L
H
H
H
H
S
S
L
L
L
I
H
H
H
H
S
S
L
L
L
L
H
H
H
H
S
S
L
L
L
L
H
H
H
H
S
S
L
L
L
L
H
H
H
H
S
S
L
L
L
L
D-5
D-5
D-3
D-3
P-l
P-l
P-4
P-4
P-3
P-3
P-5
P-5
P-2
P-2
D-2
D-2
D-4
D-4
D-l
D-l
D-5
D-5
D-3
D-3
P-l
P-l
P-4
P-4
P-3
P-3
P-5
P-5
P-2
P-2
D-2
D-2
D-4
D-4
D-l
D-l
D-5
D-5
D-3
D-3
P-l
P-l
P-4
P-4
P-3
P-3
P-5
P-5
P-2
P-2
D-2
D-2
D-4
D-4
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
4990
5280
1260
1290
2020
1640
1760
1900
39000
38400
38700
38600
112000
113000
< 300
< 300
< 300
< 300
4680
4150
5080
4760
1180
1320
1696
1324
1146
1080
34991
33550
35010
34140
118820
115359
97
100
92
96
4840
4709
4694
4520
960
960
1350
1213
1383
1478
33833
36098
32055
35567
105667
110000
89
84
65
79
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
4-
+
+
•f
•f-
+
+
+
+
•(-
+
+
+
•f-
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
•f
-------
raw data file
OBS
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
8A LEVEL SAM REP CEN
CONC ANAL EXTR CONCAT
16
16
16
16
16
16
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
1
1
1
1
1
1
1
1
1
1
1
1
H
H
H
H
S
S
L
L
L
L
H
H
H
H
S
S
L
L
L
L
H
H
H
H
S
S
L
L
L
L
H
H
H
H
S
S
L
L
L
L
H
H
H
H
S
S
L
L
L
L
H
H
H
H
S
S
L
L
D-l
D-l
0-5
D-5
D-3
D-3
P-l
P-l
P-4
P-4
P-3
P-3
P-5
P-5
P-2
P-2
D-2
D-2
D-4
D-4
D-l
D-l
D-5
D-5
D-3
D-3
P-l
P-l
P-4
P-4
P-3
P-3
P-5
P-5
P-2
P-2
D-2
D-2
D-4
D-4
D-l
D-l
D-5
D-5
D-3
D-3
P-l
P-l
P-4
P-4
P-3
P-3
P-5
P-5
P-2
P-2
D-2
D-2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
,2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
3531
4463
3191
4196
1208
1177
1700
1600
1700
1800
36000
35000
36000
37000
46000
55000
140
120
no
130
4700
4700
4800
4700
1200
1200
1310
2064
1852
2047
36594
35340
34614
35772
14010
5077
214
199
85
93
4143
3889
5241
5179
1186
1217
1542
2096
1805
1879
37699
35974
37160
37002
93532
99463
109
111
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
AA
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
NIO +
-------
raw data file 10
OBS BA LEVEL SAM REP CEN CONC ANAL EXTR CONCAT
523 1 L 0-4 1 109 AA NIO +
524 1 L D-4 2 140 AA NIO +
525 1 H D-l 1 4567 AA NIO +
526 1 H D-l 2 5014 AA NIO +
527 1 H D-5 1 5096 AA NIO +
528 1 H D-5 2 4071 AA NIO +
529 1 S 0-3 1 1199 AA NIO +
530 1 S 0-3 2 1207 AA NIO +
531 39 L P-l 1 1670 ICP EPA +
532 39 L P-l 2 1220 ICP EPA +
533 39 L P-4 1 1230 ICP EPA +
534 39 L P-4 2 1490 ICP EPA +
535 39 H P-3 1 37800 ICP EPA +
536 39 H P-3 2 38000 ICP EPA +
537 39 H P-5 1 35800 ICP EPA +
538 39 H P-5 2 38700 ICP EPA +
539 39 S P-2 1 135000 ICP EPA +
540 39 S P-2 2 123000 ICP EPA +
541 39 L D-2 1 87 ICP EPA +
542 39 L D-2 2 108 ICP EPA +
543 39 L D-4 1 97 ICP EPA +
544 39 L D-4 2 84 ICP EPA +
545 39 H D-l 1 3090 ICP EPA +
546 39 H D-l 2 3690 ICP EPA +
547 39 H D-5 1 3980 ICP EPA +
548 39 H 0-5 2 3840 ICP EPA +
549 39 S 0-3 1 1010 ICP EPA +
550 39 S D-3 2 1060 ICP EPA +
551 34 L P-l 1 1410 ICP EPA +
552 34 L P-l 2 1750 ICP EPA +
553 34 L P-4 1 1370 ICP EPA +
554 34 L P-4 2 1600 ICP EPA +
555 34 H P-3 1 34400 ICP EPA +
556 34 H P-3 2 33800 ICP EPA +
557 34 H P-5 1 35500 ICP EPA +
558 34 H P-5 2 35400 ICP EPA +
559 34 S P-2 1 116000 ICP EPA +
560 34 S P-2 2 118000 ICP EPA +
561 34 L D-2 1 107 ICP EPA +
562 34 L D-2 2 98 ICP EPA +
563 34 L D-4 1 88 ICP EPA +
564 34 L D-4 2 103 ICP EPA +
565 34 H D-l 1 3740 ICP EPA +
566 34 H D-l 2 4230 ICP EPA +
567 34 H D-5 1 3460 ICP EPA +
568 34 H D-5 2 4680 ICP EPA +
569 34 S D-3 1 1200 ICP EPA +
570 34 S D-3 2 1150 ICP EPA +
571 29 L P-l 1 1600 ICP EPA +
572 29 L P-l 2 1400 ICP EPA +
573 29 L P-4 1 2120 ICP EPA +
574 29 L P-4 2 1590 ICP EPA +
575 29 H P-3 I 35800 ICP EPA +
576 29 H P-3 2 35000 ICP EPA +
577 29 H P-5 1 39400 ICP EPA +
578 29 H P-5 2 37600 ICP EPA +
579 29 S P-2 1 116000 ICP EPA +
580 29 S P-2 2 115000 ICP EPA +
-------
raw data file 11
DBS BA LEVEL SAM REP CEN CONC ANAL EXTR CONCAT
581 29 L D-2 1 126 ICP EPA +
582 29 L D-2 2 98 ICP EPA +
583 29 L D-4 1 88 ICP EPA +
584 29 L 0-4 2 98 ICP EPA +
585 29 H 0-1 1 4260 ICP EPA +
586 29 H D-l 2 3940 ICP EPA +
587 29 H D-5 1 4720 ICP EPA +
588 29 H D-5 2 5360 ICP EPA +
589 29 S 0-3 1 1220 ICP EPA +
590 29 S D-3 2 1150 ICP EPA +
591 9 L P-l 1 1540 ICP EPA +
592 9 L P-l 2 1680 ICP EPA +
593 9 L P-4 1 1400 ICP EPA +
594 9 L P-4 2 14-10 ICP EPA +
595 9 H P-3 1 38900 ICP EPA +
596 9 H P-3 2 37600 ICP EPA +
597 9 H P-5 1 36600 ICP EPA +
598 9 H P-5 2 41000 KP EPA +
599 9 S P-2 1 119000 ICP EPA +
600 9 S P-2 2 123000 ICP EPA +
601 9 L D-2 1 74 ICP EPA +
602 9 L D-2 2 83 ICP EPA +
603 9 L D-4 1 72 ICP EPA +
604 9 L D-4 2 84 ICP EPA +
605 9 H D-l 1 5640 ICP EPA +
606 9 H D-l 2 4840 ICP EPA +
607 9 H D-5 1 4270 ICP EPA +
608 9 H D-5 2 4190 ICP EPA +
609 9 S D-3 1 950 ICP EPA +
610 9 S D-3 2 1070 ICP EPA +
611 24 L P-l 1 ICP EPA m
612 24 L P-l 2 1700 ICP EPA +
613 24 L P-4 1 1600 ICP EPA +
614 24 L P-4 2 1800 ICP EPA +
615 24 H P-3 1 39000 ICP EPA +
616 24 H P-3 2 38000 ICP EPA +
617 24 H P-5 1 39000 ICP EPA +
618 24 H P-5 2 40000 ICP EPA +
619 24 S P-2 1 120000 ICP EPA +
620 24 S P-2 2 130000 ICP EPA +
621 24 L D-2 1 < 22 ICP EPA +
622 24 L D-2 2 < 22 ICP EPA +
623 24 L D-4 1 100 ICP EPA +
624 24 L D-4 2 31 ICP EPA +
625 24 H D-l 1. 4800 ICP EPA +
626 24 H D-l 2 4300 ICP EPA +
627 24 H D-5 1 4700 ICP EPA +
628 24 H D-5 2 2500 ICP EPA +
629 24 S D-3 1 1200 ICP EPA +
630 24 S D-3 2 1200 ICP EPA +
631 19 L P-l 1 1432 ICP EPA +
632 19 L P-l 2 1408 ICP EPA +
633 19 L P-4 1 1518 ICP EPA +
634 19 L P-4 2 1502 ICP EPA +
635 19 H P-3 1 34000 ICP EPA +
636 19 H P-3 2 34100 ICP EPA +
637 19 H P-5 1 32400 ICP EPA +
638 19 H P-5 2 32600 ICP EPA +
-------
raw data file 12
OBS BA LEVEL SAM REP CEN CONC ANAL EXTR CONCAT
639 19 S P-2 1 109400 ICP EPA +
640 19 S P-2 2 109600 ICP EPA +
641 19 L D-2 1 87 ICP EPA +
642 19 L D-2 2 89 ICP EPA +
643 19 L 0-4 1 145 ICP EPA +
644 19 L D-4 2 98 ICP EPA +
645 19 H D-l 1 4160 ICP EPA +
646 19 H D-l 2 4170 ICP EPA +
647 19 H D-5 1 3960 ICP EPA +
648 19 H D-5 2 3960 ICP EPA +
649 19 S D-3 1 1142 ICP EPA +
650 19 S D-3 2 1104 ICP EPA +
651 14 L P-l 1 1896 ICP EPA +
652 14 L P-l 2 1529 ICP EPA +
653 14 L P-4 1 1995 ICP EPA +
654 14 L P-4 2 1775 ICP EPA +
655 14 H P-3 1 42112 ICP EPA +
656 14 H P-3 2 37519 ICP EPA +
657 14 H P-5 1 37685 ICP EPA +
658 14 H P-5 2 37270 ICP EPA +
659 14 S P-2 1 126637 ICP EPA +
660 14 S P-2 2 120216 ICP EPA +
661 14 L D-2 1 211 ICP EPA +
662 14 L D-2 2 101 ICP EPA +
663 14 L D-4 1 99 ICP EPA +
664 14 L D-4 2 98 ICP EPA +
665 14 H D-l 1 4980 ICP EPA +
666 14 H D-l 2 4443 ICP EPA +
667 14 H D-5 1 4258 ICP EPA +
668 14 H D-5 2 4026 ICP EPA +
669 14 S D-3 1 1192 ICP EPA +
670 14 S D-3 2 1206 ICP EPA +
671 4 L P-l 1 1500 ICP EPA +
672 4 L P-l 2. 1880 ICP EPA +
673 4 L P-4 1 1550 ICP EPA +
674 4 L P-4 2 1830 ICP EPA +
675 4 H P-3 1 35200 ICP EPA +
676 4 H P-3 2 36700 ICP EPA +
677 4 H P-5 1 33700 ICP EPA +
678 4 H P-5 2 35200 ICP EPA +
679 4 S P-2 1 117000 ICP EPA +
680 4 S P-2 2 120000 ICP EPA +
681 4 L D-2 1 80 ICP EPA +
682 4 L D-2 2 140 ICP EPA +
683 4 L D-4 1 170 ICP EPA +
684 4 L D-4 2 110 ICP EPA +
685 4 H 0-1 1 4070 ICP EPA +
686 4 H 0-12 4960 ICP EPA +
687 4 H 0-51 4110 ICP EPA +
688 4 H D-5 2 3900 ICP EPA +
689 4 S D-3 1 1170 ICP EPA +
690 4 S D-3 2 1180 ICP EPA +
691 43 L P-l 1 1640 ICP EPA +
692 43 L P-l 2 < 10 ICP EPA +
693 43 L P-4 1 1490 ICP EPA +
694 43 L P-4 2 1980 ICP EPA +
695 43 H P-3 1 36100 ICP EPA +
696 43 H P-3 2 35600 ICP EPA +
-------
raw data file 13
OBS BA LEVEL SAM REP CEN CONC ANAL EXTR CONCAT
697 43 H P-5 1 35400 ICP EPA +
698 43 H P-5 2 25000 ICP EPA +
699 43 S P-2 1 112000 ICP EPA +
700 43 S P-2 2 99200 ICP EPA +
701 43 L 0-2 1 90 ICP EPA +
702 43 L D-2 2 85 ICP EPA +
703 43 L 0-4 1 125 ICP EPA +
704 43 L D-4 2 100 ICP EPA +
705 43 H 0-11 3980 ICP EPA +
706 43 H 0-12 4620 ICP EPA +
707 43 H D-5 1 3500 ICP EPA +
708 43 H D-5 2 5010 ICP EPA +
709 43 S D-3 1 1010 ICP EPA +
710 43 S 0-3 2 1180 ICP EPA +
711 12 L P-l 1 1810 ICP EPA +
712 12 L P-l 2 1810 ICP EPA +
713 12 L P-4 1 1880 ICP EPA +
714 12 L P-4 2 2010 1C? EPA +
715 12 H P-3 1 40500 ICP EPA +
716 12 H P-3 2 41800 ICP EPA +
717 12 H P-5 1 43300 ICP EPA +
718 12 H P-5 2 46300 ICP EPA +
719 12 S P-2 1 114000 ICP EPA +
720 12 S P-2 2 116000 ICP EPA +
721 12 L 0-2 1 99 ICP EPA +
722 12 L 0-2 2 98 ICP EPA +
723 12 L 0-4 1 128 ICP EPA +
724 12 L 0-4 2 98 ICP EPA +
725 12 H 0-1 1 4870 ICP EPA +
726 12 H D-l 2 5130 ICP EPA +
727 12 H D-5 1 5190 ICP EPA +
728 12 H D-5 2 5580 ICP EPA +
729 12 S 0-3 1 1440 ICP EPA +
730 12 S D-3 2 1490 ICP EPA +
731 37 L P-l 1 1510 AA EPA +
732 37 L P-l 2 2010 AA EPA +
733 37 L P-4 1 2053 AA EPA +
734 37 L P-4 2 1640 AA EPA +
735 37 H P-3 1 34600 AA EPA +
736 37 H P-3 2 40800 AA EPA +
737 37 H P-5 1 37600 AA EPA +
738 37 H P-5 2 41900 AA EPA +
739 37 S P-2 1 110200 AA EPA +
740 37 S P-2 2 117300 AA EPA +
741 37 I D-2 1 115 AA EPA +
742 37 L D-2 2 117 AA EPA +
743 37 L D-4 1 116 AA EPA +
744 37 L D-4 2 103 AA EPA +
745 37 H D-l 1 4920 AA EPA +
746 37 H 0-12 4340 AA EPA +
747 37 H D-5 1 4630 AA EPA +
748 37 H D-5 2 4500 AA EPA +
749 37 S D-3 1 1060 AA EPA +
750 37 S D-3 2 1140 AA EPA +
751 32 L P-l 1 1500 AA EPA +
752 32 L P-l 2 1900 AA EPA +
753 32 L P-4 1 2300 AA EPA +
754 32 L P-4 2 2000 AA EPA +
-------
raw data file 14
OBS BA LEVEL SAM REP CEN CONC ANAL EXTR CONCAT
755 32 H P-3 1 52000 AA EPA +
755 32 H P-3 2 44000 AA EPA +
757 32 H P-5 1 45000 AA EPA +
758 32 H P-5 2 45000 AA EPA +
759 32 S P-2 1 164000 AA EPA +
760 32 S P-2 2 143000 AA EPA +
761 32 L D-2 1 90 AA EPA +
762 32 L D-2 2 91 AA EPA +
763 32 L D-4 1 90 AA EPA +
764 32 L D-4 2 100 AA EPA +
765 32 H 0-1 1 4800 AA EPA +
766 32 H 0-1 2 5300 AA EPA +
767 32 H 0-5 1 5100 AA EPA +
768 32 H 0-5 2 5400 AA EPA +
769 32 S 0-3 1 1100 AA EPA +
770 32 S 0-3 2 1200 AA EPA +
771 17 L P-l 1 1920 AA EPA +
772 17 L P-l 2 1720 AA EPA +
773 17 L P-4 1 2050 AA EPA +
774 17 L P-4 2 1740 AA EPA +
775 17 H P-3 1 41500 AA EPA +
776 17 H P-3 2 41300 AA EPA +
777 17 H P-5 1 42700 AA EPA +
778 17 H P-5 2 43600 AA EPA +
779 17 S P-2 1 131000 AA EPA +
780 17 S P-2 2 126000 AA EPA +
781 17 L D-2 1 130 AA EPA +
782 17 L D-2 2 130 AA EPA +
783 17 L D-4 1 140 AA EPA +
784 17 L D-4 2 140 AA EPA +
785 17 H D-l 1 4720 AA EPA +
786 17 H D-l 2 4930 AA EPA +
787 17 H 0-5 1 4800 AA EPA +
788 17 H D-5 2 5040 AA EPA +
789 17 S 0-3 1 1340 AA EPA +
790 17 S 0-3 2 1340 AA EPA +
791 7 L P-l 1 1801 AA EPA +
792 7 L P-l 2 1735 AA EPA +
793 7 L P-4 1 2165 AA EPA +
794 7 L P-4 2 2280 AA EPA +
795 7 H P-3 1 37700 AA EPA +
796 7 H P-3 2 39430 AA EPA +
797 7 H P-5 1 22440 AA EPA +
798 7 H P-5 2 22640 AA EPA +
799 7 S P-2 1 90520 AA EPA +
800 7 S P-2 2 106200 AA EPA +
801 7 L D-2 1 451 AA EPA +
802 7 L D-2 2 465 AA EPA +
803 7 L 0-41 539 AA EPA +
804 7 L 0-42 567 AA EPA +
805 7 H D-l 1 4155 AA EPA +
806 7 H D-l 2 4956 AA EPA +
807 7 H D-5 1 3929 AA EPA +
808 7 H D-5 2 4187 AA EPA +
809 7 S D-3 1 1648 AA EPA +
810 7 S D-3 2 1674 AA EPA +
811 41 L P-l 1 2130 AA EPA +
812 41 L P-l 2 2250 AA EPA +
-------
raw data file 15
OBS BA LEVEL SAM REP CEN CONC ANAL EXTR CONCAT
813 41 L P-4 1 2370 AA EPA +
814 41 L P-4 2 1960 AA EPA +
815 41 H P-3 1 43700 AA EPA +
816 41 H P-3 2 42300 AA EPA +
817 41 H P-5 1 41600 AA EPA +
818 41 H P-5 2 40200 AA EPA +
819 41 S P-2 1 130000 AA EPA +
820 41 S P-2 2 129000 AA EPA +
821 41 L D-2 1 99 AA EPA +
822 41 L D-2 2 105 AA EPA +
823 41 L D-4 1 168 AA EPA +
824 41 L D-4 2 97 AA EPA +
825 41 H D-l 1 4920 AA EPA +
826 41 H D-l 2 5450 AA EPA +
827 41 H D-5 1 5180 AA EPA +
828 41 H D-5 2 4970 AA EPA +
829 41 S D-3 1 1280 AA EPA +
830 41 S D-3 2 1300 AA EPA +
831 2 L P-l 1 1773 AA EPA +
832 2 L P-l 2 1669 AA EPA +
833 2 I P-4 1 1576 AA EPA +
834 2 L P-4 2 1522 AA EPA +
835 2 H P-3 1 38312 AA EPA +
836 2 H P-3 2 36048 AA EPA +
837 2 H P-5 1 35498 AA EPA +
838 2 H P-5 2 36621 AA EPA +
839 2 S P-2 1 109414 AA EPA +
840 2 S P-2 2 127416 AA EPA +
841 2 L D-2 1 196 AA EPA +
842 2 L D-2 2 177 AA EPA +
843 2 L D-4 1 97 AA EPA +
844 2 L D-4 2 87 AA EPA +
845 2 H D-l 1 5022 AA EPA +
846 2 H D-l 2 4210 AA EPA +
847 2 H D-5 1 4797 AA EPA +
848 2 H D-5 2 4686 AA EPA +
849 2 S 0-31 1292 AA EPA +
850 2 S D-3 2 1277 AA EPA +
-------
Appendix G-4
Missing/Censored Observations
-------
LEGEND
(Appendix G-4)
OBS = Reported Result
LAB = Laboratory Code
METH = Method Number
EXTR
ANAL
MTX
LEVEL =
CEN
CONG
TRUE
1
2
3
4
5
Microwave/Atomic Absorption Spectrometry
Hotplate/Atomic Absorption Spectrometry
Microwave/Inductively Coupled Plasma Emission Spectrometry
Hotplate/Inductively Coupled Plasma Emission Spectrometry
Laboratory X-Ray Fluorescence
Extraction Method
EPA = EPA/AREAL
NIO = NIOSH Method 7082
Analytical Method
AA = Atomic Absorption Spectrometry
ICP = Inductively Coupled Plasma Emission Spectrometry
XRF = Laboratory X-Ray Fluorescence
Matrix
P =
D =
Paint
Dust
Concentration Level
L = Low
H = High
S = Standard Reference Material (SRM)
Censored as less than or greater than the concentration reported (ug/g)
Concentration reported (ug/g)
Preliminary calculation of consensus value (without exclusion of outliers)
-------
DBS
LAB METH
Missing or censored observations
EXTR ANAL MIX LEVEL CEN
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
38
46
30
38
38
49
49
49
44
44
44
44
45
50
50
50
50
24
41
41
48
48
48
48
28
28
28
28
50
50
3
4
3
3
3
4
4
4
4
4
4
4
4
5
5
5
5
2
4
4
4
4
4
4
2
2
2
2
5
5
EPA
NIO
EPA
EPA
EPA
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
N/A
N/A
N/A
N/A
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
NIO
N/A
N/A
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
XRF
XRF
XRF
XRF
AA
ICP
ICP
ICP
ICP
ICP
ICP
AA
AA
AA
AA
XRF
XRF
P
P
P
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
0
D
D
D
P
P
L
S
L
L
L
L
L
L
L
I
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
S
S
CONC
TRUE
10
22
22
34
35
40
50
50
50
50
50
75
75
75
75
100
100
100
200
200
200
200
300
300
300
300
50000
50000
1680
113200
1680
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
113200
113200
-------
Appendix G-5
Candidate Outlying Observations
-------
LEGEND
(Appendix G-5)
OBS
LAB
METH
= Reported Result
= Laboratory Code
= Method Number
EXTR
ANAL
MTX
LEVEL
TRUE
CONC
REC
SCOREREC =
1
2
3
4
5
Microwave/Atomic Absorption Spectrometry
Hotplate/Atomic Absorption Spectrometry
Microwave/Inductively Coupled Plasma Emission Spectrometry
Hotplate/Inductively Coupled Plasma Emission Spectrometry
Laboratory X-Ray Fluorescence
Extraction Method
EPA = EPA/AREAL Method
NIO = NIOSH Method 7082
Analytical Method
AA = Atomic Absorption Spectrometry
ICP = Inductively Coupled Plasma Emission Spectrometry
XRF = Laboratory X-Ray Fluorescence
Matrix
P = Paint
D = Dust
Concentration Level
L = Low
H = High
S = Standard Reference Material (SRM)
Preliminary calculation of consensus value (without exclusion of outliers)
Concentration reported (ug/g)
Calculated recovery - ratio of reported concentration to the nominal
concentration
The recovery score calculated by subtracting the average recovery
(method/matrix/level) from the calculated recovery (REC) and dividing by
the standard deviation of recovery for a given method/matrix/level
-------
Candidate outlying observations
OBS LAB METH EXTR ANAL MIX LEVEL
TRUE
CONG
REG SCOREC
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
20
20
14
33
1
33
2
15
15
22
1
25
25
4
4
15
4
17
17
2
22
24
21
25
3
5
3
23
9
6
25
19
14
26
7
23
23
1
3
20
21
1
7
1
20
21
27
15
15
15
15
14
14
14
14
2
2
5
3
3
3
2
1
1
2
3
4
4
2
2
1
2
4
4
2
2
3
4
4
2
1
2
3
3
3
3
2
5
1
4
3
3
4
2
2
3
4
4
4
2
4
3
1
1
1
1
5
5
5
5
NIO
NIO
N/A
EPA
EPA
EPA
NIO
EPA
EPA
NIO
EPA
NIO
NIO
NIO
NIO
EPA
NIO
NIO
NIO
NIO
NIO
EPA
NIO
NIO
NIO
EPA
NIO
EPA
EPA
EPA
EPA
NIO
N/A
EPA
NIO
EPA
EPA
NIO
NIO
NIO
EPA
NIO
NIO
NIO
NIO
NIO
EPA
EPA
EPA
EPA
EPA
N/A
N/A
N/A
N/A
AA
AA
XRF
ICP
ICP
ICP
AA
AA
AA
AA
ICP
ICP
ICP
AA
AA
AA
AA
ICP
ICP
AA
AA
ICP
ICP
ICP
AA
AA.
AA
ICP
ICP
ICP
ICP
AA
XRF
AA
ICP
ICP
ICP
ICP
AA
AA
ICP
ICP
ICP
ICP
AA
ICP
ICP
AA
AA
AA
AA
XRF
XRF
XRF
XRF
P
P
D
D
P
D
D
P
P
P
P
D
P
D
D
D
P
P
P
D
P
D
D
D
D
P
P
D
D
P
P
D
D
D
P
D
P
P
P
D
D
D
P
D
D
D
D
D
D
D
D
P
P
P
P
S
S
H
L
H
H
H
H
H
S
S
H
H
S
S
H
L
L
S
H
S
H
L
L
H
S
H
S
L
S
L
L
L
L
H
S
H
L
H
L
L
S
H
H
L
L
L
L
L
L
L
L
L
L
L
113200
113200
4534
100
36611
4534
4534
36611
36611
113200
113200
4534
36611
1192
1192
4534
1680
1680
113200
4534
113200
4534
100
100
4534
113200
36611
1192
100
113200
1680
100
100
100
36611
1192
36611
1680
36611
100
100
1192
36611
4534
100
100
100
100
100
100
100
1680
1680
1680
1680
5077
14010
1400
31
25000
2500
3191
22440
22640
46000
99200
2670
28600
960
960
3929
1080
1070
46900
3531
55000
5640
160
160
5740
164000
43600
1440
145
135000
2120
171
137
196
44340
1490
46300
2330
46300
199
170
1830
47300
7150
214
270
211
451
465
539
567
5148
5434
5823
6003
0.04
0.12
0.31
0.31
0.68
0.55
0.70
0.61
0.62
0.41
0.88
0.59
0.78
0.81
0.81
0.87
0.64
0.64
0.41
0.78
0.49
1.24
1.60
1.60
1.27
1.45
1.19
1.21
1.45
1.19
1.26
1.71
1.37
1.96
1.21
1.25
1,26
1.39
1.26
1.99
1.70
1.54
1.29
1.58
2.14
2.70
2.11
4.51
4.65
5.39
5.67
3.06
3.23
3.47
3.57
-4.12
-3.75
-3.57
-3.43
-3.38
-2.79
-2.68
-2.64
-2.61
-2.42
-2.42
-2.34
-2.20
-2.19
-2.19
-2.14
-2.12
-2.07
-2.06
-2.05
-2.05
2.00
2.04
2.04
2.05
2.12
2.13
2.14
2.16
2.20
2.20
2.25
2.35
2.45
2.52
2.54
2.62
2.69
3.03
3.29
3.38
3.39
3.40
3.50
3.84
5.22
5.40
10.62
11.07
13.44
14.34
23.44
25.07
27.28
28.31
-------
Appendix G-6
Method Means, Consensus Values,
Repeatability and Reproducibility
-------
Results of Statistical Analysis
MIX LEVEL METH
SW
SB STOT MEAN
L95
U95 TRUE LT95 UT95 N NO K KO
D
0
0
0
D
0
D
D
0
D
D
D
D
D
D
P
P
P
P
P
P
P
P
P
P
P
P
P
P
P
H
H
H
H
H
L
L
L
L
L
S
S
S
S
S
H
H
H
H
H
L
L
L
L
L
S
S
S
S
S
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
296
441
501
574
98
23
14
17
23
8
38
45
55
36
24
2386
1860
1445
1708
984
217
200
196
154
34
8829
5934
5159
11239
1582
301
214
226
311
305
8
17
0
25
20
188
89
60
129
84
3150
1920
1880
2377
4118
128
197
91
196
185
16043
24194
6469
20789
32457
422
491
549
653
320
25
23
17
34
22
192
100
82
134
88
3951
2674
2372
2927
4234
252
281
216
249
188
18312
24911
8274
23633
32496
4847
4677
4281
4397
2485
114
108
98
98
93
1327
1173
1133
1112
1029
41281
36921
36654
35670
27404
1896
1661
1603
1600
1034
122432
104340
118281
94382
112721
4599
4475
4059
4136
2257
102
95
92
79
77
1186
1111
1086
1031
965
38780
35523
35336
34109
24332
1772
1517
1514
1470
885
109679
87325
113429
80620
86735
5095
4879
4503
4657
2714
125
121
104
116
109
1468
1235
1180
1194
1093
43782
38318
37972
37231
30476
2020
1806
1692
1730
1182
135185
121356
123133
108143
138708
4550
4550
4550
4550
4550
104
104
104
104
104
1186
1186
1186
1186
1186
37632
37632
37632
37632
37632
1690
1690
1690
1690
1690
109859
109859
109859
109859
109859
4316
4316
4316
4316
4316
91
91
91
91
91
1096
1096
1096
1096
1096
35872
35872
35872
35872
35872
1567
1567
1567
1567
1567
96964
96964
96964
96964
96964
4785
4785
4785
4785
4785
117
117
117
117
117
1277
1277
1277
1277
1277
39391
39391
39391
39391
39391
1814
1814
1814
1814
1814
122753
122753
122753
122753
122753
28 28
35 36
35 36
41 42
27 28
23 28
29 36
31 36
27 42
24 28
14 14
18 18
18 18
20 21
14 14
26 28
35 36
35 36
41 42
28 28
28 28
36 36
34 36
41 42
24 28
14 14
16 18
18 18
20 21
12 14
7 7
9 9
9 9
10 10
7 7
6 7
8 9
9 9
9 10
6 7
7 7
9 9
9 9
10 10
7 7
7 7
9 9
9 9
10 10
7 7
7 7
9 9
9 9
10 10
6 7
7 7
8 9
9 9
10 10
6 7
LEGEND
MTX =
Level =
Meth =
SW =
STOT =
SB =
MEAN =
L95, U95 =
TRUE =
LT95, UT95
N =
N0 =
K =
K =
Matrix (D=Dust; P=Paint)
H=High; L=Low; S=SRM
Method (1=MW/AAS; 2=HP/AAS; 3=MW/ICP; 4=HP/1CP; 5=Lab XRF)
Repeatability (within-lab standard deviation)
ReproducibiKty (within-lab and between-lab standard deviation)
Pure between-lab standard deviation
Method Mean
Lower and Upper Limits of 95% Confidence Interval of the Method Mean
Consensus Value (average of means of methods 1 through 4)
Lower and Upper Limits of 95% Confidence Interval of Consensus Value
Total sample size
Expected sample size
Number of labs for nonmissing, noncensored, and nonoutlying data
Expected number of labs
-------
Appendix G-7
Recovery and Log of Recovery Plots
by Laboratory
-------
LEGEND
(Appendix G-7)
D = Dust (low dust and high dust)
E= "Dust" SRM 2711
P = Paint (low paint and high paint)
Q = Paint SRM 1579
-------
Appendix G-7-1
MW/AAS Laboratories
-------
2.00 +
1.75
1.50 +
1.25
0.75 +
0.50
0.25
0.00
plotrec.sas 7:49 Monday, August 17, 1992 1
METH=1 LAB=10 -
Plot of REC*LOGTRUE. Symbol is value of MIX.
r
e
c
0
v 1.00 + D E
e
r
y
DO E
P D
P P
1.5 2.0
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal cone (ppm)
5.0 5.5
NOTE: 2 obs hidden.
-------
5.5
5.0
4.5 +
plotcon.sas 7:38 Monday, August 17, 1992 1
METH=1 LAB=10 -
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
Q
Q
P
P P
1
0
g
4.0 +
0
f
r
e
0 D
D D
p 3.5 +
0
r
t
e
d
P
P
P
E
3.0 + E
P
P
M
2.5
2.0 +
1.5 +
D D
0 D
1.5 2.0
2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal PPM
NOTE: 2 obs hidden.
-------
2.00
1.75
1.50
1.25
0.75
0.50
0.25
0.00 +
plotrec.sas 7:49 Monday, August 17, 1992 2
METH=1 LAB-11 -
Plot of REC*IOGTRUE. Symbol is value of MIX.
r
e
c
0
D 0
D
v 1.00 +
e
r
y
P P
EP
D
D
D
D
P P
P
P
Q
Q
1.5 2.0
NOTE: 1 obs hidden.
2.5 3.0 3.5 4.0 4.5 5.0
base 10 log of nominal cone (ppm)
5.5
-------
5.5 +
5.0
4.5
plotcon.sas 7:38 Monday, August 17, 1992 2
METH=1 LAB=11
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
P P
P
1
0
g
4.0 +
0
f
r
e
p 3.5 +
0
r
t
e
d
P P
P
P
E
3.0 + E
P
P
M
D D
2.5
2.0 +
1.5 +
D D
D
1.5 2.0
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal PPM
5.0 5.5
NOTE: 5 obs hidden.
-------
2.00
1.75
1.50
0.75
0.50
0.25
0.00 +
plotrec.sas 7:49 Monday, August 17, 1992 3
METH=1 LAB=12 —
Plot of REC*LOGTRUE. Symbol is value of HTX.
* *
1.25 •
r
e
c
0
E
EP
P
•
P D P
P P
v 1.00 +
e
r
y
D D
D
1.5 2.0
NOTE: 3 obs hidden.
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal cone (ppra)
5.0 5.5
-------
5.5
5.0
4.5 +
plotcon.sas 7:38 Monday, August 17, 1992 3
• -,•* — -.•.*.«. — «- — — -. — -.•— — •» fit I n"* J. L.MD~~Lc. — — — —— — — -•i-' — — — — -• — — — -• — «-•* — '-•-•--••• — — — •• — -••-
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
Q
Q
1
0
g
4.0 +
0
f
r
e
D
D D
p 3.5 +
0
r
t
e
d
P
E P
3.0 +
P
P
M
D
2.5
2.0 +
1.5
1.5 2.0
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal PPM
5.0 5.5
NOTE: 8 obs hidden.
-------
plotrec.sas 7:49 Monday, August 17, 1992 4
- METH=1 LAB=13 - -
Plot of REC*LOGTRUE. Symbol is value of MIX.
2.00 +
1.75
1.50
1.25 +
r
e
c
o
v 1.00
e
r
y
0.75 +
0.50
0.25
0.00
D D
E
EP
P
P
D
D D
D
P
P
P
Q
Q
1.5 2.0
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal cone (ppm)
5.0 5.5
NOTE: 3 obs hidden.
-------
5.5
5.0
4.5
plotcon.sas 7:38 Monday, August 17, 1992 4
METH=1 LAB=13
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
1
0
g
4.0 +
0
f
r
e
D D
0
p 3.5 +
0
r
t
e
d
P
P P
E
3.0 +
P
P
M
2.5
2.0
1.5
D
D D
1.5 2.0
NOTE: 7 obs hidden.
2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal PPM
-------
2.00
1.75
1.50
1.25
0.75
0.50
0.25
0.00
plotrec.sas 7:49 Monday, August 17, 1992 5
- METH=1 LAB=14
Plot of REC*LOGTRUE. Symbol is value of MTX.
r
e
c
0
D
E
P D
v 1.00 + P
e
r
y
D P
P D
D
P
P P
P
1.5 2.0
NOTE: 2 obs hidden.
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal cone (ppm)
5.0 5.5
-------
plotcon.sas 7:38 Monday, August 17, 1992 5
METH=1 LAB=14
Plot of LOGCONC*LGGTRUE. Symbol is value of MIX.
5.5
5.0 +
4.5 +
1
o
g
0
f
4.0
r
e
p 3.5 +
o
r
t
e
d
3.0 +
P
P
M
2.5
2.0 +
1.5
D
D
D
D
D D
D
Q
Q
p p
1.5 2.0
2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal PPM
NOTE: 6 obs hidden.
-------
2.00 +
1.75
1.50
1.25
0.75
0.50
0.25
0.00
plotrec.sas 7:49 Monday, August 17, 1992 6
METH=1 LAB=15 -
Plot of REC*LOGTRUE. Symbol is value of MIX.
r
e
c
0
D
v 1.00 + D
e
r
y
D
p
p
D
0
D D
P
P
P
P
Q
Q
1.5 2.0
NOTE: 1 obs hidden.
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal cone (ppm)
5.0 5.5
-------
plotcon.sas 7:38 Monday, August 17, 1992 6
METH=1 LAB=15
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
5.5
5.0 -f-
4.5 +
1
o
g
0
f
4.0 +
r
e
p 3.5
o
r
t
e
d
3.0 +
P
P
M
2.5 +
2.0
1.5 +
D 0
D D
D D
P P
1.5 2.0
2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal PPM
NOTE: 7 obs hidden.
-------
plotrec.sas 7:49 Monday, August 17, 1992 7
METH=1 LAB=16 —
Plot of REC*LOGTRUE. Symbol is value of MIX.
2.00
1.75
1.50
1.25 +
r
e
c
o
v 1.00 +
e
r
y
0.75
0.50
0.25
0.00
P
E P
P P
D
D
D D
P
P
Q
Q
1.5 2.0
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal cone (ppm)
5.0 5.5
NOTE: 4 obs hidden.
-------
5.5
5.0 +
4.5
plotcon.sas 7:38 Monday, August 17, 1992 7
METH=1 LA8=16
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
0
g
4.0 +
0
f
r
e
p 3.5 +
0
r
t
e
d
P
P P
E
3.0 +
P
P
M
D D
2.5
2.0 +
1.5
1.5 2.0
NOTE: 9 obs hidden.
2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal PPM
-------
Appendix G-7-2
HP/AAS Laboratories
-------
2.00
1.75
1.50
1.25
0.50
0.25
0.00
plotrec.sas 7:49 Monday, August 17, 1992 8
METH=2 LAB=20 -
Plot of REC*LOGTRUE. Symbol is value of MIX.
r
e
c
0
v 1.00 -i- E
e
r
y
D P
D P
D P
0.75 +
P
Q
Q
1.5 2.0
2.5 3.0 3.5 4.0 4.5 5.0
base 10 log of nominal cone (ppm)
5.5
NOTE: 1 obs hidden.
-------
plotcon.sas 7:38 Monday, August 17, 1992 8
METH=2 LAB=20
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
5.5
5.0
4.5
P P
P P
1
0
g
4.0 +
0
f
r
e
p 3.5 +
0
r
t
e
d
P P
E P
3.0 +
P
P
M
D
D D
D
2.5
2.0 +
1.5
D D
D
1.5
2.0
2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal PPM
NOTE: 4 obs hidden.
-------
2.00
1.75
1.50 +
1.25 +
r
e
c
o
v 1.00 +
e
r
y
0.75
0.50
0.25
0.00
plotrec.sas 7:49 Monday, August 17, 1992 9
METH=2 LAB=21
Plot of REC*LOGTRUE. Symbol is value of MIX.
D D
p
p
E P
E
P
D
D D
D
*
P
P
P
Q
Q
1.5
2.0 2.5 3.0 3.5 4.0 4.5 5.0
5.5
base 10 log of nominal cone (ppm)
-------
plotcon.sas 7:38 Monday, August 17, 1992 9
- METH=2 LAB=21 -
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
5.5
5.0 +
4.5
o
g
0
f
4.0 +
r
e
p 3.5 +
o
r
t
e
d
3.0 4-
P
P
M
2.5
2.0 +
1.5
D D
0 D
D D
D
Q
Q
1.5 2.0
2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal PPM
NOTE: 6 obs hidden.
-------
2.00
1.75
1.50
1.25
0.75
0.50
0.25 +
0.00 +
plotrec.sas 7:49 Monday, August 17, 1992 10
METH=2 LAB=22 — —
Plot of REC*LOGTRUE. Symbol is value of MIX.
r
e
c
0
v 1.00 + D
e
r
y
0 D
D
E P
P
P
D
D D
D
P P
P P
Q
Q
1.5 2.0
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal cone (ppm)
5.0 5.5
NOTE: 1 obs hidden.
-------
5.5
5.0
4.5
plotcon.sas 7:38 Monday, August 17, 1992 10
— METH=2 LAB=22 - - --
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
P P
P P
1
0
g
4.0 +
0
f
r
e
p 3.5 +
0
r
t
e
d
P
P P
3.0 + EP
P
P
M
D D
2.5 +
2.0
1.5
D D
1.5 2.0
NOTE: 6 obs hidden.
2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal PPM
-------
2.00
1.75
1.50
1.25
0.75
0.50
0.25
0.00
plotrec.sas 7:49 Monday, August 17, 1992 11
• METH=2 LAB=23 -
Plot of REC*LOGTRUE. Symbol is value of MIX.
r
e
c
0
D
v 1.00 + D
e
r
y
D
E P
EP P
P P
1.5 2.0
NOTE: 4 obs hidden.
2.5 3.0 3.5 4.0 4.5 5.0
base 10 log of nominal cone (ppm)
+_„
5.5
-------
5.5
5.0
4.5
plotcon.sas 7:38 Monday, August 17, 1992 11
METH=2 LAB=23
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
Q
Q
P
P P
1
0
9
4.0 +
0
f
r
e
p 3.5 +
0
r
t
e
d
E
3.0 + E
P
P
M
D D
P P
2.5
2.0
1.5
D D
1.5 2.0
NOTE: 7 obs hidden.
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal PPM
5.0 5.5
-------
2.00
1.75
1.50
1.25
0.75
0.50 +
0.25
0.00 +
plotrec.sas 7:49 Monday, August 17, 1992 12
METH=2 LAB=24 ---
Plot of REC*LOGTRUE. Symbol is value of MIX.
r
e
c
0
P
P
D P P P
E
v 1.00 + < D P Q
e
r
y
E D
P P Q
0
1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
base 10 log of nominal cone (ppm)
5.5
-------
5.5
5.0 +
4.5
plotcon.sas 7:38 Monday, August 17, 1992 12
- METH=2 LAB=24 - - —
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
Q
Q
P P
P
1
0
g
4.0 +
0
f
r
e
P 3.5 +
0
r
t
e
d
P P
P
E
3.0 +
P
P
M
D
D D
D
2.5 +
2.0
1.5
D
D
D <
1.5 2.0
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal PPM
5.0 5.5
NOTE: 3 obs hidden.
-------
2.00
1.75
1.50
1.25 +
0.75
0.50 +
0.25
0.00
plotrec.sas 7:49 Monday, August 17, 1992 13
- METH=2 LAB=25
Plot of REC*LOGTRUE. Symbol is value of MTX.
r
e
c
0
p P
D
D
p
E
v 1.00 + E P
e
r
y
P P
D D
0
D
1.5 2.0
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal cone (ppm)
5.0 5.5
NOTE: 2 obs hidden.
-------
5.5
5.0 +
4.5
plotcon.sas 7:38 Monday, August 17, 1992 13
— METH=2 LAB=25 —
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
P P
P
1
0
g
4.0 +
0
f
r
e
p 3,5 +
0
r
t
e
d
P P
P
E P
3.0 +
P
P
M
2.5 +
2.0
1.5
D
D
1.5 2.0
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal PPM
5.0 5.5
NOTE: 5 obs hidden.
-------
2.00
1.75
1.50
1.25
0.75
0.50
0.25
0.00
plotrec.sas 7:49 Monday, August 17, 1992 14
METH=2 LAB=26
Plot of REC*LOGTRUE. Symbol is value of MIX.
r
e
c
0
D P
D D
v l.OO-i- EP P P P
e
r
y
P P
Q
Q
1.5 2.0
NOTE: 4 obs hidden.
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal cone (ppm)
5.0 5.5
-------
plotcon.sas 7:38 Monday, August 17, 1992 14
- METH=2 LAB=26
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
5.5 +
5.0
4.5
P P
Q
Q
1
0
g
4.0 +
0
f
r
e
p 3.5 +
0
r
t
e
d
P P
P
E
3.0 +
P
P
M
D D
2.5
2.0
1.5
D
D D
D
1.5 2.0
2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal PPM
NOTE: 6 obs hidden.
-------
2.00
1.75
1.50
1.25
0.75
0.50 +
0.25 +
0.00
plotrec.sas 7:49 Monday, August 17, 1992 15
METH=2 LAB=27 - -
Plot of REC*LOGTRUE. Symbol is value of MIX.
r
e
c
0
0 POD
D D P
P
v 1.00 + E D P P
e
r
y
-P D
Q
Q
1.5 2.0
NOTE: 2 obs hidden.
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal cone (ppm)
5.0 5.5
-------
plotcon.sas 7:38 Monday, August 17, 1992 15
METH=2 LA8=27 -
Plot of LOGCONC*IGGTRUE. Symbol is value of MIX.
5.5
5.0
4.5
1
o
9
o
f
4.0 +
r
e
p 3.5
o
r
t
e
d
3.0 +
P
P
M
2.5
2.0 +
1.5
D
0
D D
D D
D
P P
1.5 2.0
2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal PPM
NOTE: 6 obs hidden.
-------
2.00
1.75
1.50
1.25
0.75
0.50 -t-
0.25
0.00
plotrec.sas 7:49 Monday, August 17, 1992 16
- METH=2 LAB=28 -
Plot of REC*LOGTRUE. Symbol is value of MIX.
r
e
c
o
EP D
P
P D D P P
v 1.00 + E
e
r
y
P
D
1.5 2.0
2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal cone (ppm)
NOTE: 4 obs hidden.
-------
plotcon.sas 7:38 Monday, August 17, 1992 16
METH=2 LAB=28 —
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
5.5
5.0 +
4.5
P P
1
0
g
4.0 +
0
f
r
e
p 3.5 H
0
r
t
e
d
D D
D
•
P
P P
E
3.0 +
P
P
M
2.5
2.0
1.5
1.5 2.0
NOTE: 8 obs hidden.
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal PPM
5.0 5.5
-------
Appendix G-7-3
MW/ICP Laboratories
-------
2.00 +
1.75 +
1.50
1.25
0.75 +
0.50 +
0.25 +
0.00 +
plotrec.sas 7:49 Monday, August 17, 1992 17
— METH=3 LAB=30
Plot of REC*LOGTRUE. Symbol is value of MTX.
r
e
c
0
P
0
v 1.00 + D E D P Q
e
r
y
P P P
D
P D Q
D E
1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal cone (ppm)
-------
5.5
5.0
4.5
2.5 +
2.0
1.5
plotcon.sas 7:38 Monday, August 17, 1992 17
METH=3 LAB=30 -- -
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
1
0
g
4.0 +
0
f
r
e
D D
D D
p 3.5 +
0
r
t
e
d
P
P
P
E
3.0 + E
P
P
M
D
D
Q
Q
P P
P
1.5 2.0
NOTE: 2 obs hidden.
2.5 3.0 3.5 4.0
base 10 log of nominal PPM
1 obs were out of range.
4.5
5.0 5.5
-------
2.00
1.75 +
1.50
1.25
0.75
0.50
0.25
0.00
plotrec.sas 7:49 Monday, August 17, 1992 18
METH=3 LAB=31
Plot of REC*LOGTRUE. Symbol is value of MIX.
r
e
c
0
Q
D Q
P P
v 1.00 + P
e
r
y
D P
D EP D
D EDO
1.5 2.0
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal cone (ppm)
5.0 5.5
NOTE: 1 obs hidden.
-------
5.5 +
5.0
4.5
2.5
2.0
1.5
plotcon.sas 7:38 Monday, August 17, 1992 18
-- METH=3 LAB=31
Plot of LOGCONC*LQGTRUE. Symbol is value of MIX.
1
0
g
4.0 +
0
f
r
e
D D
p 3.5 + D
0
r
t
e
d
P
P
P P
3.0 + E
P
P
M
D D
D D
Q
Q
P P
1.5 2.0
NOTE: 4 obs hidden.
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal PPM
5.0 5.5
-------
2.00
1.75
1.50
1.25
0.75
0.50
0.25
0.00
plotrec.sas 7:49 Monday, August 17, 1992 19
METH=3 LAB=32
Plot of REC*L06TRUE. Symbol is value of MIX.
r
e
c
0
0
D P D
v 1.00 + E
e
r
y
D EP P P
D P
D
P P D
1.5 2.0
NOTE: 2 obs hidden.
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal cone (ppnt)
5.0 5.5
-------
5.5
5.0
4.5 +
plotcon.sas 7:38 Monday, August 17, 1992 19
— METH=3 LAB=32
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
1
0
g
4.0 +
0
f
r
e
D D
D
p 3.5 •*• D
0
r
t
e
d
P
P P
E
3.0 +
P
P
M
2.5
2.0 +
1.5 +
D D
D
1.5 2.0
2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal PPM
NOTE: 6 obs hidden.
-------
2.00
1.75
1.50
1.25
0.75
0.50
0.25
0.00
plotrec.sas 7:49 Monday, August 17, 1992 20
METH=3 LAB=33 - —
Plot of REC*LOGTRUE. Symbol is value of MIX.
r
e
c
0
v 1.00 + D
e
r
y
E
EP
D P
1.5 2.0
NOTE: 7 obs hidden.
2.5 3.0 3.5 4.0 4.5 5.0
base 10 log of nominal cone (ppm)
5.5
-------
plotcon.sas 7:38 Monday, August 17, 1992 20
METH=3 LAB=33 -
Plot of LOGCGNC*LOGTRU£. Symbol is value of MTX.
5.5
5.0
4.5 +
1
o
9
o
f
4.0
r
e
p 3.5 +
o
r
t
e
d
3.0 +
P
P
M
2.5
2.0
1.5 +
D
D
P P
D D
P P
1.5 2.0
2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal PPM
NOTE: 9 obs hidden.
-------
plotrec.sas 7:49 Monday, August 17, 1992 21
METH=3 LAB=34
Plot of REC*LOGTRUE. Symbol is value of MIX.
2.00
1.75
1.50
1.25
r
e
c
o
v 1.00
e
r
y
0.75 +
0.50
0.25
0.00
D
P D D
P D
P
P P
P
1.5 2.0
2.5 3.0 3.5 4.0 4.5 5.0
base 10 log of nominal cone (ppm)
5.5
NOTE: 2 obs hidden.
-------
5.5 +
5.0
4.5
plotcon.sas 7:38 Monday, August 17, 1992 21
METH=3 LAB=34
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
P P
P
1
0
g
4.0 +
0
f
r
e
p 3.5 +
0
r
t
e
d
P P
P P
E
3.0 +
P
P
M
D
D D
2.5
2.0
1.5
1.5 2.0
NOTE: 4 obs hidden.
2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal PPM
-------
2.00
1.75
1.50
1.25
0.50
0.25
0.00
plotrec.sas 7:49 Monday, August 17, 1992 22
- - METH=3 LAB=35
Plot of REC*LOGTRUE. Symbol is value of MTX.
r
e
c
0
P
D P Q
P Q
v 1.00 + P P
e
r
y
D
E P D
D D P
E
0.75 + D
D
1.5 2.0
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal cone (ppm)
5.0 5.5
NOTE: 1 obs hidden.
-------
5.5
5.0
4.5 +
plotcon.sas 7:38 Monday, August 17, 1992 22
• — METH=3 LAB=35 - —
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
P P
1
0
g
4.0 +
0
f
r
e
D
D D
D
p 3.5 +
0
r
t
e
d
P
P P
3.0 + E
P
P
M
2.5
2.0
1.5
D D
D D
1.5
2.0
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal PPM
5.0 5.5
NOTE: 5 obs hidden.
-------
2.00
1.75
1.50 +
1.25
0.75
0.50 +
0.25
0.00
plotrec.sas 7:49 Monday, August 17, 1992 23
METH-3 LAB=36
Plot of REC*LOGTRUE. Symbol is value of MIX.
r
e
c
0
D
E D
v 1.00 + DO
e
r
y
EP P 0
D D
P
P Q
Q
P
1.5 2.0
KOTE: 1 obs hidden.
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal cone (ppm)
5.0 5.5
-------
5.5
5.0
4.5
2.5 +
2.0
1.5
plotcon.sas 7:38 Monday, August 17, 1992 23
- METH=3 LAB=36 -- -
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
1
0
g
4.0 +
0
f
r
e
D
D D
D
p 3,5 +
0
r
t
e
d
P
p p
E
3.0 +
P
P
M
D
D D
D
P P
1.5 2.0
NOTE: 5 obs hidden.
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal PPM
5.0 5.5
-------
2.00
1.75
1.50
1.25 +
0.75
0.50
0.25
0.00
plotrec.sas 7:49 Monday, August 17, 1992 24
METH=3 LA8=37 —
Plot of REC*LOGTRUE. Symbol is value of MTX.
r
e
c
0
v 1.00 + D D
e
r
y
D
0
D
D
D
P P
P
Q
Q
1.5 2.0
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal cone (ppm)
+ +_„
5.0 5.5
NOTE: 1 obs hidden.
-------
5.5
5.0
4.5
plotcon.sas 7:38 Monday, August 17, 1992 24
— METH=3 LAB=37 --
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
P P
1
0
g
4.0 +
0
f
r
e
D D
D
p 3.5 +
0
r
t
e
d
P
P P
P
E
3.0 +
P
P
M
2.5 +
2.0
1.5
D D
1.5 2.0
NOTE: 6 obs hidden.
2.5 3.0 3.5 4.0 4.5
base 10 Tog of nominal PPM
5.0 5.5
-------
2.00
1.75
1.50
1.25
0.75
0.50
0.25
0.00
plotrec.sas 7:49 Monday, August 17, 1992 25
METH=3 LAB=38
Plot of REC*LOGTRUE. Symbol is value of MIX.
r
e
c
0
Q
P P P
D D P Q
v 1.00 + D E P
e
r
y
P D
1.5 2.0
NOTE: 3 obs hidden.
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal cone (ppm)
5.0 5.5
-------
5.5
5.0 t
4.5
2.5
2.0
1.5
plotcon.sas 7:38 Monday, August 17, 1992 25
METH=3 LAB=38 -
Plot of LOGCQNC*LOGTRUE. Symbol is value of MTX.
1
0
g
4.0 +
0
f
r
e
D D
p 3.5 +
0
r
t
e
d
D
P P
P
E
3.0 +
P
P
M
P p
1.5
2,0
NOTE: 5 obs hidden.
2.5 3.0 3.5 4.0
base 10 log of nominal PPM
3 obs were out of range.
4.5
5.0 5.5
-------
Appendix G-7-4
HP/ICP Laboratories
-------
2.00
1.75
1.50
1.25
0.75
0.50
0.25
0.00
plotrec.sas 7:49 Monday, August 17, 1992 26
METH=4 LAB=40
Plot of REC*LOGTRUE. Symbol is value of MIX.
r
e
c
0
P
P
v 1.00 + D
e
r
y
D
D D
1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal cone (ppm)
-------
plotcon.sas 7:38 Monday, August 17, 1992 26
METH=4 LAB=40 - -
Plot of LOGCONC*LOGTRUE. Symbol is value of MTX.
5.5
5.0
4.5
1
o
g
0
f
4.0 -f
r
e
P 3.5
o
r
t
e
d
3.0
P
P
M
2.5
2.0
1.5
Q
Q
P
P
P P
0
D
D 0
P P
EP P
1.5
2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal PPM
-------
2.00
1.75
1.50
1.25 +
0.75
0.50
0.25
0.00
plotrec.sas 7:49 Monday, August 17, 1992 27
METH=4 LAB=41
Plot of REC*LOGTRUE. Symbol is value of MIX.
D
D
r
e
c
0
D P
EP D D P
E P
v 1.00 + < P
e
r
y
P
1.5 2.0
NOTE: 2 obs hidden.
2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal cone (ppm)
-------
5.5
5.0
4.5
plotcon.sas 7:38 Monday, August 17, 1992 27
METH=4 LAB=41 -
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
P P
P P
1
0
g
4.0 +
0
f
r
e
D
D
p 3.5 + D
0
r
t
e
d
P P
P
E
3.0 +
P
P
M
2.5
2.0
1.5
1.5 2.0
NOTE: 6 obs hidden.
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal PPM
5.0 5.5
-------
plotrec.sas 7:49 Monday, August 17, 1992 28
METH=4 LAB=42 —
Plot of REC*LOGTRUE. Symbol is value of MIX.
2.00
1.75
1.50
1.25
r
e
c
o
v 1.00 +
e
r
y
0.75 +
0.50
0.25
0.00 +
D D
EP P
E
P
D
D
D
P P
P
Q
Q
1.5 2.0
NOTE: 2 obs hidden.
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal cone (ppm)
5.0 5.5
-------
5.5
5.0 +
4.5 +
plotcon.sas 7:38 Monday, August 17, 1992 28
METH=4 LAB=42 -
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
P P
1
0
g
4.0 +
0
f
r
e
p 3.5 +
0
r
t
e
d
P P
P P
E
3.0 + E
P
P
M
D D
D 0
2.5
2.0 +
1.5
D
D
1.5 2.0
NOTE: 4 obs hidden.
2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal PPM
-------
plotrec.sas 7:49 Monday, August 17, 1992 29
METH=4 LAB=43 -
Plot of REC*LOGTRUE. Symbol is value of MIX.
2.00
1.75
1.50
1.25
r
e
c
o
v 1.00 +
e
r
y
0.75
0.50
0.25
0.00
P
EP
D
D
P P
Q
Q
1.5 2.0
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal cone (ppm)
5.0
5.5
NOTE: 6 obs hidden.
-------
plotcon.sas 7:38 Monday, August 17, 1992 29
- METH=4 LAB=43 — — -
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
5.5
5.0
4.5
1
o
g
0
f
4.0
r
e
p 3.5 +
o
r
t
e
d
3.0 +
P
P
M
2.5
2.0
1.5
D D
D D
D
P P
1.5 2.0
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal PPM
5.0 5.5
NOTE: 9 obs hidden.
-------
plotrec.sas 7:49 Monday, August 17, 1992 30
METH=4 LA8=44
Plot of REC*LQGTRUE. Symbol is value of MIX.
2.00
1.75
1.50
1.25
r
e
c
o
v 1.00
e
r
y
0.75 +
0.50
0.25
0.00
EP
E P
P
P
P
D
D D
P P
P P
Q
Q
1.5 2.0
2.5 3.0 3.5 4.0 4.5 5.0
base 10 log of nominal cone (ppm)
5.5
NOTE: 8 obs hidden.
-------
5.5
5.0
4.5
plotcon.sas 7:38 Monday, August 17, 1992 30
METH=4 LAB=44 -- -
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
P
P P
Q
Q
1
0
g
4.0 +
0
f
r
e
p 3.5 +
0
r
t
e
d
p p
P
3.0 + E P
P
P
M
D
0 D
2.5
2.0 +
1.5 +
D
< <
1.5 2.0
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal PPM
5.0 5.5
NOTE: 14 obs hidden.
-------
2.00
1.75
1.50
1.25 +
0.75 +
0.50
0.25
0.00 +
plotrec.sas 7:49 Monday, August 17, 1992 31
METH=4 LAB=45
Plot of REC*LOGTRUE. Symbol is value of MIX.
r
e
c
0
P
P
P
v 1.00 + E
e
r
y
E
P
P P
1.5 2.0
NOTE: 3 obs hidden.
2.5 3.0 3.5 4.0 4.5 5.0
base 10 log of nominal cone (ppm)
5.5
-------
5.5
5.0 +
4.5
plotcon.sas 7:38 Monday, August 17, 1992 31
METH=4 LAB=45
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
Q
Q
P P
1
0
g
4.0 +
0
f
r
e
D D
D
p 3.5 +
0
r
t
e
d
P
P P
E
3.0 +
P
P
M
2.5 -
-
D
D D
2.0 +
1.5
1.5 2.0
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal PPM
5.0 5.5
NOTE: 5 obs hidden.
-------
plotrec.sas 7:49 Monday, August 17, 1992 32
METH=4 LAB=46 —
Plot of REC*LOGTRUE. Symbol is value of MIX.
2.00
1.75 +
1.50
1.25 +
r
e
c
o
v 1.00 +
e
r
y
0.75 +
0.50
0.25 +
0.00
P D
P P
E P D
P
P P
P
1.5 2.0
2.5 3.0 3.5 4.0 4.5 5.0
base 10 log of nominal cone (ppm)
5.5
NOTE: 2 obs hidden.
-------
plotcon.sas 7:38 Monday, August 17, 1992 32
METH=4 LAB=46
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
5.5
5.0
4.5 +
1
o
g
0
f
4.0 +
r
e
p 3.5 +
o
r
t
e
d
3.0 +
P
P
M
2.5
2.0
1.5
D
D
0 D
D
D
P P
P P
P
1.5 2.0
2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal PPM
NOTE: 5 obs hidden.
-------
2.00
1.75
1.50
1.25
0.75
0.50
0.25
0.00
plotrec.sas 7:49 Monday, August 17, 1992 33
HETH=4 LAB=47
Plot of REC*LOGTRUE. Symbol is value of MTX.
r
e
c
0
D
p
D
v 1.00 + D EP D D
e
r
y
P 0
P
P p
1.5 2.0
NOTE: 4 obs hidden.
2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal cone (ppm)
-------
5.5
5.0
4.5
1
o
g
0
f
4.0
r
e
p 3.5 +
o
r
t
e
d
3.0
P
P
M
2.5
2.0
plotcon.sas 7:38 Konday, August 17, 1992 33
- - METH=4 LAB=47 — --
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
1.5 +
1.5 2.0
D
D D
P P
2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal PPM
NOTE: 8 obs hidden.
-------
plotrec.sas 7:49 Monday, August 17, 1992 34
METH=4 LAB=48
Plot of REC*LOGTRUE. Symbol is value of MIX.
2.00
1.75
1.50
1.25 +
r
e
c
0
D
E
v 1.00 + E P P P
e
r
y
0.75 •
P D P
P
P 0 D
Q
Q
0.50 +
0.25 +
0.00
1.5 2.0
NOTE: 3 obs hidden.
2.5 3.0 3.5 4.0 4.5 5.0
base 10 log of nominal cone (ppm)
4---
5.5
-------
plotcon.sas 7:38 Monday, August 17, 1992 34
METH=4 LA8=48 -
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
5.5 +
5.0
4.5
1
o
g
0
f
4.0 +
r
e
p 3.5 +
o
r
t
e
d
3.0 +
P
P
M
2.5
2.0
1.5
D D
D D
Q
Q
P P
-4*—« — "^™-™^"^ — -' — — — —H~—™»^»«-^—4*—-™ — — — -™+—™ — "- — — —+—- — — — — — -H-— — "- — — — —+- — — •"' — •-•—— 4"
1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal PPM
NOTE: 7 obs hidden.
-------
plotrec.sas 7:49 Monday, August 17, 1992 35
- METH=4 LAB=49 -
Plot of REC*LOGTRUE. Symbol is value of MIX.
2.00
1.75
1.50
1.25
r
e
c
0
D
v 1.00 + P P
e
r
y
P D P P
E
E P DO
0.75
0.50
0.25
0.00
1.5 2,0
NOTE: 1 obs hidden.
2.5 3.0 3.5 4.0 4.5 5.0
base 10 log of nominal cone (ppm)
4-.
5.5
-------
5.5
5.0
4.5 +
2.5 +
2.0
1.5
plotcon.sas 7:38 Monday, August 17, 1992 35
- - METH=4 LAB=49 -
Plot of LOGCQNC*LOGTRUE. Symbol is value of MIX.
1
0
g
4.0 +
0
f
r
e
D D
D D
p 3.5 +
0
r
t
e
d
P
P
P P
3.0 + E
P
P
M
Q
Q
P P
1.5 2.0
NOTE: 3 obs hidden.
2.5 3.0 3.5 4.0
base 10 log of nominal PPM
4.5 5.0 5.5
-------
Appendix G-7-5
Laboratory XRF Laboratories
-------
plotrec.sas 7:49 Monday, August 17, 1992 36
METH=5 LAB=50
Plot of REC*LOGTRUE. Symbol is value of MIX.
2.00
1.75
1.50
1.25 +
r
e
c
o
v 1.00 +
e
r
y
0.75
0.50
0.25
0.00
P P
P P
P P
D D
D D
1.5 2.0
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal cone (ppm)
5.0 5.5
NOTE: 6 obs hidden.
-------
5.5
5.0 +
4.5
0
g
0
f
4.0 +
r
e
p 3.5 +
o
r
t
e
d
3.0 +
P
P
M
2.5 +
2.0 +
1.5
plotcon.sas 7:38 Monday, August 17, 1992 36
• — METH=5 LAB=50
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
p p
D D
D
P P
1.5 2.0
NOTE: 9 obs hidden.
2.5 3.0 3.5 4.0 4.5
base 10 !og of nominal PPM
5.0 5.5
-------
plotrec.sas 7:49 Monday, August 17, 1992 37
METH=5 LAB=51
Plot of REC*LOGTRUE. Symbol is value of MIX.
2.00
1.75
1.50
1.25
r
e
c
0
v 1.00 •
e
r
y
0.75 •
•
E
D E
D
D D P P
P P P
D D P
D D
0.50
0.25
0.00
1.5 2.0
NOTE: 3 obs hidden.
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal cone (ppm)
5.0 5.5
-------
plotcon.sas 7:38 Monday, August 17, 1992 37
METH=5 LA8=51
Plot of LOGCONC*I_OGTRUE. Symbol is value of MIX.
5.5 4-
5.0
4.5 +
1
o
g
0
f
4.0
r
e
p 3.5 +
o
r
t
e
d
3.0 +
P
P
M
2.5 4-
2.0 +
1.5
0 0
D D
P P
P
D D
P P
1.5 2.0
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal PPM
5.0 5.5
NOTE: 7 obs hidden.
-------
plotrec.sas 7:49 Monday, August 17, 1992 38
— METH=5 LAB=52
Plot of REC*LOGTRUE. Symbol is value of MIX.
2.00
1.75
1.50 +
1.25
r
e
c
o
v 1.00 +
e
r
y
0.75
0.50 +
0.25
0.00
D
D D
D
P P
D
0 D
Q
Q
P
P P
1.5 2.0
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal cone (ppm)
5.0 5.5
NOTE: 5 obs hidden.
-------
plotcon.sas 7:38 Monday, August 17, 1992 38
METH=5 LAB=52 — -
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
5.5
5.0
4.5
1
o
g
0
f
4.0
r
e
p 3.5 +
o
r
t
e
d
3.0
P
P
H
2.5 +
2.0
1.5 +
P P
D D
E P
P P
1.5 2.0
2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal PPM
NOTE: 9 obs hidden.
-------
2.00
1.75
1.50
plotrec.sas 7:49 Monday, August 17, 1992 39
- METH=5 LAB=53
Plot of REC*LOGTRUE. Symbol is value of MTX.
1.25 -
r
e
c
0
D
D D
D
v 1.00 +
e
r
y
0.75 •
E
E
0.50
0.25
0.00
P
P
1.5 2.0
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal cone (ppm)
5.0 5.5
NOTE: 5 obs hidden.
-------
5.5
5.0 +
4.5 +
2.5 +
2.0 +
1.5
plotcon.sas 7:38 Monday, August 17, 1992 39
METH=5 LAB=53 - —
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
1
0
g
4.0 +
0
f
r
e
P P
p 3.5 +
0
r
t
e
d
D
D
3.0 4- E
P
P
M
E
D
D D
P
P P
1.5
2.0
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal PPM
5.0 5.5
NOTE: 6 obs hidden.
-------
2.00
1.75 +
1.50
1.25
0.75
0.50
0.25
0.00
plotrec.sas 7:49 Monday, August 17, 1992 40
METH=5 LAB=54 -
Plot of REC*LOGTRUE. Symbol is value of MIX.
r
e
c
0
D
D 0
v 1.00 +
e
r
y
D
P P
D D
P
P P
1.5 2.0
NOTE: 7 obs hidden.
2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal cone (ppm)
-------
plotcon.sas 7:38 Monday, August 17, 1992 40
— METH=5 LAB=54 -—
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
5.5 +
5.0 +
4.5
1
o
g
0
f
4.0 +
r
e
p 3.5
o
r
t
e
d
3.0 +
P
P
M
2.5 +
2.0
1.5
D D
D
P P
D D
P P
1.5 2.0
2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal PPM
NOTE: 9 obs hidden.
-------
plotrec.sas 7:49 Monday, August 17, 1992 41
METH=5 LAB=55
Plot of REC*LOGTRUE. Symbol is value of MIX.
2.00
1.75
1.50
1.25
r
e
c
0
D
v 1.00 +
e
r
y
D
E
DO E
0.75 +
0.50
0.25
0.00
P
p p
P
D D
P
P P
P
1.5 2.0
2.5 3.0 3.5 4.0 4.5
base 10 log of nominal cone (ppm)
5.0 5.5
NOTE: 3 obs hidden.
-------
plotcon.sas 7:38 Monday, August 17, 1992 41
METH=5 LAB=55 - -
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
5.5
5.0 +
4.5 +
1
o
g
0
f
4.0
r
e
p 3.5 +
o
r
t
e
d
3.0 +
P
P
M
2.5
2.0 +
1.5
D
D D
P P
0 D
EP P
1.5 2.0
2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal PPM
NOTE: 9 obs hidden.
-------
2.00
1.75
1.50
1.25
0.75
0.50
0.25
0.00
plotrec.sas 7:49 Monday, August 17, 1992 42
METH=5 LAB=56 -
Plot of REC*LOGTRUE. Symbol is value of MIX.
r
e
c
0
v 1.00 +
e
r
y
D
D D
D
D
D D
Q
Q
p
p P
1.5 2.0
2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal cone (ppm)
NOTE: 4 obs hidden.
-------
plotcon.sas 7:38 Monday, August 17, 1992 42
Plot of LOGCONC*LOGTRUE. Symbol is value of MIX.
5.5
5.0
4.5
1
o
g
0
f
4.0
r
e
p 3.5 +
o
r
t
e
d
3.0
P
P
M
2.5 +
2.0
D
D D
1.5 +
l +
1.5 2.0
P P
D
D D
P p
2.5 3.0 3.5 4.0 4.5 5.0 5.5
base 10 log of nominal PPM
NOTE: 8 obs hidden.
-------
Appendix G-8
Plots of Repeatability/Reproducibility
versus Lead Concentration
-------
I
10 +
4
1
9 +
2
3
A
R 2
I 3
TH 7 +
M
0 4
F 2
6 +
R
E 1
P 2
A 5 "*" 4
T
A
6
I
i 4 -I- 3
T ,«
1
4 5
3 +
3
2
4
6
8
10
12
LOGARITHM OF METHOD MEAN
Plot of Log Repeatability versus Log Method Mean.
Legend
1 = MS/AAS
2 = HP/AAS
3 = MW/ICP
4 = HP/ICP
5 = Lab XRF
-------
s
o
f
X
I
I
o>
DC
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
I
0.0
HP/MS
MW/ICP
MW/AAS
HP/MS
Lab XRF
I
0.2 0.4 0.6 0.8 1.0
Method Mean x 10's (|ig/g)
Figure 1. Repeatability versus lead concentration by method.
1.2
-------
11 +
L
0 10 +
G
A
R
T
H
M
0
F
9 +
8 +
7 +
R
E
P
R 6 +
0
D
U
C 5 +
I
B
I
L 4
I
T
Y
5
3
3
2
421
3
5 1
3
2
5
2
5 1
3
2
2 +
I
6 8 10
LOGARITHM OF METHOD MEAN
12
Plot of log of Reproducibility versus log of Method Mean (pg/g).
-------
1.8
X
£
5
u
Q.
0)
cc
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
HP/ICP
Lab XRF
MW/AAS
HP/AAS
MW/ICP
I
I
I
0.0
0.2
0.8
1.0
0.4 0.6
Method Mean x
Figure 2. Reproducibility versus lead concentration by method.
1.2
-------
Appendix G-9
Geometric Mean Recovery by Method
-------
-H-
-l-
-L-
U-
-M---U-
56 +-- - - L—
55 + L M
54 +L—M—-U
53 + -L M- -U
52 +
51 + - L
50 +
48 +
46 + L- — —
45 + -
44 +--
43 +
42 +
41 + - L
40 +
M 38 +
E 36 + -
T 35 + -
H 34 +
C 33 + L M—U-
0 32 + -
N 31 +— -
30 +
28 *
-M---U-
-L-
-M-
-M—U-
.„U—
-M-
-U-
-U-
M-
26 +-
25 +-
24 +-
23 +-
22 +•
21 +-
20 +
18 +
16 +-
15 +-
14 +-
13 +-
12 +-
11 +-
-H-
-M-
-L-
-U-
-M—U-
-M
-U-
-M-
+ + __ +__ _ + __ _+ + __+ +_
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3
95% conf int for geom mean rec
95% Confidence Intervals for Geometric Mean Recovery
for Each Method and Sample.
LEGEND
95% Confidence Interval for Geometric
Mean Recovery
L = Lower Limit
M = Mean
U = Upper Limit
METHCON
The first digit denotes method
number
1= MW/AAS
2= HP/AAS
3= MW/ICP
4 = HP/ICP
5 = Laboratory XRF
The second digit denotes rank of
concentration for sample
1 = Low Dust
2= DustSRM
3 = Low Paint
4 = High Dust
5 = High Paint
6= Paint SRM
-------
1 .(. 1
G
E
0 11
V 1 • * '
M
E
T
R
1
C
M
1 A
r- 1 .V
A
N
R
E
C
0
Vfi Q •
v » y '
E
R
Y
Op .
• O
1
1
1 1
L ...... _ . .. 1 ........... .. . . ,
1
1
2
3
2
2
2
2 4
3
4 5
3 3
4
3
4 4
5
L ,, ...... ..................
2
5
4
L _ _C C > C
+
1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
BASE 10 LOG OF CONSENSUS VALUE
Demonstration of method effect via geometric mean recovery
versus log consensus value (Method 5 censored at .8).
-------
1
2
L
0
G
A
R
I
T
H
M
11
10 +
0
F
M
E
T
H
0
D
1
3
1
2
1
2
M
E
A
N
6 +
(ng/g)
1
2
4
6 7 8 9 10 11 12
LOGARITHM OF CONSENSUS VALUE
Legend
•t _
2-
3 =
4 =
5 =
MS/AAS
HP/AAS
MW/ICP
HP/ICP
LabXRF
Log of Method Mean versus Consensus Value.
-------
Appendix G-10
Method Effects and Pairwise
Comparison of Method Means
-------
Tests for method effects and pairwise comparison of method means.
The overall F-tests for significance of method effects were not significant
for the low dust sample (p = .44), were only marginally significant for the paint
SUM (p = .08), but were highly significant on the other four samples (p < .001 in
all cases).
For pairwise comparisons of method means within each of the six samples,
ordinary nonsimultaneous t-tests at the 5% significance level were used. There
are ten possible paired comparisons of methods within each of the six samples, so
that three false rejections of the hypothesis of no difference would be expected by
chance alone.
The results of the pairwise comparisons are summarized below. No
differences were declared in connections with the low dust sample, and only two
differences were declared on the paint SRM samples. It is clear from the table
below that the differences primarily involve methods 1 and 5. Of 28 declared
differences, 26 involve methods 1 and 5. These results confirm those obtained by
the simple nonparametric logic, namely, method 1 is generally higher and method
5 is generally lower than the other methods. There are, of course, exceptions,
notably the low dust sample.
-------
Results of sample-specific pairwise method comparisons.
M E
1
2 CDE
3 ACDE
4 ACDEF
5 ACDE
T
2
XXXXX
X
A
None
ACDE
H O D
3 4
XXXXXXXXXXXXXX
XXXXXXXXXXXXXX
xxxxxxxxxx
xxxxxxx
F XXX
ADE ADE
Table entries indicate samples for which method comparisons are significantly
different using ordinary nonsimultaneous t-tests at the 5% significance level. For
instance, methods 3 and 5 were declared different on samples A, D, E.
Legend
A = High Dust
B = Low Dust
C = Dust SRM
D = High Paint
E = Low Paint
F = Paint SRM
-------
Tests for method effects
Several other effects are suggested. In addition to the facts that MW/AAS
is uniformly higher and XRF uniformly lower than the other methods, there
appear to be other effects due to analytic method or extraction method, as
indicated by the results of comparisons using the SAS General Linear Model
procedure. These comparisons were limited to non-XRF methods. Low p-values
indicate significant effects.
Tests for effect of method of analysis, by matrix and method of extraction
Extraction Matrix p-value
MW dust <.01
MW paint <.01
HP dust .06
HP paint .36
Tests for effect of method of extraction by matrix and method of analysis
Analysis Matrix p-value
AAS dust .02
AAS paint <.01
ICP dust .92
ICP paint .03
-------
Appendix H
Total Microwave Digestion Method
-------
RTI Method for Total Digestion of Lead in Paint and Dust
Procedure 1: U. S. Fish and Wildlife Service Digestion
• Weigh 100 mg of ground paint into a clean Teflon digestion vessel.
Add 5 mL of cone. HNO3 and 1 mL of 49% HP.
• Cap the vessel and microwave at the following conditions:
- 3 min at 255 power,
- 3 min at 50% power,
- 3 min at 100% power.
• Allow solution to cool to room temperature; uncap Teflon digestion
vessel.
• Evaporate residue to a volume of 2 - 3 mL.
Procedure 2: Institute of Chemical Industry and Metallurgy of China Digestion
Prepare 12 digestates as follows:
• Transfer contents from Procedure 1 into a 120 mL Teflon PFA vessel,
rinsing walls of vessel with DI water.
Add 10 mL cone. HC1 and 0.5 mL HP.
• Microwave at the following conditions for ICP analysis:
- 10 minutes at 80% power,
- 8 minutes at 60% power, or
Microwave at the following conditions for AAS analysis:
- 10 minutes at 80 % power,
- 5 minutes at 60% power.
• Allow solution to cool to room temperature; uncapTeflon digestion
vessel.
• Add 6 mL of 4% boric acid, and 15 ml of cone. HC1.
• Transfer to 100 mL volumetric flask and dilute to volume.
10-4
-------
U.S. Fish and Wildlife Service Procedure
DIGESTION OF ANIMAL TISSUE
Method 201 - ICP
Digestion of Animal Tissue
Metals of Reference: Al, Sb, Ba, Be, B, Cd, Co, Cr, Cu, Fe, Pb, Mg, Mn, Mo, Ni, Ag,
Sr, Sn, V, Zn
1.0 Reagents
1. Concentrated nitric acid - instra-analyzed
2. Source of laboratory pure water; Type II, etc.
1.1. Materials and Apparatus
1. CEM MDS-81D microwave oven
2. Top loader analytical balance accurate to 0.001 grams
3. 120 mL digestion vessels - PFA Teflon
4. 50 mL polypropylene volumetric flasks
5. 60 mL polypropylene sample bottles
6. Disposable polypropylene funnels - 55 mm
1.2 Method
1. Weigh out 0.5 grams freeze-dried, homogenized material accurately to
0.001 grams into a clean 120 mL microwave digestion vessel.
2. Add 5 mL Baker Instra-Analyzed concentrated nitric acid.
3. Place cap on vessels and torque to 12 ft-lbs using CEM capping station
or torque wrench.
4. Place vessels onto turntable and load into CEM MDS-81D microwave
oven.
5. Heat the vessels:
a) 3 minutes at 20% power
b) 3 minutes at 50% power
c) 15 minutes at 75% power
-------
6. Upon completion of heating cycle, wait 1 minute, then remove vessels from
oven and cool in a fume hood.
7. When cool, uncap vessels using capping station and carefully evaporate vessel
contents to 0.5 - 1.0 mL residue and dilute to 10 mL with deionized water.
-------
Determination of Si, Al, Ca, Mg, Fe, Ti, Mn, Cu, Ci and Ni In
Vanadium • Titanium • Iron Ore by Microwave Oven Digestion,
ICP, AA and Chemical Analysis Methods
Li Bao-hou
Yu Zhong-quan
Han Kai
Institute of Chemical Industry and Metallurgy
The Academy of Sciences of China
June 1988
Beijing, China
-------
Institute of Chemical Industry and Metallurgy
Acid Digestion of Samples by Microwave Oven
1. Standard Samples of Pan Zhi Hua, Academy of Iron and Steel, Ministry of
Metallurgy, China
BH 0102 vanadium - titanium fine ore
BH 0104 titanium fine ore
2. Microwave Oven Equipment:
Model MDS - 81D Microwave Oven (product of CEM, U.S.A.) with capping
station, cooling groove and 120 mL Teflon PFA vessel
Settings of MDS - 8 ID operation program:
ICP ANALYSIS AA ANALYSIS
Time Power Time Power
Program 1: 10 minutes 80% 10 minutes 80%
Program 2: 8 minutes 60% 5 minutes 60%
3, Methods of Sample Dissolution:
Put 0.1 gram accurately weighed standard sample (BH 0102) into a 120 mL
Teflon PFA vessel, rinse the wall of the vessel with small quantity of deionized water
and add 10 mL concentrated hydrochloric acid and 0.5 mL hydrofluoric acid. Secure
the safety valve on the vessel and tighten the vessel cap on the capping station.
Place the vessel on the carousel and connect the exhaust tubes. The operation for the
BH0104 standard sample is the same as above except 10 mL concentrated acid and
2 mL hydrofluoric acid are added.
------- |